PDA

View Full Version : New PECOTA's are out



PaulieIsAwesome
01-15-07, 11:43 PM
I don't think I can post them all, but just to post a few Yankee highlights:

Phil Hughes 130 IP, 3.91 ERA. By VORP, he'd be the 21st best pitcher in the AL next year. That'd be awesome.

In general, of the top 25 pitchers in the AL, 3 are Yankees, Moose, Pettitte, Hughes, and 3 are Red Sox, Curt, Matsuzaka and Papelbon. Josh Beckett's not that far from Hughes.

They've got Jeter with another great year: by WARP, 3rd best player in baseball behind Pujols and Mauer. A-Rod's 8th, Ortiz is first Red Sock on the list, at 18th. They've got Cano staying hot, .308/.345/.472, 27th best player in baseball. I'll run a couple of numbers to see where this generally puts the Yankees compared to the Sox.

gdn
01-15-07, 11:52 PM
Interesting!

PaulieIsAwesome
01-15-07, 11:52 PM
Oh, and the worst free agent contracts? Juan Pierre, who PECOTA has with a .707 OPS, Gil Meche, with a 5.48 ERA. Zito's isn't that good, with a 4.22 ERA in the NL. Soriano turns out ok: .918 OPS.

26 and counting
01-16-07, 12:00 AM
What are VORP and WARP? I know that VORP stands for "Value over Replacement Player", and WARP stands for "Wins Above Replacement Player". But how are they measured, and how good indicators of determining player success are they?

PaulieIsAwesome
01-16-07, 12:09 AM
What are VORP and WARP? I know that VORP stands for "Value over Replacement Player", and WARP stands for "Wins Above Replacement Player". But how are they measured, and how good indicators of determining player success are they?

VORP takes runs created, or some similar formula for total run value, and compares that run total to what a bench player at the same position would create. It's a good statistic. It makes sense and describes total offensive value pretty well.

WARP is VORP plus defense, translated into wins rather than runs, an attempt to quantify total wins added. WARP is not great at doing what it does: Baseball Prospectus' defensive statistics are black box (they don't show their data or their methods) and it's known that they are not based on play by play data, so their defensive stats are pretty weak.q

Hughes2.50
01-16-07, 12:29 AM
Well this is interesting. They have Hughes at 3.91. That won't be close.<p>Paulie, if you have Sanchez and Clippard #s can you post them.

Hughes2.50
01-16-07, 12:33 AM
I don't think I can post them all, but just to post a few Yankee highlights:

Phil Hughes 130 IP, 3.91 ERA. By VORP, he'd be the 21st best pitcher in the AL next year. That'd be awesome.

In general, of the top 25 pitchers in the AL, 3 are Yankees, Moose, Pettitte, Hughes, and 3 are Red Sox, Curt, Matsuzaka and Papelbon. Josh Beckett's not that far from Hughes.

They've got Jeter with another great year: by WARP, 3rd best player in baseball behind Pujols and Mauer. A-Rod's 8th, Ortiz is first Red Sock on the list, at 18th. They've got Cano staying hot, .308/.345/.472, 27th best player in baseball. I'll run a couple of numbers to see where this generally puts the Yankees compared to the Sox.What equivalent ERA stat are you posting, the PERA or the ERA? I'd prefer seeing the PERA if you are posting the ERA.

PaulieIsAwesome
01-16-07, 12:45 AM
Well this is interesting. They have Hughes at 3.91. That won't be close.Paulie, if you have Sanchez and Clippard #s can you post them.




This was all ERA. Hughes' PERA was 3.80.


Clippard had 137 IP, 5.14 ERA, 5.03 PERA.

Sanchez had 122 IP, 5.16 ERA, 5.04 PERA.



Hughes2.50, do you really think that Hughes will have an ERA lower than 3.91, or 3.80? Jake Peavy took 2 years to get really good on the big league level, Brad Penny took a couple of years: they're two of his top comparables.

flymick24
01-16-07, 12:51 AM
This was all ERA. Hughes' PERA was 3.80.


Clippard had 137 IP, 5.14 ERA, 5.03 PERA.

Sanchez had 122 IP, 5.16 ERA, 5.04 PERA.



Hughes2.50, do you really think that Hughes will have an ERA lower than 3.91, or 3.80? Jake Peavy took 2 years to get really good on the big league level, Brad Penny took a couple of years: they're two of his top comparables.

i'll take a healthy 3.80 from hughes... that could easily be ROY numbers.

as for clippard and sanchez, i'm not holding my breath on either of those projections.

PaulieIsAwesome
01-16-07, 12:52 AM
i'll take a healthy 3.80 from hughes... that could easily be ROY numbers.

as for clippard and sanchez, i'm not holding my breath on either of those projections.

Both of those are equivalent on the major league level.

gdn
01-16-07, 12:54 AM
What's a PERA?

This basically means that if they were both to pitch in the majors next year, they'd post almost identical stats....interesting.

flymick24
01-16-07, 12:57 AM
What's a PERA?

This basically means that if they were both to pitch in the majors next year, they'd post almost identical stats....interesting.

it's the ERA projected from the pitcher's peripherals (EqH9, EqHR9, EqBB9, EqK9)

gdn
01-16-07, 12:58 AM
it's the ERA projected from the pitcher's peripherals (EqH9, EqHR9, EqBB9, EqK9)Thanks, that makes sense.

PaulieIsAwesome
01-16-07, 12:58 AM
What's a PERA?

This basically means that if they were both to pitch in the majors next year, they'd post almost identical stats....interesting.

It's peripheral ERA: only based on HRs, Hs, Ks, and BBs, so it's like a luck independent (or DIPS) version of PECOTA ERA.

flymick24
01-16-07, 01:01 AM
Both of those are equivalent on the major league level.

the PECOTA gizmo machine was clearly malfunctioning while printing out clippard and sanchez's numbers. :(

will carroll probably broked it.

PaulieIsAwesome
01-16-07, 01:05 AM
the PECOTA gizmo machine was clearly malfunctioning while printing out clippard and sanchez's numbers. :(

will carroll probably broked it.

To break down some more of the periphs:

Sanchez: 6.4 K/9, 4.1 BB/9, 1.1 HR/9; Clippard: 6.6 K/9, 3.4 BB/9, 1.4 HR/9

Hughes2.50
01-16-07, 01:11 AM
Hughes2.50, do you really think that Hughes will have an ERA lower than 3.91, or 3.80? Jake Peavy took 2 years to get really good on the big league level, Brad Penny took a couple of years: they're two of his top comparables.Yes I do. My MLE DIPS ERA for Hughes is 2.95. So there is a considerable difference in what I expect from my analysis and what Pecota posted. I don't see Brad Penny as any reasonable comparable for Hughes. <p>At 20 Penny had a 2.96 ERA in A+ ball, and at 21 he was way over 4.00 in AA. Peavy is a better comparable, but not close enough to provide a true, close to equivalent comparison. For example at 20 years old, Peavy only pitched 28 innings. He pitched again at 21 in the same league (Mobile - Southern, AA) forabout 80 more innings and immediately got promoted and pitched 100 innings in the majors the same year. Do you see the problem? 30 innings one year, nearly 200 the next - do you think tired arm and other issues were in play that year, and the next for Peavy, affecting his performance? I do.<p>For any sabermetrically savvy people, check out Hughes' defense independent statistics, find a formula to calculate ERA for it and compare that number with Hughes' actual ERA for both Tampa and Trenton. I trust that you will find that for both it is significantly lower than the regular ERA he actually had. That should lead anyone doing that kind of analysis to question how Pecota's PERA of 3.80 and ERA of 3.91 can be so close. <p>Paulie on BP's non-paying website they list all equivalent PERA stats and ERA stats as associated with particular minor league teams for a given year. Where do they derive the 3.80 and 3.91 figures from? Tampa, Trenton, or cumulative?

Hughes2.50
01-16-07, 01:16 AM
It's peripheral ERA: only based on HRs, Hs, Ks, and BBs, so it's like a luck independent (or DIPS) version of PECOTA ERA.Thanks for reminding me with this post. The biggest problem I have with Pecota's analysis of PERA stats is that they don't conform to my idea of truly independent of fielding statistic. Bases on Balls are, strikeouts are, home runs are, but other hits aren't necessarily. The truest form of such analytical approaches doesn't use hits but does include hit batters along with bases on balls and the others.

PaulieIsAwesome
01-16-07, 01:32 AM
Yes I do. My MLE DIPS ERA for Hughes is 2.95. So there is a considerable difference in what I expect from my analysis and what Pecota posted.For any sabermetrically savvy people, check out Hughes' defense independent statistics, find a formula to calculate ERA for it and compare that number with Hughes' actual ERA for both Tampa and Trenton. I trust that you will find that for both it is significantly lower than the regular ERA he actually had. That should lead anyone doing that kind of analysis to question how Pecota's PERA of 3.80 and ERA of 3.91 can be so close.
Paulie on BP's non-paying website they list all equivalent PERA stats and ERA stats as associated with particular minor league teams for a given year. Where do they derive the 3.80 and 3.91 figures from? Tampa, Trenton, or cumulative?



Ok, but if Hughes puts up a 2.95 ERA in 130 IP in the bigs, that would make him one of the top 10 pitchers in baseball according to PECOTA. Could you run your MLEs on Justin Verlander's 2005? In 2005, in 86 A+ innings, Verlander had 10.9 K/9, 1.99 BB/9, .314 HR/9, with a 1.67 ERA, and in 32.7 AA IP, 8.8 K/9, 1.92 BB/9, .276 HR/9, for a .28 ERA.

Hughes last year threw 30 A+ Innings, 9 K/9, .6 BB/9, and 0 HR/9, with a 1.80 ERA, and 116 innings, 10.7 K/9, 2.48 BB/9, .388 HR/9 with a 2.25 ERA. I think that puts them pretty close. Maybe I'd give Hughes a little advantage, from throwing more innings, at AA. But I can't think that Hughes will do much better than Verlander did last year, and expect him to have the same problem Verlander had, moving from 145 innings to 170-180 innings total, and to struggle some at the end of the year. I'd be really surprised for Hughes to have an ERA below 3.4 with more than 130 IP. Just 4 AL pitchers did that last year. He's going to have some growing pains, like Liriano (injuries) and Felix (unluckiness, mainly: his defense independent stats were some of the best in baseball.)

PaulieIsAwesome
01-16-07, 01:41 AM
Here's Hughes list of Comps (it doesn't have the full list of comps and years, I assume those won't come until they update the cards):

Bobby Bradley: This would suck, and is almost certainly the source of the negative parts of Hughes' projection. Bradley was an always good pitching prospect for Pittsburgh, rising to A+ ball in 2001 at the age of 20, but getting hurt: over and over and over again.

Peavy, Penny, we discussed these.

Edwin Jackson: A similar problem as Bradley, though Jackson was even better. For a couple of years there, he was considered a dominant prospect, and he has not been able to pitch at all effectively in the majors.

Hughes2.50
01-16-07, 01:50 AM
Ok, but if Hughes puts up a 2.95 ERA in 130 IP in the bigs, that would make him one of the top 10 pitchers in baseball according to PECOTA. Could you run your MLEs on Justin Verlander's 2005? In 2005, in 86 A+ innings, Verlander had 10.9 K/9, 1.99 BB/9, .314 HR/9, with a 1.67 ERA, and in 32.7 AA IP, 8.8 K/9, 1.92 BB/9, .276 HR/9, for a .28 ERA.

Hughes last year threw 30 A+ Innings, 9 K/9, .6 BB/9, and 0 HR/9, with a 1.80 ERA, and 116 innings, 10.7 K/9, 2.48 BB/9, .388 HR/9 with a 2.25 ERA. I think that puts them pretty close. Maybe I'd give Hughes a little advantage, from throwing more innings, at AA. But I can't think that Hughes will do much better than Verlander did last year, and expect him to have the same problem Verlander had, moving from 145 innings to 170-180 innings total, and to struggle some at the end of the year. I'd be really surprised for Hughes to have an ERA below 3.4 with more than 130 IP. Just 4 AL pitchers did that last year. He's going to have some growing pains, like Liriano (injuries) and Felix (unluckiness, mainly: his defense independent stats were some of the best in baseball.)Are you aware that Verlander was shut down after 119 innings in '05 with a tired arm? <p>And that he experienced the same thing after about 130 innings in the majors in '06 (where after that the Tigers had to really manage his innings and time between starts carefully). The point being that Verlander had a 2.67 ERA or so, before he started getting tired arm. After that his ERA climbed precipitiously as the Tigers tried to manage his exposure. In any event, I have Verlander's MLE DIPS ERA for A+ at 4.02 and at AA = 3.12, the weighted cumulative MLE DIPS ERA = 3.76 off Verlander's '05 season. Still, he was pitching closer to his AA results before tired arm set in.<p>Contrast that with Hughes who in '06 pitched 152 innings 30 in A+, 116 in AA, and 6 in the postseason without any issues whatsoever. I believe the Yankees will limit Hughes to a total of 180 innings between the season and the postseason, and the minors and majors. And, I believe that he will not experience arm fatigue or other issues due to some extent by the well coordinated exposure that Eiland allowed Hughes to undergo in '06. The long and short of it is that yes, Verlander's '05 and Hughes' '06 were similar statistically. And, unlike Verlander who hit the wall due to improper preparation by the Tigers pitching handlers, I predict that Hughes will be able to sustain his level of excellence throughout the entire season and post-season.

Hughes2.50
01-16-07, 02:10 AM
The two best comparable for Hughes based on level of accomplishment at a similar age, and stuff along with pitching maturity are Dwight Gooden and Pedro Martinez. To compare Philip Hughes with guys like Penny, Bradley, Jackson,and even Peavy is to totally miss what makes Hughes able to do what he has been able to do.<p>Like Gooden and to a lesser extent Martinez, the 20 year old Hughes' greatest asset lies between his ears. His knowledge of pitching, his awareness of what he can and cannnot do at any particular time, and his command of his stuff puts him head and shoulders above other really good prospects - just like Martinez, and especially Dwight Gooden were. If you want a true comparable, check out what Dwight Gooden did before he was 21. <p>I don't think that Hughes will put up those numbers because he pitches in a league with a DH and it is a tougher league, and because the Yankees won't expose him to as many innings or starts. But if you want to know how good Philip Hughes is as a 20/21 year old major leaguer, look no farther than the 20 year old Dwight Gooden.

PaulieIsAwesome
01-16-07, 02:13 AM
Are you aware that Verlander was shut down after 119 innings in '05 with a tired arm? And that he experienced the same thing after about 130 innings in the majors in '06 (where after that the Tigers had to really manage his innings and time between starts carefully). The point being that Verlander had a 2.67 ERA or so, before he started getting tired arm. After that his ERA climbed precipitiously as the Tigers tried to manage his exposure.
Contrast that with Hughes who in '06 pitched 152 innings 30 in A+, 116 in AA, and 6 in the postseason without any issues whatsoever. I believe the Yankees will limit Hughes to a total of 180 innings between the season and the postseason, and the minors and majors. And, I believe that he will not experience arm fatigue or other issues due to some extent by the well coordinated exposure that Eiland allowed Hughes to undergo in '06. The long and short of it is that yes, Verlander's '05 and Hughes' '06 were similar statistically. And, unlike Verlander who hit the wall due to improper preparation by the Tigers pitching handlers, I predict that Hughes will be able to sustain his level of excellence throughout the entire season and post-season.



I'd forgotten about the shutdown in 2005, and you're right about that. Maybe Hughes does pitch like Verlander and sustain that, finishing with an ERA like Verlander had before mid-August, 3.2 or something.

But I still think it's very unlikely. Assuming Hughes pitches exactly like he did last year in his transition to the big leagues, he puts up a 2.95 ERA. That's if pretty much everything that went exactly right last year goes right again this year.

But if a couple of small things go wrong for him this year (walk rate goes up a little, he's forced to throw more innings because of an unexpected injury to another Yankees starter, he gets hurt: and this is the biggest one, he's still not through the injury nexus yet,) his ERA could easily balloon to 4.2 or 4.4. That's why I think an ERA prediction of 3.8 or so is probably correct. And, his FIP ERA, based on his periphs from PECOTA, predict a 3.91 FIP ERA.


Once they post his PECOTA card, we'll be able to see what his 90th degree PECOTA percentile is, and see how much uncertainty is built into Hughes profile. For a taste, he's in the top quartile for COLLAPSE rate among the pitchers in the PECOTA spreadsheet.

Just for a quick reason for why I'm not totally convinced: Top 3 Pitching prospects from Baseball America since 1990:

1990: Steve Avery, Ben McDonald, Kiki Jones
1991: Todd Van Poppel, Roger Salkeld, Arthur Rhodes
1992: Brien Taylor, Todd Van Poppel, Roger Salkeld
1993: Brien Taylor, Todd Van Poppel, Jason Bere
1994: James Baldwin, Jose Silva, Darren Dreifort
1995: Armando Benitez, Bill Pulsipher, Alan Benes
1996: Paul Wilson, Alan Benes, Livan Hernandez
1997: Kerry Wood, Matt White, Kris Benson
1998: Kerry Wood, Matt White, Kris Benson
1999: Rick Ankiel, Bruce Chen, Brad Penny
2000: Rick Ankiel, Ryan Anderson, John Patterson
2001: Josh Beckett, Jon Rauch, Ben Sheets
2002: Josh Beckett, Mark Prior, Juan Cruz
2003: Jesse Foppert, Jose Contreras, Gavin Floyd
2004: Edwin Jackson, Greg Miller, Scott Kazmir
2005: Felix Hernandez, Scott Kazmir, Matt Cain

There are some pretty decent pitchers on that list, and some serious disappointments too. I think Phil's going to go through growing pains just like all of them.

Intrepid22
01-16-07, 02:16 AM
I don't think I can post them all, but just to post a few Yankee highlights:

Phil Hughes 130 IP, 3.91 ERA. By VORP, he'd be the 21st best pitcher in the AL next year. That'd be awesome.

In general, of the top 25 pitchers in the AL, 3 are Yankees, Moose, Pettitte, Hughes, and 3 are Red Sox, Curt, Matsuzaka and Papelbon. Josh Beckett's not that far from Hughes.

They've got Jeter with another great year: by WARP, 3rd best player in baseball behind Pujols and Mauer. A-Rod's 8th, Ortiz is first Red Sock on the list, at 18th. They've got Cano staying hot, .308/.345/.472, 27th best player in baseball. I'll run a couple of numbers to see where this generally puts the Yankees compared to the Sox.


any particular reason wang isnt a top 25 AL pitcher

PaulieIsAwesome
01-16-07, 02:23 AM
The two best comparable for Hughes based on level of accomplishment at a similar age, and stuff along with pitching maturity are Dwight Gooden and Pedro Martinez. To compare Philip Hughes with guys like Penny, Bradley, Jackson,and even Peavy is to totally miss what makes Hughes able to do what he has been able to do.Like Gooden and to a lesser extent Martinez, the 20 year old Hughes' greatest asset lies between his ears. His knowledge of pitching, his awareness of what he can and cannnot do at any particular time, and his command of his stuff puts him head and shoulders above other really good prospects - just like Martinez, and especially Dwight Gooden were. If you want a true comparable, check out what Dwight Gooden did before he was 21.
I don't think that Hughes will put up those numbers because he pitches in a league with a DH and it is a tougher league, and because the Yankees won't expose him to as many innings or starts. But if you want to know how good Philip Hughes is as a 20/21 year old major leaguer, look no farther than the 20 year old Dwight Gooden.



Where's the evidence that Hughes is on the same plane as Gooden? In A+ ball at the age of 18, Gooden struck out 300 batters in 191 innings pitched. That's just absurd. Park factors and era factors bring that down some, but to argue that Phil Hughes' 10 K/9 could equal Gooden's 14 K/9 seems ludicrous to me. Re: Pedro: Pedro's minor league numbers don't really jump out at me. A K/9 rate right around 9, good walk numbers. But his FIP numbers in his last minor league season, his 20 year old season at Albuquerque was only 3.62.

It's possible that the way Phil approaches the game is as good as Gooden or Pedro's. But that's a really subjective judgement. I don't see how Phil's numbers are appreciably better than Josh Beckett's minor league stat line (in fact, Beckett's is better,) or Mark Prior's college stat line (Prior's is probably better) or Rick Ankiel's. There's a lot of variance in minor league pitchers lines, and PECOTA needs to put that variance into their model. And it spits out a 3.80 PERA.

PaulieIsAwesome
01-16-07, 02:25 AM
any particular reason wang isnt a top 25 AL pitcher

They don't like Wang's strikeout rate, and see his ERA going up to 4.35.

Hughes2.50
01-16-07, 02:33 AM
They don't like Wang's strikeout rate, and see his ERA going up to 4.35.Even with Wang's strikeout rates his DIPS ERA + was 117 compared to the 122 it was measured off his era. Pecota doesn't have to like his strikeout rates, they ought to love his BB/9 and HR/9 rates though.

PaulieIsAwesome
01-16-07, 02:38 AM
Even with Wang's strikeout rates his DIPS ERA + was 117 compared to the 122 it was measured off his era. Pecota doesn't have to like his strikeout rates, they ought to love his BB/9 and HR/9 rates though.

They have his walk rate and home run rate both going up, his walk rate minorly, his home run rate quite a bit, giving him a DIPS ERA of 4.35.

Hughes2.50
01-16-07, 02:55 AM
Where's the evidence that Hughes is on the same plane as Gooden? In A+ ball at the age of 18, Gooden struck out 300 batters in 191 innings pitched. That's just absurd. Park factors and era factors bring that down some, but to argue that Phil Hughes' 10 K/9 could equal Gooden's 14 K/9 seems ludicrous to me. Re: Pedro: Pedro's minor league numbers don't really jump out at me. A K/9 rate right around 9, good walk numbers. But his FIP numbers in his last minor league season, his 20 year old season at Albuquerque was only 3.62.

It's possible that the way Phil approaches the game is as good as Gooden or Pedro's. But that's a really subjective judgement. I don't see how Phil's numbers are appreciably better than Josh Beckett's minor league stat line (in fact, Beckett's is better,) or Mark Prior's college stat line (Prior's is probably better) or Rick Ankiel's. There's a lot of variance in minor league pitchers lines, and PECOTA needs to put that variance into their model. And it spits out a 3.80 PERA.Any variance is correctly accounted for in a confidence interval around an aveage projection which should be reported. Therefore, the correct interpretation of Pecota's 3.80 projection is that is what they think his average result would be if there were many trials of a 2007 season. A large variance would mean a wider confidence interval. Less variance a narrower confidence interval around that 3.80 value.<p>Those K rates you are hung up on (i.e., "but to argue that Phil Hughes' 10 K/9 could equal Gooden's 14 K/9 seems ludicrous to me") are not terribly germane to this topic, because even Pecota concedes that Hughes will likely put up some pretty strong k rates as did Gooden. And of course, I could go on and on about what is the most important statistical consideration(s) when it comes to projecting pitchers but that is for another day.<p>What Pedro did at San Antonio should have caught your eye. Still, yes Pedro, but even moreso, the best comparable for Philip Hughes is Dwight Gooden. To compare someone with the composure and maturity that Philip Hughes has as a pitcher with Josh (I'm only gonna throw my heater in the AL) Beckett and Rick Ankiel is so off the mark I just hope someone as bright as you never brings them up again. I concede that both Beckett and Ankiel were very promising, just as Ryan Leaf was a very promising quarterback once upon a time, but like Leaf both Beckett (to a lesser extent) and Ankiel (to a much greater extent - sadly) have major issues in the area of makeup, which prevent them from reaching their awesome potential. Philip Hughes doesn't share much in common with those two, but he does share a lot in common with the young Dwight Gooden, and older pitchers like Clemens, Halladay, and the Mussina's of the world who are impeccable professionals who know their strengths and weaknesses as well as their opponents. Hughes already brings that to the table at the tender age of 20.

Hughes2.50
01-16-07, 02:59 AM
They have his walk rate and home run rate both going up, his walk rate minorly, his home run rate quite a bit, giving him a DIPS ERA of 4.35.Arbitrary inflation of a pitchers peripherals while he is approaching peak performance period. I'd say their projections are an excessively pessimistic interpretation of Wang's 2007 performance.

Reggievision
01-16-07, 12:46 PM
I've got to say, Hughes 2.50 has done more than his share of homework in projecting Hughes in pinstripes. Make sure to keep all these threads archived; frankly, if Phillip Hughes even comes close to what H2.50's models indicate, I'll be the first name on the petition to get him a job in the Bronx.

PaulieIsAwesome
01-16-07, 01:39 PM
Any variance is correctly accounted for in a confidence interval around an aveage projection which should be reported. Therefore, the correct interpretation of Pecota's 3.80 projection is that is what they think his average result would be if there were many trials of a 2007 season. A large variance would mean a wider confidence interval. Less variance a narrower confidence interval around that 3.80 value.
Those K rates you are hung up on (i.e., "but to argue that Phil Hughes' 10 K/9 could equal Gooden's 14 K/9 seems ludicrous to me") are not terribly germane to this topic, because even Pecota concedes that Hughes will likely put up some pretty strong k rates as did Gooden. And of course, I could go on and on about what is the most important statistical consideration(s) when it comes to projecting pitchers but that is for another day.
What Pedro did at San Antonio should have caught your eye. Still, yes Pedro, but even moreso, the best comparable for Philip Hughes is Dwight Gooden. To compare someone with the composure and maturity that Philip Hughes has as a pitcher with Josh (I'm only gonna throw my heater in the AL) Beckett and Rick Ankiel is so off the mark I just hope someone as bright as you never brings them up again. I concede that both Beckett and Ankiel were very promising, just as Ryan Leaf was a very promising quarterback once upon a time, but like Leaf both Beckett (to a lesser extent) and Ankiel (to a much greater extent - sadly) have major issues in the area of makeup, which prevent them from reaching their awesome potential. Philip Hughes doesn't share much in common with those two, but he does share a lot in common with the young Dwight Gooden, and older pitchers like Clemens, Halladay, and the Mussina's of the world who are impeccable professionals who know their strengths and weaknesses as well as their opponents. Hughes already brings that to the table at the tender age of 20.

But you're using hindsight to look at Beckett and Ankiel's makeup. If you went back to 2000, not many people were badmouthing Ankiel's makeup: at most they were saying that he was a little wild. Baseball America loved him even more than they love Hughes. I do remember a couple of things being said about Beckett's makeup, but that makeup can go two ways: he's a bulldog, he never gives in, and also, he's an idiot, he isn't cerebral enough.

I think I misspoke re: variance. There is some list of comparables for Hughes. Some are spectacularly great, others are merely good, some are pretty bad. We don't necessarily know which of those tracks Hughes is on. You keep speaking in the absolutes about Hughes' makeup, but how do you know? Have you talked to him? Even if you have, isn't there some way it's not going to go that perfectly for him. If his makeup isn't as good as you say it is, or even if it is, and he still has growing pains, we're going to have to look at a greater list of comparables than saying: He's Dwight Gooden, no question. Given that, his projection is 3.8 or whatever. He then has some confidence band around that projection.

I would think that his 90th degree projection is extremely good, maybe even better than 2.95, and he has a 90th degree projection better than every pitcher in the league outside of Santana, Halladay, Liriano, Felix, and maybe 2 or 3 more. So his confidence band is also pretty high. But I don't think PECOTA's numbers are crazy, or even particularly wrong:

Some other projection models have him at:

ZIPS has him throwing 164 Innings, with 164 hits, 15 home runs, 58 walks, and 127 strikeouts, for a predicted ERA of 4.06, and a FIP ERA of 3.90.

First Inning predicts on the basis of AAA, and has him throwing 159 Innings, with 134 hits, 46 walks, 150 strikeouts, and 12 home runs, for an ERA of 2.86, and a FIP ERA of about 3.00.

I also think you're doing good work, as Reggievision says, but there's nothing wrong with a little bit of caution, which is what PECOTA and ZIPS are trying to do. First Inning has a list of comps making up their projection that you might not like: http://firstinning.com/players/Philip-Hughes-595/, at the bottom.

THEBOSS84
01-16-07, 01:44 PM
Can someone please let me know how accurate these PECOTA'S generally are? I would like to see some examples of what they projected in the 2006 season for some players vs. what the player actually did...

thanks

PaulieIsAwesome
01-16-07, 01:45 PM
I've got to say, Hughes 2.50 has done more than his share of homework in projecting Hughes in pinstripes. Make sure to keep all these threads archived; frankly, if Phillip Hughes even comes close to what H2.50's models indicate, I'll be the first name on the petition to get him a job in the Bronx.

If Hughes is were PECOTA says he'll be, I want him in the Bronx. The 20th or so most valuable pitcher in the AL? That would mean he's a #2 starter right now.

Oh, and I agree with Hughes2.50 about Wang. His DIPS ERA was right near 4.00, and I don't see his periphs really falling too much. PECOTA is probably going on something like Hardball Times' xFIP, which says that most pitchers with his groundball rates still give up more home runs that Wang did. Hardball Times xFIP ERA is FIP, Fielding Independent Pitching, but instead of using home runs, finds that the average major league pitcher has 11% of all flyballs turn into home runs, and that pitchers have little control over whether or not a flyball turns into a home run. It adjusts home runs to 11% of flyballs, which would bring Wang's FIP up to, you guessed it, 4.35.

I'm pretty sure Hardball Times never actually wrote up their data for why pitchers don't control home runs outside of flyballs, and I've read some things which have disputed that. I'd bet for Wang's ERA to be 3.7 or lower before I'd bet it to be 4.3 or higher.

PaulieIsAwesome
01-16-07, 01:49 PM
Can someone please let me know how accurate these PECOTA'S generally are? I would like to see some examples of what they projected in the 2006 season for some players vs. what the player actually did...

thanks

There are a bunch of older articles on Baseball Prospectus, some of which you don't need a subscription for, but I can't find one for 2006 yet:

http://www.baseballprospectus.com/article.php?articleid=2659

http://www.baseballprospectus.com/article.php?articleid=2515

On a team level: http://www.baseballprospectus.com/article.php?articleid=5609

Reggievision
01-16-07, 02:01 PM
If Hughes is were PECOTA says he'll be, I want him in the Bronx. The 20th or so most valuable pitcher in the AL? That would mean he's a #2 starter right now.

Heh. I meant getting Hughes2.50 a job in the Bronx! Damn ambiguity.

His projections for Phillip Hughes are so incredibly optimistic; he's basically saying that Hughes will be an elite HOF-caliber starter almost immediately, and that's an incredibly bold claim to make. The proof will be in the pudding, however, and I'd like nothing more than for him to be absolutely correct. We will see.

pedromartinezfan
01-16-07, 06:15 PM
Hughes and Gooden equal at age 20? If he has a 1.53 ERA and 14.2 WARP I'll become a Yankees fan.

Hughes2.50
01-16-07, 06:46 PM
First Inning predicts on the basis of AAA, and has him throwing 159 Innings, with 134 hits, 46 walks, 150 strikeouts, and 12 home runs, for an ERA of 2.86, and a FIP ERA of about 3.00.As for Triple A, I'm projecting that Hughes ERA would be 2.50 if he accumulated over a 100 innings down there this year.

Skars
01-16-07, 10:22 PM
Hughes and Gooden equal at age 20? If he has a 1.53 ERA and 14.2 WARP I'll become a Yankees fan.

Amen to that. I'll become an Eagles fan :lol:

Hughes2.50
01-16-07, 10:54 PM
Hughes and Gooden equal at age 20? If he has a 1.53 ERA and 14.2 WARP I'll become a Yankees fan.I didn't say 1.53 ERA, I said:
If you want a true comparable, check out what Dwight Gooden did before he was 21.

I don't think that Hughes will put up those numbers because he pitches in a league with a DH and it is a tougher league, and because the Yankees won't expose him to as many innings or starts. But if you want to know how good Philip Hughes is as a 20/21 year old major leaguer, look no farther than the 20 year old Dwight Gooden.<p>However at 20 and 21 Gooden did average a 175 ERA+. I can see Hughes doing the same. But to expect Hughes to get down to a 1.53 is very unlikely, even if he is better than Gooden was back then, and I doubt that is true.<p>Dwight Gooden is one of the saddest examples of what drug abuse can do to a great athletic talent.

Hughes2.50
01-16-07, 11:02 PM
Heh. I meant getting Hughes2.50 a job in the Bronx! Damn ambiguity.

His projections for Phillip Hughes are so incredibly optimistic; he's basically saying that Hughes will be an elite HOF-caliber starter almost immediately, and that's an incredibly bold claim to make. The proof will be in the pudding, however, and I'd like nothing more than for him to be absolutely correct. We will see.They may seem optimistic, and in Hughes case they are more extreme than the other analytical projections are estimating. Yet, they are what they are and I stand by them (to the extent anyone can with projections). Yet, as you said the proof is in the pudding. We will find out soon about these various projections. For example, Paulie pointed out the projections that Pecota has for Hughes, Sanchez, and Clippard. Pecota and I agree on only one of the three (Clippard), and, fortunately for an empirical test of the two projections both Sanchez and Hughes ought to get some chances to show what they can do in the majors. Maybe they will post upwards of a hundred or more innings each, and our respective model projections can be examined.

pedromartinezfan
01-17-07, 01:35 PM
Yeah, Hughes is really capable of of a 175 ERA+ right now. Do you know how many players in MLB history have ever done that in a single season? That would put him just outside the top 100 best seasons ever. Do you know how many of them were 20 or 21?

Dwight Gooden was 20.
Silver King was also 20. 119 years ago.
Jack Stivetts, one year after Silver King.
Billy Rhines in 1890.

The only pitchers over a 178 ERA+ this century are Pedro Martinez (2000, 2001, 2002, and 2003), Roger Clemens once, Randy Johnson twice, Jason Schmidt, and of course, Johan Santana. The average age there was 33 years old. Gooden has no comparables.

Hughes2.50
01-17-07, 05:54 PM
Yeah, Hughes is really capable of of a 175 ERA+ right now. Do you know how many players in MLB history have ever done that in a single season? That would put him just outside the top 100 best seasons ever. Do you know how many of them were 20 or 21?

Dwight Gooden was 20.
Silver King was also 20. 119 years ago.
Jack Stivetts, one year after Silver King.
Billy Rhines in 1890.

The only pitchers over a 178 ERA+ this century are Pedro Martinez (2000, 2001, 2002, and 2003), Roger Clemens once, Randy Johnson twice, Jason Schmidt, and of course, Johan Santana. The average age there was 33 years old. Gooden has no comparables.Look, you don't think that Hughes can do what the projections say he can - after all you're a Pedro Martinez fan, I can understand where you are coming from.<p>What happened to all of the years between 1900-1999? Why didn't you list the pitchers who had years with ERA+ over your arbitrary cutoff of 178, during those years?

pedromartinezfan
01-17-07, 07:05 PM
It has nothing to do with me being a Pedro Martinez fan. You're projections for a prospect are just unreasonable. You have Hughes having a career 175 ERA+, Betances matching Pedro at his peak. You also said the average year for Sanchez and Chamberlains would be Smoltz at his peak.

178 isn't really an arbitrary cutoff, its the 100th best ever.

PaulieIsAwesome
01-17-07, 07:07 PM
Look, you don't think that Hughes can do what the projections say he can - after all you're a Pedro Martinez fan, I can understand where you are coming from.What happened to all of the years between 1900-1999? Why didn't you list the pitchers who had years with ERA+ over your arbitrary cutoff of 178, during those years?



He did: "That would put him just outside the top 100 best seasons ever. Do you know how many of them were 20 or 21?" From 1876-2006, 100 pitchers had at least a 178 ERA+. If Hughes had a 175 ERA+, he would be just outside the top 100 seasons ever, dating back to before 1901. Since 1901, just 88 pitchers have had an ERA+ of 178. I mean, it's fine to hope that Phil can do that this year, but there's literally no evidence for it.


He chose 178 because that's close to 175 and the cutoff for top 100 seasons by Adjusted ERA+ on Baseball Reference.

Hughes2.50
01-17-07, 09:43 PM
He did: "That would put him just outside the top 100 best seasons ever. Do you know how many of them were 20 or 21?" From 1876-2006, 100 pitchers had at least a 178 ERA+. If Hughes had a 175 ERA+, he would be just outside the top 100 seasons ever, dating back to before 1901. Since 1901, just 88 pitchers have had an ERA+ of 178. I mean, it's fine to hope that Phil can do that this year, but there's literally no evidence for it.PaulieIsAwesome, the evidence is contained within even BP's confidence interval. More specifically, I am not projecting a 175 ERA+ for Hughes this year - more accurately it would be about 152.5 assuming a 2.95 ERA and a 4.50 league and park adjusted average for him. <p>Do I know how many were 21? I don't have to look that up to estimate damn few if none. What does that matter? My question to you and pedromartinezfan is how many of those pitchers have pitched as well as Hughes has before they got to the majors? How many had Hughes' stuff? How many had Hughes' makeup?<p>I agree that the projections for Hughes and the rest of the Royal Flush are difficult to swallow for some. That doesn't change what the projections are though. Since this appears to be a classic case of minds not meeting on a topic, I respectfully ask that we all wait until we actually see how the results fall before we get into back-and-forths about the probabilities of it happening, or how rare it has been historically.<p>As for pedromartinezfan, I would most certainly say that it does happen to have to do with you being a fan of pedro. You yourself said (I might say in an implied bait at that) 'that if Hughes posts a 177 ERA+ this year you would become a Yankee fan.' It would appear to me that you acknowledge that you are no fan of the Yankees. Given that, wouldn't it be more natural to say that you have trouble accepting what I am projecting the Royal Flush to do because of your own bias. Or that it is easier for you to suggest that the projections are 'just unreasonable,' because you don't want to see them actually develop? A more favorable to you interpretation would be you think it is unlikely, I have no problem with that point, as I agree that it hasn't happened often in the past. Yet, I did present the projections and we can wait to see how it turns out. <p>But just who do you think you are to call someone else unreasonable in what they present?
He chose 178 because that's close to 175 and the cutoff for top 100 seasons by Adjusted ERA+ on Baseball Reference.That's interesting, because I thought that he didn't include 175 as his cutoff for ERA+ because just 4 years ago a 22 year old pitcher finished the season 3rd in the Cy Young Voting, and had a 175 ERA+ for the year.

Hughes2.50
01-17-07, 10:03 PM
They've got Jeter with another great year: by WARP, 3rd best player in baseball behind Pujols and Mauer. A-Rod's 8th, Ortiz is first Red Sock on the list, at 18th. They've got Cano staying hot, .308/.345/.472, 27th best player in baseball. I'll run a couple of numbers to see where this generally puts the Yankees compared to the Sox.Paulie, if you could - I'd like to see the VORP # projected for '07 for the players mentioned in this paragraph. Thanks.

NJASDJDH
01-17-07, 10:41 PM
PaulieIsAwesome, the evidence is contained within even BP's confidence interval. More specifically, I am not projecting a 175 ERA+ for Hughes this year - more accurately it would be about 152.5 assuming a 2.95 ERA and a 4.50 league and park adjusted average for him. Do I know how many were 21? I don't have to look that up to estimate damn few if none. What does that matter? My question to you and pedromartinezfan is how many of those pitchers have pitched as well as Hughes has before they got to the majors? How many had Hughes' stuff? How many had Hughes' makeup?
I agree that the projections for Hughes and the rest of the Royal Flush are difficult to swallow for some. That doesn't change what the projections are though. Since this appears to be a classic case of minds not meeting on a topic, I respectfully ask that we all wait until we actually see how the results fall before we get into back-and-forths about the probabilities of it happening, or how rare it has been historically.
As for pedromartinezfan, I would most certainly say that it does happen to have to do with you being a fan of pedro. You yourself said (I might say in an implied bait at that) 'that if Hughes posts a 177 ERA+ this year you would become a Yankee fan.' It would appear to me that you acknowledge that you are no fan of the Yankees. Given that, wouldn't it be more natural to say that you have trouble accepting what I am projecting the Royal Flush to do because of your own bias. Or that it is easier for you to suggest that the projections are 'just unreasonable,' because you don't want to see them actually develop? A more favorable to you interpretation would be you think it is unlikely, I have no problem with that point, as I agree that it hasn't happened often in the past. Yet, I did present the projections and we can wait to see how it turns out.
But just who do you think you are to call someone else unreasonable in what they present?That's interesting, because I thought that he didn't include 175 as his cutoff for ERA+ because just 4 years ago a 22 year old pitcher finished the season 3rd in the Cy Young Voting, and had a 175 ERA+ for the year.



With all due respect, I'm a Yankee fan and I think your projections are outlandish at best. Not only your projections, but the tact or lack thereof you use when discussing them. It seems you have no use for those who would disagree upon reasonable grounds. If you are planning on having your statistical analysis taken seriously then you should be able to answer general "sniff test" questions without backing out of the discussion. For instance, do you really believe much faith should be placed in your projection of Dellin Betances? On what basis are you projecting Joba Chamberlain? What is the body of evidence? What would your projection system have spit out for Felix, Prior, or Beckett when they were on the cusp of the majors? What about Kerry Wood?

Skars
01-17-07, 11:27 PM
WIf you are planning on having your statistical analysis taken seriously then you should be able to answer general "sniff test" questions without backing out of the discussion.

Don't count on it. I appreciate his general knowledge of our minor leaguers, but he's never going to go there. I've exhausted that route already, so its easier to just laugh the predictions off

PaulieIsAwesome
01-17-07, 11:52 PM
Just because you're a Red Sox fan doesn't mean you can't make non-biased judgements of Yankee prospects. I think pedromartinezfan would readily admit that Hughes is one of the top 2 pitching prospects in baseball, and has a bright future if he can avoid injuries.

They've got Jeter with a 50.1 VORP. Pujols 81.4, Mauer 53.4, Ortiz 53.4, Cano 30.7, A-Rod 48.0.

Hughes2.50
01-18-07, 12:38 AM
Just because you're a Red Sox fan doesn't mean you can't make non-biased judgements of Yankee prospects. I think pedromartinezfan would readily admit that Hughes is one of the top 2 pitching prospects in baseball, and has a bright future if he can avoid injuries.

They've got Jeter with a 50.1 VORP. Pujols 81.4, Mauer 53.4, Ortiz 53.4, Cano 30.7, A-Rod 48.0.Thanks. And I agree that being a fan of another team doesn't disqualify you from making non-biased judgments of players on your own, or other teams. It takes an awareness of, and implementation of rigourous safeguards to stay non-biased. An often overlooked external bias is the one where social pressures from others can compromise a systematic, non-biased approach. The key is that one remains true to a system that treats all that is evaluatd in a rigorous, standardized way. I have no way of knowing if pedromartinez fan is trained, or understands how to remain unbiased. I do know that he had a reaction to my projections which betrayed a bias against the projections because they seemed very, very, unlikely or possibly absurd to him/her. That reaction is the type of reaction one must guard oneself from following if one wants to be truly non-biased.

Hughes2.50
01-18-07, 12:46 AM
Don't count on it. I appreciate his general knowledge of our minor leaguers, but he's never going to go there. I've exhausted that route already, so its easier to just laugh the predictions offSkars, you are right about me not going there. I will go there on my own schedule and on my terms, and then everyone can open up the black-box and dissect it. That will happen before the season starts and will allow people to examine the actual projections that come from the system. Still, I thought that I had conceded to you in our last discussion that the Betances and Chamberlain projections carry a very large confidence interval. Indeed on the original thread where I posted those projections for the Royal Flush I was asked by someone, who had the highest bust potential and I noted that it was Sanchez and Betances. So obviously, much still needs to happen for the projections to come to fruition, as I projected them to be.<p>As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them. I'll explain more to you later, before the season starts. Remind me to post all of that before the season starts. But kindly wait until at least the last week of March to remind me.

PaulieIsAwesome
01-18-07, 03:46 AM
Skars, you are right about me not going there. I will go there on my own schedule and on my terms, and then everyone can open up the black-box and dissect it. That will happen before the season starts and will allow people to examine the actual projections that come from the system. Still, I thought that I had conceded to you in our last discussion that the Betances and Chamberlain projections carry a very large confidence interval. Indeed on the original thread where I posted those projections for the Royal Flush I was asked by someone, who had the highest bust potential and I noted that it was Sanchez and Betances. So obviously, much still needs to happen for the projections to come to fruition, as I projected them to be.As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them. I'll explain more to you later, before the season starts. Remind me to post all of that before the season starts. But kindly wait until at least the last week of March to remind me.

I have a lot of speculation about the value of your projections given some of your posts here, but I'm more than willing to keep an open mind. I think it's pretty cool that you've made your own projection scheme, and the numbers should be extremely interesting.



However, I don't think it'd be very productive or even particular rational to accuse pedromartinezfan or any non-Yankee fan poster who disagrees with you of bias. You want to better your projection model? Listen to all the criticism you can possibly get, and try and figure out where the disagreement comes from.

pedromartinezfan
01-18-07, 07:40 AM
I'm not being biased at all. You're projections have 4 minor leaguers and Wang in the Majors for many years posing as Pedro, Clemens, Smoltz, another Smoltz, and Pete Alexander.

Hughes2.50
01-18-07, 09:39 AM
I have a lot of speculation about the value of your projections given some of your posts here, but I'm more than willing to keep an open mind. I think it's pretty cool that you've made your own projection scheme, and the numbers should be extremely interesting.



However, I don't think it'd be very productive or even particular rational to accuse pedromartinezfan or any non-Yankee fan poster who disagrees with you of bias. You want to better your projection model? Listen to all the criticism you can possibly get, and try and figure out where the disagreement comes from.I am not being contentious when I say that he is biased. We are all biased, that should be known as a given. The person you know almost certainly hasn't controlled their own bias, is the person who doesn't acknowledge that they are biased.<p>To be biased means that you have a disposition to respond/think in a particular way to a situation, information, etc. We all have that disposition once we've experienced anything.<p>Now, when being confronted with new information, or old for that matter, if we want to remain open we need to acknowledge that we have particular tendencies and guard against them prejudicing our perceptions. As a researcher one needs to guard against tainting a systematic and rigorous analysis with bias. And as a consumer of information one needs to guard against bias from closing off some possibilities.<p>So, I think that my discussion of bias needs to be put in the context it is intended, that is to acknowledge that we are all biased, and that both the researcher and the consumer of research both have biases and one needs to be aware of that and do the right thing by planning to minimize the affects of that bias. My perspective is that of a trained researcher, hopefully my explanation of where I am coming from clears up some difficulties. <p>Given the response to some of the projections here, I think it is probably best now to follow up in the manner I told Skars I would do, and not provide any new (i.e., not discussed here yet) projections until then.

Jasbro
01-18-07, 03:05 PM
I love it.

Create a new projection formula, present your conclusions, insist your model is better than other methods, attempt to end the discussion about the validity of your method when it is questioned, and then accuse dissenters of falling victim to their own biases while positioning yourself as being too "trained" to exhibit any bias of your own.

For what it is worth, arrogance itself can be a form of bias.

ppa79
01-18-07, 03:46 PM
I'm not being biased at all. You're projections have 4 minor leaguers and Wang in the Majors for many years posing as Pedro, Clemens, Smoltz, another Smoltz, and Pete Alexander.

Good enough for me. :D

ma-gaga
01-18-07, 05:33 PM
I just put together a quick and dirty "projected vs actual" from last year. Just to show where they did good and bad. I highlighted a couple of players who got hurt, or didn't get the same amount of play as expected, so their numbers are a little goofy.

http://www.tc.umn.edu/~dunca016/2006%20Yankees%20-%20PECOTA%20Projections%20vs%20Actual.jpg

They were pretty far off with Wang. But I think the entire AL pitched a lot better than anyone could have expected.

Hughes2.50
01-18-07, 05:34 PM
I love it.

Create a new projection formula, present your conclusions, insist your model is better than other methods, attempt to end the discussion about the validity of your method when it is questioned, and then accuse dissenters of falling victim to their own biases while positioning yourself as being too "trained" to exhibit any bias of your own.

For what it is worth, arrogance itself can be a form of bias.Did you even read the quote below which was posted a few posts above yours? If you had I see no reason for your tone or your comments. In awhile everyone can look at the specifics of how the projections are made and what they are, and you will even be able to do it at home with someone who can read directions properly. Sheesh.<p>
Skars, you are right about me not going there. I will go there on my own schedule and on my terms, and then everyone can open up the black-box and dissect it. That will happen before the season starts and will allow people to examine the actual projections that come from the system. Still, I thought that I had conceded to you in our last discussion that the Betances and Chamberlain projections carry a very large confidence interval. Indeed on the original thread where I posted those projections for the Royal Flush I was asked by someone, who had the highest bust potential and I noted that it was Sanchez and Betances. So obviously, much still needs to happen for the projections to come to fruition, as I projected them to be.<p>As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them. I'll explain more to you later, before the season starts. Remind me to post all of that before the season starts. But kindly wait until at least the last week of March to remind me.

Jasbro
01-18-07, 08:33 PM
Did you even read the quote below which was posted a few posts above yours? If you had I see no reason for your tone or your comments. In awhile everyone can look at the specifics of how the projections are made and what they are, and you will even be able to do it at home with someone who can read directions properly. Sheesh.

Condescending much? I promise you that there are dozens of people on this site with as much "training" as you say you have, and more. And most are able to communicate their message and deal with the feedback they receive without the pedantic tone.

I am not surprised you see no reason for my tone or my comments, you are seemingly unable to allow contrary opinions or legitimate queries to pierce your shield of arrogance.

And now that you mention reading comprehension, read my post again. Nowhere do I criticize you because you haven't released the magical, other-worldly secrets of your methodology -- instead, I refer to the manner in which you have been dealing with those who find your conclusions a bit exuberant.

Next time, I can break it down into shorter sentences and use a larger font, if that will help you. Just let me know.

If you want to be taken seriously, you need to learn how to take seriously the opinions, findings and questions of others.

destiNY
01-18-07, 10:04 PM
We need to let in some cool NYC wind in this thread

shroud
01-18-07, 10:06 PM
Again, not worried with what PECOTA or any projection system says about Wang, he's one of a kind. Good to see Cano with another solid year, another reason why it's not wise to project ML performance with MiLB numbers.

How dare you imply that Hughes2.50 might be wrong?? shame on you...

PaulieIsAwesome
01-19-07, 12:37 AM
Some number crunching:

If the lineup is going to be Damon, Jeter, Abreu, Giambi, A-Rod, Matsui, Posada, Cano, Mientkiewicz, based on the number of plate appearances the Yankees got as a team last year, if no injuries or sitting out occurs, the Yankees will score 944 runs according to PECOTA. If you replaced Mientkiewicz with Andy Phillips (who has nearly the exactly the same projection as Craig Wilson does,) the Yankees would be on pace to score 951 runs. Since PECOTA predicts that Mienkiewicz will be just 1 run better defensively than Wilson or Phillips, they would strongly suggest that course of action. Josh Phelps is even better, and not that much worse defensively, leading to the Yankees scoring 958 runs.

After running some numbers, we can see that since Posada won't play every day, at the least, that costs the Yankees 16 runs, assuming that our backup catchers hit like they did last year, which was pretty bad: .189/.252/.288

PECOTA predicts that Matsui will only get 420 plate appearances. Melky will replace the other 270 so (of Matsui's, he'll get some from Abreu and maybe Damon too,) which loses us another 7 runs.

PECOTA predicts that A-Rod, Jeter, and Cano will not play about 225 plate appearances, and will replace that production with the Nick Green/Miguel Cairo crowd. That loses another 14 runs.

Melky and Kevin Thompson will have to replace another 330 PAs of Abreu's and Damon's, according to PECOTA (this is probably how Melky gets 570 total plate appearances.) That loses 7 runs.

Giambi will need 225 plate appearances replaced. Replacing those with Phelps/Phillips costs the Yankees another 9 runs.

I think that might be it. If anyone can think of any other likely lineup changes, I'll make the changes. Using 951 as my preliminary number, (the average of the Mientkiewicz, Phelps, and Phillips platoon at first base, ) the Yankees, by PECOTA, are predicted to score 951-16-7-14-7-9 runs = 898 runs.

Oh crap, I forgot about interleague games. That'll reduce that number by a little: since Yankee pitchers only got 22 at bats last year, 24 plate appearances, it's only about 3 runs. So, 895 runs. They scored 930 last year, and 886 runs in 2005. PECOTA is always a little more conservative on a team like the Yankees, who had a couple of offensive players really break out last year (Cano, most obviously.) Since PECOTA thinks Jeter will regress some, Matsui get hurt or at least have to be babied some, Cano falls some, I think this is a pretty solid prediction.

I'll do pitching some other time. But, just very quickly, if the Yankees score 895 runs, and they prevent runs at league average (last year,) their pythagorean expectation is 89 wins. So, they're +8 wins in run scoring. If they can be at least +6 wins in pitching/defense, this could be another 95 win season (which the Yankees have done every year since 2000, and in the 10 seasons since 1997, they've done every year outside of 2000.)

Hughes2.50
01-19-07, 01:00 AM
Some number crunching:

If the lineup is going to be Damon, Jeter, Abreu, Giambi, A-Rod, Matsui, Posada, Cano, Mientkiewicz, based on the number of plate appearances the Yankees got as a team last year, if no injuries or sitting out occurs, the Yankees will score 944 runs according to PECOTA. If you replaced Mientkiewicz with Andy Phillips (who has nearly the exactly the same projection as Craig Wilson does,) the Yankees would be on pace to score 951 runs. Since PECOTA predicts that Mienkiewicz will be just 1 run better defensively than Wilson or Phillips, they would strongly suggest that course of action. Josh Phelps is even better, and not that much worse defensively, leading to the Yankees scoring 958 runs.

After running some numbers, we can see that since Posada won't play every day, at the least, that costs the Yankees 16 runs, assuming that our backup catchers hit like they did last year, which was pretty bad: .189/.252/.288

PECOTA predicts that Matsui will only get 420 plate appearances. Melky will replace the other 270 so (of Matsui's, he'll get some from Abreu and maybe Damon too,) which loses us another 7 runs.

PECOTA predicts that A-Rod, Jeter, and Cano will not play about 225 plate appearances, and will replace that production with the Nick Green/Miguel Cairo crowd. That loses another 14 runs.

Melky and Kevin Thompson will have to replace another 330 PAs of Abreu's and Damon's, according to PECOTA (this is probably how Melky gets 570 total plate appearances.) That loses 7 runs.

Giambi will need 225 plate appearances replaced. Replacing those with Phelps/Phillips costs the Yankees another 9 runs.

I think that might be it. If anyone can think of any other likely lineup changes, I'll make the changes. Using 951 as my preliminary number, (the average of the Mientkiewicz, Phelps, and Phillips platoon at first base, ) the Yankees, by PECOTA, are predicted to score 951-16-7-14-7-9 runs = 898 runs.

Oh crap, I forgot about interleague games. That'll reduce that number by a little: since Yankee pitchers only got 22 at bats last year, 24 plate appearances, it's only about 3 runs. So, 895 runs. They scored 930 last year, and 886 runs in 2005. PECOTA is always a little more conservative on a team like the Yankees, who had a couple of offensive players really break out last year (Cano, most obviously.) Since PECOTA thinks Jeter will regress some, Matsui get hurt or at least have to be babied some, Cano falls some, I think this is a pretty solid prediction.

I'll do pitching some other time. But, just very quickly, if the Yankees score 895 runs, and they prevent runs at league average (last year,) their pythagorean expectation is 89 wins. So, they're +8 wins in run scoring. If they can be at least +6 wins in pitching/defense, this could be another 95 win season (which the Yankees have done every year since 2000, and in the 10 seasons since 1997, they've done every year outside of 2000.)Thanks, nice job.

PaulieIsAwesome
01-20-07, 01:50 PM
When Nate Silver ran the PECOTA numbers, here's what he got for the Yanks/Sox:

http://www.baseballprospectus.com/unfiltered/?p=160

If you can't read it, that's 918 RS for the Yankees, 774 RA, 913 RS for the Sox, 772 RA. That puts the Yankees at 93.544 wins, the Sox at 93.336 wins. Of course, PECOTA loves JD Drew, which is the only way the Sox score that many runs.

Prickly Pete
01-20-07, 03:04 PM
When Nate Silver ran the PECOTA numbers, here's what he got for the Yanks/Sox:

http://www.baseballprospectus.com/unfiltered/?p=160

If you can't read it, that's 918 RS for the Yankees, 774 RA, 913 RS for the Sox, 772 RA. That puts the Yankees at 93.544 wins, the Sox at 93.336 wins. Of course, PECOTA loves JD Drew, which is the only way the Sox score that many runs.
I was a little surprised to see that projection for Boston's offense, too. As for Drew, PECOTA loves his rate stats (projected .868 OPS and.303 EqA) but they don't love his ability to stay on the field -- they only have him playing 115 games and putting up a VORP of 19.6, so his overall contribution isn't that huge. PECOTA is predicting pretty big bounce-back years from Crisp and Varitek, and a boost in the middle infield from Pedroia/Lugo replacing Loretta/Gonzalez.

JamieMadrox
01-20-07, 03:10 PM
I just put together a quick and dirty "projected vs actual" from last year. Just to show where they did good and bad. I highlighted a couple of players who got hurt, or didn't get the same amount of play as expected, so their numbers are a little goofy.

http://www.tc.umn.edu/%7Edunca016/2006%20Yankees%20-%20PECOTA%20Projections%20vs%20Actual.jpg

They were pretty far off with Wang. But I think the entire AL pitched a lot better than anyone could have expected.

wow, with the exception of arod, our offense really overperformed

Cuban Connection
01-21-07, 08:13 PM
wow, with the exception of arod, our offense really overperformed
It depends on your point of view.

You could also say the system really underestitmated. PECOTA was way off on the pitching staff as well.

KLJ
01-24-07, 10:21 AM
I just put together a quick and dirty "projected vs actual" from last year.
what's up with the projected rivera numbers? when was the last time he put that kind of era up? is this strictly do to age? can i assume that he will be projected even higher this year?

PaulieIsAwesome
01-24-07, 12:32 PM
what's up with the projected rivera numbers? when was the last time he put that kind of era up? is this strictly do to age? can i assume that he will be projected even higher this year?

Which ERA? Last years or this years? PECOTA never really likes Mo that much because he always outperforms his Fielding Independent Pitching stats by a lot: his DIPS ERA is normally closer to 3 than 2, and yet he keeps putting up 2.00 ERA years.

THEBOSS84
01-25-07, 10:55 AM
Can someone let me know Betance's comparables from PECOTA? Based on Nate Silver's chat answer below, I can't tell who they are going to be...


G-MOTA (Bumpus, MA): Dellin Betances's comparables!?!?! Yowza!

Nate Silver: Betances is another guy that falls below the 100 PA/BFP threshold, so that projection needs an asterisk. Still, to come like that out of a New York City high school and dominate professional hitters at age 18 is pretty impressive.

pedromartinezfan
01-25-07, 12:32 PM
In order

Matt Cain 2004
Francisco Liriano 2003
Joel Zumaya 2004
Francisco Rodriguez 2001
Travis Blackley 2002
Yusmeiro Petit 2004
Jerome Williams 2001
Justin Jones 2004
Alan Webb 1999
Joel Pineiro 1998
Angel Rocha 2004
Ruben Quevedo 1998
Craig Anderson 2000
Julio Caesar Pimentel 2005
Gonzalo Lopez 2003
Ricardo Aramboles 2001
Paul Stewart 1998
Jacob McGee 2005
Geronimo Mendoza 1999
Fabio Castro 2004

Hughes2.50
01-29-07, 07:12 PM
This was all ERA. Hughes' PERA was 3.80.


Clippard had 137 IP, 5.14 ERA, 5.03 PERA.

Sanchez had 122 IP, 5.16 ERA, 5.04 PERA.



Hughes2.50, do you really think that Hughes will have an ERA lower than 3.91, or 3.80? Jake Peavy took 2 years to get really good on the big league level, Brad Penny took a couple of years: they're two of his top comparables.I'm curious what your response to a question is, if you read this. Take a look at <a href="http://www.baseballprospectus.com/pecota/HUGHES19860624A.php">Hughes Pecota</a> from 2005 and 2006. You can see whether you look at his PERA or ERA he has never been over 3.91 or 3.80 for that matter. Given expected natural maturation and improvement, what justification do you think BP would/does give for having Hughes' projections for '07 over the parameter estimates their own calculations indicate? <p>Aside from the hackneyed rationalizations (Hughes is only 21, Hughes has never pitched over AA ball; no one posts ERA's like that in their first year; etc.) that one would expect from those not committed to the Pecota models how do you suppose the people over at BP would explain why their MLE models consistently show Hughes better than 3.91 or even 3.80 ERA yet, their projection for 2007 shows what it does? Thanks

PaulieIsAwesome
01-30-07, 03:26 PM
I'm curious what your response to a question is, if you read this. Take a look at Hughes Pecota (http://www.baseballprospectus.com/pecota/HUGHES19860624A.php) from 2005 and 2006. You can see whether you look at his PERA or ERA he has never been over 3.91 or 3.80 for that matter. Given expected natural maturation and improvement, what justification do you think BP would/does give for having Hughes' projections for '07 over the parameter estimates their own calculations indicate? Aside from the hackneyed rationalizations (Hughes is only 21, Hughes has never pitched over AA ball; no one posts ERA's like that in their first year; etc.) that one would expect from those not committed to the Pecota models how do you suppose the people over at BP would explain why their MLE models consistently show Hughes better than 3.91 or even 3.80 ERA yet, their projection for 2007 shows what it does? Thanks

Because those hackneyed rationalizations actually do matter. PECOTA does nearly every other projections schemes one better, by finding not just statistical comparables, but height and body comparables too. You can write this off, as you did earlier, and say that the only comparable that really matters is Doc Gooden, but given, as you say, no ones posts ERAs like this in their first year, they have to regress their MLE projections back.

Hughes2.50
01-30-07, 06:25 PM
Because those hackneyed rationalizations actually do matter. PECOTA does nearly every other projections schemes one better, by finding not just statistical comparables, but height and body comparables too. You can write this off, as you did earlier, and say that the only comparable that really matters is Doc Gooden, but given, as you say, no ones posts ERAs like this in their first year, they have to regress their MLE projections back.So what do P. Martinez 5-11 170; Randy Johnson 6-10 225; and R. Clemens 6-4 220, have in common? Not much except for the fact that they can throw harder than most mortals, do funny things to the spin of a baseball, control the destination of a thrown baseball, and, will be first ballot HOF awardees.<p>Here is what Baseball Prospectus says about their Peripheral Earned Run Average (PERA) statistic:<p>
PERA is a pitcher's ERA as estimated from his peripheral statistics (EqH9, EqHR9, EqBB9, EqK9). Because it is not sensitive to the timing of batting events, PERA is less subject to luck than ERA, and is a better predictor of ERA going-forward than ERA itself. Like the rest of a pitcher's equivalent stats, his PERA is calibrated to an ideal league with an average PERA of 4.50.<p>Since those statistics are already <b>major league adjusted</b> when they put them up for prospects it is useful to see what they find using their model for Hughes in 2005 and 2006.<p>In 2006 they find that Hughes has a PERA of 2.58 in 30.67 innings atA+ Tampa and 3.39 at AA Trenton. Calculating by weighting by the innings pitched I get a <b>cumulative MLE adjusted PERA of 3.22 for Hughes in '06</b> over his 146+ innings in the minors last year- according to Baseball Prospectus' own figures.<p>For 2005 I get a MLE PERA of 3.32 at A- Charleston, and 3.56 at A+ Tampa. Again, using BP's own numbers calculating a weighted by innings total you get a cumulative year result for 2005, of <b>MLE PERA =3.37for Hughes in '05</b>.<p>So, by BP"s own standards, Hughes' equivalent (MLE) PERA results from '05 and '06 ought to be a "better predictor of ERA going-forward than ERA itself."<p>So, when you use the phrase "they have to regress their MLE projections back," I agree. They have to regress their projections back to statistical norms and use the model fits not some hackneyed rationalizations - for why the results should be different than the models indicate.<b>BTW, PECOTA's results for their PERA cumulatively for 2006 is only 0.26 different than my model. And the reason for that is because my model calculates peripheral statistics differently than theirs does.</b>

PaulieIsAwesome
01-30-07, 08:27 PM
So what do P. Martinez 5-11 170; Randy Johnson 6-10 225; and R. Clemens 6-4 220, have in common? Not much except for the fact that they can throw harder than most mortals, do funny things to the spin of a baseball, control the destination of a thrown baseball, and, will be first ballot HOF awardees.

So, these guys should be the full list of comparables for Philip Hughes?


Here is what Baseball Prospectus says about their Peripheral Earned Run Average (PERA) statistic:
Since those statistics are already major league adjusted when they put them up for prospects it is useful to see what they find using their model for Hughes in 2005 and 2006.
In 2006 they find that Hughes has a PERA of 2.58 in 30.67 innings atA+ Tampa and 3.39 at AA Trenton. Calculating by weighting by the innings pitched I get a cumulative MLE adjusted PERA of 3.22 for Hughes in '06 over his 146+ innings in the minors last year- according to Baseball Prospectus' own figures.
For 2005 I get a MLE PERA of 3.32 at A- Charleston, and 3.56 at A+ Tampa. Again, using BP's own numbers calculating a weighted by innings total you get a cumulative year result for 2005, of MLE PERA =3.37for Hughes in '05.
So, by BP"s own standards, Hughes' equivalent (MLE) PERA results from '05 and '06 ought to be a "better predictor of ERA going-forward than ERA itself."
So, when you use the phrase "they have to regress their MLE projections back," I agree. They have to regress their projections back to statistical norms and use the model fits not some hackneyed rationalizations - for why the results should be different than the models indicate.BTW, PECOTA's results for their PERA cumulatively for 2006 is only 0.26 different than my model. And the reason for that is because my model calculates peripheral statistics differently than theirs does.




When you're willing to show your model, I'll be glad to compare it to PECOTA. Since you haven't given your list of comps yet, I can't compare those to PECOTA's.

Hughes2.50
01-30-07, 08:56 PM
So, these guys should be the full list of comparables for Philip Hughes?Where did I say those guys were comparables for Hughes? I don't even care for comparables, particularly when they take people away from empirical evidence. My point was that it is relatively useless to use comparables (after all can't P. Martinez, Johnson, and Clemens be argued to provide an excellent comparable for each other - in hindsight? And, because we can see that hard throwing, high ERA+ pitchers can come in drastically different physical packages, don't we need to focus on results and not at the expense of body type. This is particularly so when we can see that the pitcher has the right stuff and isn't producing results on smoke and mirrors). Moreso, when adjustments are made away from the estimates of an empirical model to satisfy height, weight, whatever constraints. <p>Using the three pitchers above, all strikeout aces, all high ERA+ career pitchers screams for their comparables if your suggestions about what motivates Pecota are to be taken at face value. For example, is Pedro Martinez a comparable of Clemens, Johnson, or Hughes for that matter? I think if they pitched in the same league/level at the same age (<a href="http://forums.nyyfans.com/showpost.php?p=3869041&postcount=1141"><b>check out this link for an example of a way to compare players based on age/level as the primary components</b><a>, and produced similar MLE DIPS ERA's or PERA"s that you can make a hell of a lot more of a case that they are comparables to each other, than to say - well wait a minute, Hughes is 6-5 220 lets match him up with Brett Myers because he is more similar physiologically, or with Brad Penny - even though Myers nor Penny ever came close to producing the same MLE PERAs at the same age.<p>

When you're willing to show your model, I'll be glad to compare it to PECOTA. Since you haven't given your list of comps yet, I can't compare those to PECOTA's.Don't expect a particularly long list for a player with few peers when it comes to doing what he did at his age. And, as I said earlier, I will show the model and whatever constraints I use to include estimates before the season. I see little value in providing comparables beyond their use as an illustrative tool so that people can see a potential career trajectory. However, given a players age, current performance level (whether CR+ or MLE DIPS ERA+) one can estimate peak performances and probable career averages given certain constraints (e.g., years played and innings pitched or plate appearances made).<p>BTW, your lack of response to my criticisms of your interpretation of why PECOTA's results from '05 and '06 are so much different than the projections for Hughes in '07 are in no way connected to whether or not I produce a model for projecting prospects (which I intend to do before the season starts). I don't believe that you have provided a statistically valid defense for what BP has done to Hughes' projections (would you refuse to discuss this with another poster who did't say they had a model?). If you don't want to defend what BP apparently did with the Hughes' '07 projections anymore, I certainly wouldn't blame you.

PaulieIsAwesome
01-30-07, 10:46 PM
Where did I say those guys were comparables for Hughes? I don't even care for comparables, particularly when they take people away from empirical evidence. My point was that it is relatively useless to use comparables (after all can't P. Martinez, Johnson, and Clemens be argued to provide an excellent comparable for each other - in hindsight? And, because we can see that hard throwing, high ERA+ pitchers can come in drastically different physical packages, don't we need to focus on results and not at the expense of body type. This is particularly so when we can see that the pitcher has the right stuff and isn't producing results on smoke and mirrors). Moreso, when adjustments are made away from the estimates of an empirical model to satisfy height, weight, whatever constraints. Using the three pitchers above, all strikeout aces, all high ERA+ career pitchers screams for their comparables if your suggestions about what motivates Pecota are to be taken at face value. For example, is Pedro Martinez a comparable of Clemens, Johnson, or Hughes for that matter? I think if they pitched in the same league/level at the same age (check out this link for an example of a way to compare players based on age/level as the primary components (http://forums.nyyfans.com/showpost.php?p=3869041&postcount=1141), and produced similar MLE DIPS ERA's or PERA"s that you can make a hell of a lot more of a case that they are comparables to each other, than to say - well wait a minute, Hughes is 6-5 220 lets match him up with Brett Myers because he is more similar physiologically, or with Brad Penny - even though Myers nor Penny ever came close to producing the same MLE PERAs at the same age. (http://forums.nyyfans.com/)

Don't expect a particularly long list for a player with few peers when it comes to doing what he did at his age. And, as I said earlier, I will show the model and whatever constraints I use to include estimates before the season. I see little value in providing comparables beyond their use as an illustrative tool so that people can see a potential career trajectory. However, given a players age, current performance level (whether CR+ or MLE DIPS ERA+) one can estimate peak performances and probable career averages given certain constraints (e.g., years played and innings pitched or plate appearances made). (http://forums.nyyfans.com/)
BTW, your lack of response to my criticisms of your interpretation of why PECOTA's results from '05 and '06 are so much different than the projections for Hughes in '07 are in no way connected to whether or not I produce a model for projecting prospects (which I intend to do before the season starts). I don't believe that you have provided a statistically valid defense for what BP has done to Hughes' projections (would you refuse to discuss this with another poster who did't say they had a model?). If you don't want to defend what BP apparently did with the Hughes' '07 projections anymore, I certainly wouldn't blame you. (http://forums.nyyfans.com/)


PECOTA's a black box, and I can't defend it any serious way. However, I'm not so quick to reject it like you have.

Yes, your tone is certainly one of the reasons that I've disagreed with you as sharply. Sorry, it's true. I wasn't here when you joined, but having read the responses by some people on this thread, it's clear that you generally don't take criticism that well. Your willingness to make declarative statements of opinion as absolute fact certainly doesn't help my opinion of your postings.

Hughes2.50
01-30-07, 11:08 PM
PECOTA's a black box, and I can't defend it any serious way. However, I'm not so quick to reject it like you have.I don't reject it at all. If you think about it, I respect their black-box more than they do. I am defending the results of their PECOTA results (although mine differs slightly from theirs), and wonder why they abandon the results for some undescribed reasons. Actually the calculations are pretty straightforward and I will describe mine before the season starts and you will be able to take any player, any year, and compare the respective results. Pecota is a good system when you look at the consistency of the mathematical results, when you have a systematic way to compare it.


Yes, your tone is certainly one of the reasons that I've disagreed with you as sharply. Sorry, it's true. I wasn't here when you joined, but having read the responses by some people on this thread, it's clear that you generally don't take criticism that well. Your willingness to make declarative statements of opinion as absolute fact certainly doesn't help my opinion of your postings.If you look back at this thread, and think back to our discussion on the Stat Help thread you will notice that two posters came to this thread (neither have engaged me in any discussion in the past, at least to my memory) criticized me for being unwilling to defend myself (that is absurd if you take the time to look over my posts - not responding and saying I will respond later is not the same as someone who refuses to discuss, or won't even set a time when something will be discussed further) and claim I am pedantic (guilty as charged) and don't take criticism well. Well with respect to taking criticism well, I take it very well indeed, but don't agree just for agreements sake. I will counter what I think are false, or baseless claims. That can obviously lead to people feeling some discomfirt. Sorry if you felt it was personal, it isn't.

Saxmania
01-31-07, 06:32 AM
As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them.

I'm looking forward to seeing Hughes' methodology in March, but frankly for me this is the key sentence.

Be seeing you,

Saxmania

Hughes2.50
01-31-07, 09:53 AM
I'm looking forward to seeing Hughes' methodology in March, but frankly for me this is the key sentence.

"As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them."

Be seeing you,

SaxmaniaIn context, I was referring to the Betances and Chamberlain projections as they are not based on A+ or better results to work with. Baseball Prospectus made a similar comment about their report on Betances (and due to small sample size issues).

Saxmania
01-31-07, 11:45 AM
In context, I was referring to the Betances and Chamberlain projections as they are not based on A+ or better results to work with. Baseball Prospectus made a similar comment about their report on Betances (and due to small sample size issues).

Then I will continue to wait until March; I think I've made my feelings known on this subject.

Be seeing you,

Saxmania

Hughes2.50
02-02-07, 11:56 PM
Ok, but if Hughes puts up a 2.95 ERA in 130 IP in the bigs, that would make him one of the top 10 pitchers in baseball according to PECOTA. Could you run your MLEs on Justin Verlander's 2005? In 2005, in 86 A+ innings, Verlander had 10.9 K/9, 1.99 BB/9, .314 HR/9, with a 1.67 ERA, and in 32.7 AA IP, 8.8 K/9, 1.92 BB/9, .276 HR/9, for a .28 ERA.

Hughes last year threw 30 A+ Innings, 9 K/9, .6 BB/9, and 0 HR/9, with a 1.80 ERA, and 116 innings, 10.7 K/9, 2.48 BB/9, .388 HR/9 with a 2.25 ERA. I think that puts them pretty close. Maybe I'd give Hughes a little advantage, from throwing more innings, at AA. But I can't think that Hughes will do much better than Verlander did last year, and expect him to have the same problem Verlander had, moving from 145 innings to 170-180 innings total, and to struggle some at the end of the year. I'd be really surprised for Hughes to have an ERA below 3.4 with more than 130 IP. Just 4 AL pitchers did that last year. He's going to have some growing pains, like Liriano (injuries) and Felix (unluckiness, mainly: his defense independent stats were some of the best in baseball.)Now that Pecota has released their cards we can look at <a href="http://www.baseballprospectus.com/pecota/verlaju01.php"><b>Verlander's minor league results from '05</b></a> and see how well it projects for Verlander's 2006 season. You can see by that link that Verlander's A+ performance is equal to a PERA of 3.71 and his AA performanceis 3.26. Weighting for innings as I did with Hughes you get a cumulative minor league PERA in '05 of 3.59. What was his actual ERA in '06? 3.63. So the PERA's from '05 predicted the 24 year old Verlander's '06 pretty well. You would expect that because as BP says in their literature, PERA predicts future ERA better than ERA does, but of course this could merely be a coincidence. However, as a young pitcher approaches 26 years old, he rarely makes a large leap forward in performance over what he has accomplished up to then. There is some inflation of performance but it isn't as much as you'd expect from a 20/21 year old like Hughes. <p>In addition with Verlander he was shutdown for a time with tired arm after his minor league season was over, so his '05 ML results ought not be a surprise. Indeed, Verlander was hitting the wall in '06 around the dog days of summer and the Tigers saw Verlander's ERA rise from 2.60s, all the way to the 3.60s at the end of the season. For that reason, I believe that there is some hope for Verlander to post a sub 3.00 ERA for a whole season in the majors, once he gets completely acclimated to the workload. Interestingly, Hughes' showed in '06 better long-term pitching endurance without any tired arm, or soreness issues of note, and was actually his most dominant in his last game, a six inning stint in the playoffs when he struck out 13 batters. This all bodes well for Hughes to pitch without issue up to his prescribed limit of 180 innings (as mentioned by VP of Minor League Operations Mark Newman earlier this week).<p> If you check you will see that Verlander got shutdown in the minors after 118.66 innings in the minors in '05, and then hit the wall with tired arm in '06 after about 130 innings. Since Hughes had no such issues in 152 total innings in '06, he ought to be able to handle the step-up to 180 with no problem. If he had hit the wall at 150 innings then we would expect him to possibly hit the wall a bit later, say at about 165-170, but that is unlikely to be an issue for Hughes in '06 because of the well designed path the Yankees have articulated for Hughes.

PaulieIsAwesome
02-03-07, 10:50 AM
Now that Pecota has released their cards we can look at Verlander's minor league results from '05 (http://www.baseballprospectus.com/pecota/verlaju01.php) and see how well it projects for Verlander's 2006 season. You can see by that link that Verlander's A+ performance is equal to a PERA of 3.71 and his AA performanceis 3.26. Weighting for innings as I did with Hughes you get a cumulative minor league PERA in '05 of 3.59. What was his actual ERA in '06? 3.63. So the PERA's from '05 predicted the 24 year old Verlander's '06 pretty well. You would expect that because as BP says in their literature, PERA predicts future ERA better than ERA does, but of course this could merely be a coincidence. However, as a young pitcher approaches 26 years old, he rarely makes a large leap forward in performance over what he has accomplished up to then. There is some inflation of performance but it isn't as much as you'd expect from a 20/21 year old like Hughes. In addition with Verlander he was shutdown for a time with tired arm after his minor league season was over, so his '05 ML results ought not be a surprise. Indeed, Verlander was hitting the wall in '06 around the dog days of summer and the Tigers saw Verlander's ERA rise from 2.60s, all the way to the 3.60s at the end of the season. For that reason, I believe that there is some hope for Verlander to post a sub 3.00 ERA for a whole season in the majors, once he gets completely acclimated to the workload. Interestingly, Hughes' showed in '06 better long-term pitching endurance without any tired arm, or soreness issues of note, and was actually his most dominant in his last game, a six inning stint in the playoffs when he struck out 13 batters. This all bodes well for Hughes to pitch without issue up to his prescribed limit of 180 innings (as mentioned by VP of Minor League Operations Mark Newman earlier this week).
If you check you will see that Verlander got shutdown in the minors after 118.66 innings in the minors in '05, and then hit the wall with tired arm in '06 after about 130 innings. Since Hughes had no such issues in 152 total innings in '06, he ought to be able to handle the step-up to 180 with no problem. If he had hit the wall at 150 innings then we would expect him to possibly hit the wall a bit later, say at about 165-170, but that is unlikely to be an issue for Hughes in '06 because of the well designed path the Yankees have articulated for Hughes.

Why didn't you look at Verlander's 2006 PERA, which was 4.27, rather than his ERA, which was dependent on his home ball park, which generally favors pitchers, and factors out some of the defensive help Verlander got from the great Detroit defense, or his 2006 DIPS ERA, where his FIP was 4.44, and his ESPN DIPS2.0 ERA which was 4.37? Aren't those numbers evidence that had Verlander thrown in a park neutral, defense and luck neutral environment, his ERA would have risen from his 3.67 2005 PERA about 6/10 of a run?

Hughes2.50
02-03-07, 04:12 PM
Why didn't you look at Verlander's 2006 PERA, which was 4.27, rather than his ERA, which was dependent on his home ball park, which generally favors pitchers, and factors out some of the defensive help Verlander got from the great Detroit defense, or his 2006 DIPS ERA, where his FIP was 4.44, and his ESPN DIPS2.0 ERA which was 4.37? Aren't those numbers evidence that had Verlander thrown in a park neutral, defense and luck neutral environment, his ERA would have risen from his 3.67 2005 PERA about 6/10 of a run?Because the evidence is clear that after Verlander had tired arm he struggled mightily. That is he got shut down in '05 after 119 innings and in 2006 he got tired arm during late July/early August. In fact, Verlander's ERA stood at 2.69 at the end of July (after 130.33 innings pitched) and, his DIPS ERA at that point was equal to 3.79. I believe that once Verlander gets some arm endurance (he'd be happy to have the endurance that Hughes showed in '06 and will have for '07) then he ought to post similar results or better to those he posted through the end of July last year. Still, that isn't as good as Hughes, but it is about what was predicted by Verlander's (pre-tired arm cumulative PERA) PERA posted by BP.

Pats93
02-04-07, 01:05 AM
So what do P. Martinez 5-11 170; Randy Johnson 6-10 225; and R. Clemens 6-4 220, have in common? Not much except for the fact that they can throw harder than most mortals, do funny things to the spin of a baseball, control the destination of a thrown baseball, and, will be first ballot HOF awardees.
I'm not terribly knowledgeable about the Philip Hughes particulars...
Is his heater in the high 90's consistently?

Yankees fans aren't the first, nor will they be the last to be hyperbolically optimistic about their young talent, but as an ardent fan I'm always a bit offended when people trot out some of the greatest players in baseball's storied history as comparisons for a player who hasn't seen/thrown a single pitch above AA ball.

WHIP
02-04-07, 12:29 PM
I'm not terribly knowledgeable about the Philip Hughes particulars...
Is his heater in the high 90's consistently?

Yankees fans aren't the first, nor will they be the last to be hyperbolically optimistic about their young talent, but as an ardent fan I'm always a bit offended when people trot out some of the greatest players in baseball's storied history as comparisons for a player who hasn't seen/thrown a single pitch above AA ball.

You should be offended. Just like you should be offended by K-Fed's Superbowl commercial.

Saxmania
03-26-07, 08:29 AM
As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them. I'll explain more to you later, before the season starts. Remind me to post all of that before the season starts. But kindly wait until at least the last week of March to remind me.

Ahem. Cough, cough. Loud throat-clearing noise.

Be seeing you,

Saxmania

Soriambi
03-26-07, 10:22 AM
Ahem. Cough, cough. Loud throat-clearing noise.

Be seeing you,

Saxmania

Did you have today marked on your calendar or something? :lol:

Saxmania
03-26-07, 11:54 AM
Did you have today marked on your calendar or something? :lol:

Oh, hush. I don't want a Squeak out of you.

Be seeing you,

Saxmania

MaineSoxFan
03-26-07, 08:18 PM
Skars, you are right about me not going there. I will go there on my own schedule and on my terms, and then everyone can open up the black-box and dissect it. That will happen before the season starts and will allow people to examine the actual projections that come from the system. Still, I thought that I had conceded to you in our last discussion that the Betances and Chamberlain projections carry a very large confidence interval. Indeed on the original thread where I posted those projections for the Royal Flush I was asked by someone, who had the highest bust potential and I noted that it was Sanchez and Betances. So obviously, much still needs to happen for the projections to come to fruition, as I projected them to be.As to the specifics of those projections I will say that the methodology relies much more heavily on subjective analysis on my part than the stricter MLE projections for players with at least A+ ball behind them. I'll explain more to you later, before the season starts. Remind me to post all of that before the season starts. But kindly wait until at least the last week of March to remind me.



Here we are, the last week of March and the season starts next week. Open up that black box and let's see what is inside and take it for a spin.

CPatriarch
03-28-07, 01:56 PM
Hughes2.50, Quick question...are you Phil's agent?

Hughes2.50
03-31-07, 06:46 PM
The estimation of pitching efficiency is an important task in evaluating pitching prospects. In 2001 Clay Dreslough created what he called a defense indendent component era. The variables of interest ate homeruns, walks, hit batters and strikeouts. Those four statistics are under the control of the pitcher and are 'independent' of fielding and the vagaries of hitting performance (luckly bounces, etc.). It has been argued that defense indpendent statistics are a better predictor of future ERA when used instead of regular era/. Baseball Prospectus for example uses what they call PERA which does include hits allowed. I chose to not use that additional statistic for the reason given above.<p>Given all of that, I have found that Dreslough's statistic is a valuable component when used in conjunction with major league translations (here I use MLE's generated by <b>MinorLeagueSplits.com statistics and use the cumulative year numbers not broken up by level). At the botton of a pitcher's line, for cumulative stats, you can find his MLE's. I use only the four statistics listed above.<p>Dreslough's formula is <br>3.00 + (13HR + 3(BB + HBP)) - 2K / IP<p>Using the MLE's from the site listed above and using the above formula below I list selected pitchers who pitched in the minors last year, pitch for American League teams, and generated results better than league average ERA+ > 100. <p>The assumption here is that when a pitcher produces an MLE result which is above league average, as calculated using MLE's it is a matter of establishing consistency to do it again and again, and of course one would expect improved command, in general over time. <p>I list the pitcher, his age, the team he pitches for his handedness, his results from last year on the major league equivalent ERA+ (league average standardized to ERA = 4.50) and then I calculate his career ERA by taking his raw walk and hit batter rate for the year and replace his MLE's with those numbers (Pitchers generally improve their command for the bulk of their career over the MLE values provided in any given year); then I calculate the difference in ERA+ between the two values and do a cubic transformation to normalize the distribution of the scores (I'll describe this in more detail if you are interested in the rationale).

1) Nick Adenhart* 21 LAA RHP (2006 MDERA160; Career DERA 184)<br>2) <b>Hughes</b> 21 NYY RHP (2006 MDERA 153; Career DERA 167)<br>3) <b>Sanchez</b>** 24 NYY RHP (2006 MDERA 130; Career DERA 142)<br>4) Garza 24 MT RHP (2006 MDERA 129; Career DERA 137)<br>5) Windsor 25 OA RHP (2006 MDERA 128; Career DERA134)<br>6) Buchholz 23 BRS RHP (2006 MDERA 114; Career DERA 128)<br>7) Erbe 20 BO RHP (2006 MDERA 108; Career DERA 128)<br>8) Slowey 23 MT RHP (2006 MDERA 122; Career DERA 127)<br><b>9) Jackson</b> 25 NYY RHP (2006 MDERA 116; Career DERA 123)<br>10) Penn 23 BO RHP (2006 MDERA 116; Career DERA 122)<br>11) Mendoza 20 RHP LAA (2006 MDERA 111; Career DERA 121)<br>12) Masset (2006 25 WS RHP (2006 MDERA 113; Career DERA 121) <br>13) Miller 23 CI RHP (2005 MDERA 112, Career DERA 119)<br>14) Bowden 21 BRS RHP (2006 MDERA 100; Career DERA 116)<br>15) <b>Clippard</b>** 22 NYY RHP (2006 MDERA 105; Career DERA 113)<br>16) <b>Ohlendorf</b> 25 NYY RHP (2006 MDERA107; Career DERA 112)<p>* I mentioned here somewhere that another pitcher than Hughes got the top spot based off last years results. Staying true to the model, I list Adenhart first above, eventhough when you look at all of the data, it appeared that Hughes was the better pitcher, when using just the model to inform (something that should never be done, by the way, it is correct to list Adenhart first. Still, this may be a harbinger for Adenhart's career, and shouldn't surprise many informed fans who remember that he was the highest rated pitcher before he succumbed to TJ surgery before the 2004 draft. And, in all fairness late last year, before I did these calculations I doubted that Adenhart would be an ace, and even argued that Tommy Mendoza would end up better (Mendoza is listed at #11 above.** Clippard's values are generated from July, Aug, and Sept 2006; Sanchez from Apr, May and June, both due to injuries.<p>Pitchers in short season or developmental leagues are not listed above.

Hughes2.50
03-31-07, 11:21 PM
As to the Betances and Chamberlain estimates. When I provided those estimates in the fall of last year only Baseball America (of the commercial evaluators) had provided their top ten of the Yankees prospects. My evaluations of both pitchers were heavily dependent upon the reports of Baseball America, and other, non-commercial sources. <p>One criticism levied against those projections was that it was impossible to suggest that a pitcher who had never pitched above short-season ball could not be evaluated as better than a prospect who had pitched in the high minors. I suspect that most people if they give it a bit of thought can see why that argument is faulty. If it were true, how could baseball talent evaluators expect Andrew Miller to be an ace level prospect in the majors off 5 innings in the low minors? How could a high school pitcher get drafted before a college pitcher? The point here is that when scouts suggest that a pitcher has a 'once in a generation set of tools' (Betances) and has 'high end ace stuff' (Chamberlain) you have to adjust your expectations accordingly if you value the scouts opinions (and I do). That doesn't mean that everything will work out, but it does mean that a talent like what Betances has, should be carefully nurtured by the wise organization. Because the return on investment in such situations could be extraordinary. <p>BTW, the pitcher that Betances has most often been compared to is JR Richard. Interestingly, over his last 406 innings (last two years) Richard had an ERA+ of 142 and a WHIP of 1.04. What I have been led to believe is that Betances seems to be a talent equal to Richard's but with much better command for his age, indeed Betances appears to have better command than Richard ever demonstrated! Given that, I have a very high estimate of Betances' future, even if he still has a ways to go, and even if he still has a significant bust possibility due to his still tender age (with all of the attendant injury possibilities, and developmental failures which still could rear their ugly heads). <p>Chamberlain, because he is a bit older, and because his skill set appears to be more settled is more predictable than Betances. Whereas Betances' skill set appears to be very, very, rare - Chamberlain appears to be more the protypical power pitcher. Still, with a heater that can reach 97 and which sat comfortably above 94, Chamberlain's fastball grades out at a 75 on the scouts 20-80 scale. Now it must be noted that these measures occured in Hawaii, in warm weather, and after Chamberlain worked with Nardi Contreras after last year's draft. My guess is that Chamberlain has a very good chance to have a career on a par with Curt Schilling's. <p>So, when push comes to shove, I'd say that Betances has better than a 50/50 chance to have a career ERA+ over 150 (maybe somewhat higher, if his command is really good - he could end up like a lefthanded Randy Johnson for his career - minus the poor results to start his career), and Chamberlain has a better than 50/50 chance to have a career ERA+ better than 130. <p>And of course, everything I just said is subject to change when we start to see significant innings with MLE's attached to those innings, beginning this year.

Hughes2.50
04-02-07, 08:26 PM
Sabermetricians, have been interested in estimating the effectiveness of offensive producers (independent of runs batted in, and runs scored - because those two statistics are so dependent upon the team that the player plays on). Bill James asked this question, in this way:<p>
With regard to an offensive player, the first key question is how many runs have resulted from what he has done with the bat and on the basepaths. Willie McCovey hit .270 in his career, with 353 doubles, 46 triples, 521 home runs and 1,345 walks -- but his job was not to hit doubles, nor to hit singles, nor to hit triples, nor to draw walks or even hit home runs, but rather to put runs on the scoreboard. How many runs resulted from all of these things? James suggested the equation for created runs; CR = (A * B)/C, where <p>A = On-Base Factor<br>B = Advancement factor<br>C = Opportunity factor. Over the intervening years, others have analyzed this problem, as has James. <p>In 2002, another revision of the formula for created runs emerged, which placed the player's run creation in the context of his team (that is, this statistic does not make the assumption that all players will be similarly efficient - as is the player being evaluated, or, the team being evaluated if it is a team composite statistic). The equation is CR = ((2.4C + A) (3C + B)/9C) - .9C; Where <p>A: H + BB - CS + HBP - GIDP<p>TB + (.24 * (BB - IBB + HBP)) + (.62 * SB) + (.5 * (SH + SF)) - (.03 * K)<p>C: AB + BB + HBP + SH + SF (is equal to the total number of plate appearances)<p>This formula does not include situational hitting.<p>The above formula produces a number representing the total number of created runs. <p>CR+ is calculated by taking the total number of created runs, and dividing the number of runs that would have been generated by a league average run creator, playing instead of the evaluated player. The calculation depends upon obtaining the park factor (found on Baseballreference.com) and the league average for runs. <p>An example:<br>Let us say that the league averages 5.00 runs per 27 outs. And lets say that park A has a .95 hitting factor (that would mean that the average number of runs produced in that park with average run producers over 27 outs in an average game, for the season, would be 4.75 runs). If we know how many outs player A had in the season in question, we would know how many runs the average run producer would create in that park, that year. That number is divided into the runs created by the player in question and gives you a number something like this 100/80 = 125 or a CR+ of 125.<p>Over the last few months, I've introduced the above statistic and called it Runs Created Plus (CR+). The statistic is analogous to the ERA+ statistic used to evaluate pitching efficiency. Like ERA+, CR+ is designed to measure runs, in this case runs created by offensive performers. ERA+ is designed to measure runs prevented by pitchers. Like ERA+, CR+ is standardized to have the score 100 represent league average, and scores higher that that represent better run production. Like ERA+, CR+ is particularly useful when comparing players, because it is adjusted for league and park effects, and is intended to provide evaluators with the ability to compare performers where the basic question is, how effective is this player, or team, as run creators?<p>Since 100 represents league average, it is convienent to know that a score of 200 is twice as productive at creating runs as its 100 counterpart is. And, a score of 50 is half as productive as is league average of 100. Earlier this year, I pointed out that, like ERA+ CR+ can be part of a formula that can be used to calculate winning percentages for a team (given the composite team ERA+ and CR+). Or, that a pitcher's winning percentage is directly related to his ERA+ and his teams run support (essentially his team's CR+ when he pitches).<p>Below I provide a list of run creators in the Anerican League for the year 2006. The first number (after their name, position, and team are listed, is the year CR+ score for that player, and the second number is the number of runs that player is calculated to have produced over the league average performer (i.e., CR+ = 100), if that league average performer had taken that particular player's place for all of their offensive appearances at the plate and the bases. <p>

Travis Hafner DH Indians - 208 (64)<br>
Manny Ramirez LF Red Sox - 184 (53)<br>
David Ortiz DH Red Sox - 176 (60)<br>
Jim Thome DH White Sox - 175 (52)<br>
<b>Jason Giambi</b> DH Yankees - 169 (43)<br>
<b>Bobby Abreu</b> RF Yankees - 166 (18)<br>
<b>Derek Jeter</b> SS Yankees - 164 (49)<br>
Jermaine Dye RF White Sox - 158 (44)<br>
Vladimir Guerrero RF Angels - 155 (42)<br>
Joe Mauer C Twins - 154 (37)<br>
Grady Sizemore CF Indians - 153 (46)<br>
Justin Morneau 1B Twins - 152 (41)<br>
Frank Thomas DH Athletics - 148 (31)<br>
<b>Alex Rodriquez</b> 3b Yankees - 145 (35)<br>
<b>Hideki Matsui</b> LF Yankees - 143 (10)<br>
<b>Jorge Posada</b> C Yankees - 139 (24)<br>
Ichiro Suzuki RF Mariners - 138 (32)<br>
<b>Robinson Cano</b> 2B Yankees - 136 (22)<br>
Paul Konerko 1B White Sox - 135 (28)<br>
<b>Johnny Damon</b> CF Yankees - 134 (26)<br>
Rocco Baldelli CF Devil Rays 134 (17)<br>
Marcus Thames DH Tigers - 134 (16)<br>
Vernon Wells CF Blue Jays - 133 (29)<br>
Michael Cuddyer RF Twins - 133 (25)<br>
Raul Ibanez LF Mariners - 132 (27)<br>
Gary Matthews CF Rangers - 128 (24)<br>
Victor Martinez C Indians - 128 (21)<br>
Mark Teixeira 1B Rangers - 126 (24)<br>
Carl Crawford LF Devil Rays - 125 (21)<br>
Casey Blake RF Indians - 122 (12)<br>
Alex Rios RF Blue Jays - 121 (14)<br>
Ryan Garko 1B Indians - 121 (5)<br>
Milton Bradley RF Athletics - 119 (9)<br>
Kevin Youkulis 1B Red Sox - 118 (15)<br>
Maglio Ordonez RF Tigers - 117 (13)<br>
Ben Broussard 1B Indians/Mariners - 117 (10)<br>
Delmon Young RF Devil Rays - 116 (3)<br>
Kevin Millar 1B Orioles - 115 (9)<br>
Alex Gordon* 3B Royals - 114 (10)<br>
Frank Catalano LF Blue Jays - 114 (9)<br>
<b>Gary Sheffield</b> RF Yankees - 114 (3)<br>
Tim Salmon DH Angels - 110 (3)<br>
Torri Hunter CF Twins - 109 (7)<br>
Nick Markolis RF Orioles - 109 (6)<br>
Mark Derosa RF Rangers - 108 (6)<br>
<b>Melky Cabrera</b> LF Yankees - 107 (4)<br>
Chris Shelton 1B Tigers - 107 (3)<br>
Emil Brown LF Royals - 105 (4)<br>
Curtis Granderson CF Tigers - 105 (4)<br>
Corey Patterson CF Orioles - 105 (3)<br>
Garrett Anderson LF Angels - 104<br>
Trott Nixon RF Red Sox - 103 (2)<br>
Doug Mientkiewiz 1B Royals - 102 (1)<br>
<b>Bernie Williams</b> CF Yankees - 102 (1)<p>
*Alex Gordon, the highly rated rookie third baseman for the Royals produced a 114 CR+ when I did his calculations using MLE's (he played in the minors in 2006) from MinorLeagueSplits.com.<p>Travis Hafner, was an overwhelming creator of runs in 2006, and surely did not receive the attention for MVP that he deserved. Voters for MVP awards are heavily influenced by how well a team does, and then they tend to overrate the best performances on those teams. For example, last year, the Twins were a tremendous success story in the second half of the season, Justin Morneau benefitted by his teams success - eventhough it is clear that Derek Jeter, Joe Mauer (both he and Jeter play much more difficult positions than does Morneau) and even the pitcher Johan Santana had better years than Morneau did. In fact, Hafner and Ortiz did as well, but were hurt by the aforementioned poor overall seasons their teams had.<p>Based on results of last season for the created runs measure CR+, I project that the American League first team all star team at the end of the season will be<p>C - Joe Mauer, Twins<br>IB - Justin Morneau, Twins<br>2B - Robinson Cano, Yankees<br>SS - Derek Jeter, Yankees<br>3B - Alex Rodriquez, Yankees<br>LF - Manny Ramirez, Red Sox<br>CF - Grady Sizemore, Indians<br>RF - Bobby Abreu, Yankees<p>I expect that the AL MVP will be the CF Grady Sizemore, of the Indians, if the Indians make a strong comeback this year, and if Sizemore continues to improve (mostly by adding more power to his already outstanding overall game)<p>I then calculated the 2005 results for David Ortiz and Alex Rodriquez, here are the results<p>Alex Rodriquez 3B Yankees - 190 (71)<br>David Ortiz DH Red Sox - 173 (58)<p>Clearly, given the fact that Rodriquez was a position player who had a decent year (2005 at third base) and, that Ortiz was strictly a designated hitter it is clear by these measures that Alex Rodriquez clearly earned the 2005 AL MVP.<p>I believe that CR+ is a parsimonious way to examine how individual offensive performers effect the probability that their team will win baseball games, and that the composite CR+ along with the composite ERA+ for any given team is a very strong predictor of that teams winning percentage.<p>Given that, should the Yankees remain healthy and produce similarly on the CR+ as they did last year, one can see that their composite CR+ has a chance to be even higher this year than last year's league leading 126 team CR+. A quick way to estimate the Yankees CR+ is to look at the nine starters add up their CR+ numbers and divide by nine - that would give you a rough estimate, if we assume that all players get approximately the same number of offensive opportunities (of course this is a gross oversimplification but allows anyone to quickly estimate how productive a given team would be, or, how much better a given team might be by replacing a player with a given CR+ with another player with a different CR+).

pedromartinezfan
04-03-07, 05:08 PM
No pitcher has a 50/50 chance of having a career ERA+ in between Lefty Grove and Pedro Martinez. Where do you have Felix? 240?

ABCBaseball
04-03-07, 11:43 PM
The estimation of pitching efficiency is an important task in evaluating pitching prospects. In 2001 Clay Dreslough created what he called a defense indendent component era. The variables of interest ate homeruns, walks, hit batters and strikeouts. Those four statistics are under the control of the pitcher and are 'independent' of fielding and the vagaries of hitting performance (luckly bounces, etc.). It has been argued that defense indpendent statistics are a better predictor of future ERA when used instead of regular era/. Baseball Prospectus for example uses what they call PERA which does include hits allowed. I chose to not use that additional statistic for the reason given above.Given all of that, I have found that Dreslough's statistic is a valuable component when used in conjunction with major league translations (here I use MLE's generated by MinorLeagueSplits.com statistics and use the cumulative year numbers not broken up by level). At the botton of a pitcher's line, for cumulative stats, you can find his MLE's. I use only the four statistics listed above.
Dreslough's formula is
3.00 + (13HR + 3(BB + HBP)) - 2K / IP
Using the MLE's from the site listed above and using the above formula below I list selected pitchers who pitched in the minors last year, pitch for American League teams, and generated results better than league average ERA+ > 100.
The assumption here is that when a pitcher produces an MLE result which is above league average, as calculated using MLE's it is a matter of establishing consistency to do it again and again, and of course one would expect improved command, in general over time.
I list the pitcher, his age, the team he pitches for his handedness, his results from last year on the major league equivalent ERA+ (league average standardized to ERA = 4.50) and then I calculate his career ERA by taking his raw walk and hit batter rate for the year and replace his MLE's with those numbers (Pitchers generally improve their command for the bulk of their career over the MLE values provided in any given year); then I calculate the difference in ERA+ between the two values and do a cubic transformation to normalize the distribution of the scores (I'll describe this in more detail if you are interested in the rationale).

1) Nick Adenhart* 21 LAA RHP (2006 MDERA160; Career DERA 184)
2) Hughes 21 NYY RHP (2006 MDERA 153; Career DERA 167)
3) Sanchez** 24 NYY RHP (2006 MDERA 130; Career DERA 142)
4) Garza 24 MT RHP (2006 MDERA 129; Career DERA 137)
5) Windsor 25 OA RHP (2006 MDERA 128; Career DERA134)
6) Buchholz 23 BRS RHP (2006 MDERA 114; Career DERA 128)
7) Erbe 20 BO RHP (2006 MDERA 108; Career DERA 128)
8) Slowey 23 MT RHP (2006 MDERA 122; Career DERA 127)
9) Jackson 25 NYY RHP (2006 MDERA 116; Career DERA 123)
10) Penn 23 BO RHP (2006 MDERA 116; Career DERA 122)
11) Mendoza 20 RHP LAA (2006 MDERA 111; Career DERA 121)
12) Masset (2006 25 WS RHP (2006 MDERA 113; Career DERA 121)
13) Miller 23 CI RHP (2005 MDERA 112, Career DERA 119)
14) Bowden 21 BRS RHP (2006 MDERA 100; Career DERA 116)
15) Clippard** 22 NYY RHP (2006 MDERA 105; Career DERA 113)
16) Ohlendorf 25 NYY RHP (2006 MDERA107; Career DERA 112)
* I mentioned here somewhere that another pitcher than Hughes got the top spot based off last years results. Staying true to the model, I list Adenhart first above, eventhough when you look at all of the data, it appeared that Hughes was the better pitcher, when using just the model to inform (something that should never be done, by the way, it is correct to list Adenhart first. Still, this may be a harbinger for Adenhart's career, and shouldn't surprise many informed fans who remember that he was the highest rated pitcher before he succumbed to TJ surgery before the 2004 draft. And, in all fairness late last year, before I did these calculations I doubted that Adenhart would be an ace, and even argued that Tommy Mendoza would end up better (Mendoza is listed at #11 above.** Clippard's values are generated from July, Aug, and Sept 2006; Sanchez from Apr, May and June, both due to injuries.
Pitchers in short season or developmental leagues are not listed above.If your projections list 2 pitchers in the minors as having the best ERA+'s OF ALL TIME wouldn't that make you think the problem is with the projections not the common sense of knowing that won't happen?

Hughes2.50
04-04-07, 12:17 AM
[/B]If your projections list 2 pitchers in the minors as having the best ERA+'s OF ALL TIME wouldn't that make you think the problem is with the projections not the common sense of knowing that won't happen?You really should read the fine print<p>
when using just the model to inform (something that should never be done, by the way

ABCBaseball
04-04-07, 01:31 AM
You really should read the fine printYou seem to use it to inform in the vast majority of your posts. You basically said that Betances has a 50% chance of being a top 3 pitcher of all time.

Hughes2.50
04-04-07, 09:57 AM
You seem to use it to inform in the vast majority of your posts. You basically said that Betances has a 50% chance of being a top 3 pitcher of all time.And from my comments regarding Betances:<p>
And of course, everything I just said is subject to change when we start to see significant innings with MLE's attached to those innings, beginning this year.

ABCBaseball
04-04-07, 01:06 PM
And from my comments regarding Betances:So if your going to backpedal off those statements when you are challenged why make them in the first place?

Hughes2.50
04-04-07, 04:01 PM
So if your going to backpedal off those statements when you are challenged why make them in the first place?I'm not backpeddling at all. You see ABCbaseball some people are able to entertain multiple possiblities as potential realities. The model suggests one thing, scouting reports another, and additional information from the newest season yet some more. You on the other hand are content to say that:
If your projections list 2 pitchers in the minors as having the best ERA+'s OF ALL TIME wouldn't that make you think the problem is with the projections not the common sense of knowing that won't happen?<p>You say the above
as if the possibility is closed off to the young pitchers because they would have to be gods to accomplish that. The fact is that there are many more people who can do certain things, and never do them (whereas some do). These posts are an attempt to use objective analysis as a leading indicator of what we might expect from these pitchers (given the MLE and DIPS analysis). It doesn't mean that these pitchers will reach the heights that are projected for them by the model. Yet it is useful and informative to see where these young pitchers stand relative to other pitchers when analyzed using MLE and DIPS analysis. <p>It is also important when having such discussions to actually understand what these measures are conceptually designed to do. Your answer to my PERA question last week left me with the impression that you actually didn't understand what this type of analysis means. Also, last week you seemed to be under the impression that such comparisons (between the minor leagues and the major leagues) were absolutely useless.<p>
Thats exactly what it is. They have no idea what the real disparity between the NL and AL is how can they know what it is between minor leagues and MLB? Anyone trying to do that is merely guessing. And if they are trying to make stats based on that it's even more so guess work.If you still believe that, are you just here to bait me?<p>BTW, I would be more than happy to answer questions, or get into a discussion about these issues. But don't think that it is perfectly ok to be totally uninformed, throw out thinly veiled baiting comments, and then claim that the person you are baiting can't stand criticism.<p>If you have specific criticisms lets hear them, and let me respond to them.

ABCBaseball
04-05-07, 12:02 AM
I'm not backpeddling at all. You see ABCbaseball some people are able to entertain multiple possiblities as potential realities. The model suggests one thing, scouting reports another, and additional information from the newest season yet some more. You on the other hand are content to say that: You say the above
as if the possibility is closed off to the young pitchers because they would have to be gods to accomplish that. The fact is that there are many more people who can do certain things, and never do them (whereas some do). These posts are an attempt to use objective analysis as a leading indicator of what we might expect from these pitchers (given the MLE and DIPS analysis). It doesn't mean that these pitchers will reach the heights that are projected for them by the model. Yet it is useful and informative to see where these young pitchers stand relative to other pitchers when analyzed using MLE and DIPS analysis.If thats the case then they are worthless as predictors because they can change based on one AB. The problem you have is that you come out with these statements (like that Betances will be Randy Johnson, only better, or that Steven Jackson will have a career smiler to andy pettitte) and don't then say how amazingly speculative your predictions are and how volatile they can be. So volatile in fact that reading ANYTHING into them makes little sense at all. You have TWENTY THREE INNINGS to go on. You're taking some random comment from a random scout from BA and trying to assign a value to it. It's COMPLETE speculation.

It is also important when having such discussions to actually understand what these measures are conceptually designed to do. Your answer to my PERA question last week left me with the impression that you actually didn't understand what this type of analysis means. Also, last week you seemed to be under the impression that such comparisons (between the minor leagues and the major leagues) were absolutely useless.
If you still believe that, are you just here to bait me?
BTW, I would be more than happy to answer questions, or get into a discussion about these issues. But don't think that it is perfectly ok to be totally uninformed, throw out thinly veiled baiting comments, and then claim that the person you are baiting can't stand criticism.
If you have specific criticisms lets hear them, and let me respond to them.I think you're the one whos having the problem understanding the facts. Stats like ERA+ are adjusted against the league average. That kind of stat CANNOT with 100% accuracy adjust for the competitive imbalance between the NL and the AL because the number with witch to multiply the stats by is not a constant every year. Some years the disparity is much bigger then others. Thats why you take stats like that with a grain of salt. It's a FACT that the league that Jackson put up his numbers in was one of the best pitchers leagues in the country and it's also a FACT that Johnson put up his numbers in the best hitting league in the world. Thinking you have a fool proof way to compare those stats just shows a misunderstanding of the way these stats are compiled and the disparity in the level of competition.

ABCBaseball
04-05-07, 12:40 AM
If you still believe that, are you just here to bait me?
BTW, I would be more than happy to answer questions, or get into a discussion about these issues. But don't think that it is perfectly ok to be totally uninformed, throw out thinly veiled baiting comments, and then claim that the person you are baiting can't stand criticism.
If you have specific criticisms lets hear them, and let me respond to them.If you come out with statements that many people consider to be very overly optimistic to say the least then you can't accuse someone who questions them of bating you. Thats not being honest at all. And just because i disagree with you doesn't mean i'm uninformed.

Hughes2.50
04-05-07, 01:13 AM
If you come out with statements that many people consider to be very overly optimistic to say the least then you can't accuse someone who questions them of bating you. Thats not being honest at all. And just because i disagree with you doesn't mean i'm uninformed.See the following for my response.

If thats the case then they are worthless as predictors because they can change based on one AB. Where do you get this stuff? Who said that anything would change significantly based on one at bat?
The problem you have is that you come out with these statements (like that Betances will be Randy Johnson, only better, or that Steven Jackson will have a career smiler to andy pettitte) and don't then say how amazingly speculative your predictions are and how volatile they can be.Apparently you still are mistaking the comments about Betances and Chamberlain, with the comments about the 16 pitchers above (including Jackson), which is based on two different approaches. One relies more on a statistical model (the one with Jackson in it), the other based on scouts' (that's plural sources fyi) assessnments about the projectability of Betances and Chamberlain - that is why they were presented in two separate posts.
So volatile in fact that reading ANYTHING into them makes little sense at all. You have TWENTY THREE INNINGS to go on. You're taking some random comment from a random scout from BA and trying to assign a value to it. It's COMPLETE speculation. Like I said in the post about Betances and Chamberlain, apparently some argue (you appear to fall in this category) that it is impossible for scouts to assess that a high school player has ace potential, or that a guy who has five innings in the minors like Miller of the Tigers, has ace potential. I happen to believe that scouts can make very good assessments that project several years into the future, and, that some scouts are better at this than others. You apparently think this is no more useful that reading tea leaves I guess.
I think you're the one whos having the problem understanding the facts. Stats like ERA+ are adjusted against the league average. That kind of stat CANNOT with 100% accuracy adjust for the competitive imbalance between the NL and the AL because the number with witch to multiply the stats by is not a constant every year.Sorry to inform you of this but ERA+ does not try to compare ERA's in the two major leagues. Since it doesn't, you once again show your lack of knowledge about these matters. Now, unfortunately, you need to be told that so that you can go back and get a bit more knowledgable about these matters before you can make meaningful criticisms. Please actually take a few days, try to work through the definitions, and then you ought to have a better appreciation for what these statistics are designed to do. It would seem that you are confusing my statistic MDERA and Baseball Prospectus' PERA, for ERA or ERA+. If so, I suggest that you re-read what methodology I followed, or get ahold of a web page, or a book that walks you through BP's process for getting PERA stats. You will find that once you learn what they do, you will realize that such numbers are not absolute, or facts, or even necessarily accurate - they are merely estimates based on following a standardized process with the intent of describing the effectiveness of pitchers independent of park, league, and level. That means, that any given estimate might be significantly off, but in aggregate, the estimates are generally pretty accurate (here I am referring primarily to the PERA's and MDERA stats. So, in conjunction with scouting reports, and additional information, over time these estimates paint a fairly accurate picture of most pitchers.
Some years the disparity is much bigger then others. Thats why you take stats like that with a grain of salt. It's a FACT that the league that Jackson put up his numbers in was one of the best pitchers leagues in the country and it's also a FACT that Johnson put up his numbers in the best hitting league in the world.Your comments here, reflect your lack of appreciation for what MLE analysis is all about. If you would just take the time to seriously study this stuff we could have a much better discussion, and you would be less inclined to just assume that the process is useless. No method for estimating something of interest is foolproof, no one would with any working knowledge of such processes would ever suggest otherwise. However, these models can go a long way in providing insight, particularly when looking at muliple measures over several years for a particular player, or, in larger aggregate estimates. However, one thing should be noted, when you have a pitcher getting an MLE estimate that is very strong, based on over 130 innings in any league, whether it is in the minors or majors, it is a major marker indicating strong potential. The idea is, that superior results as measured by this process is unlikely to occur unless there is underlying talent. That is, weak talent never produces exceptional results. However, strong talent, can produce weak performances. The idea is to identify those players who have talent.
Thinking you have a fool proof way to compare those stats...As I said before, there are no foolproof methods, there are only attempts to use scientific methods based in solid methods, to estimate a process of interest. In this case, we are referring to the process of measuring major league talent, before it gets to the majors.
just shows a misunderstanding of the way these stats are compiled and the disparity in the level of competition.The misunderstanding is all yours, and I expect that you will educate yourself about what PERA means, how it is derived, and what it suggests when comparing pitchers in the major leagues and the minors with each other. Until you indicate you actually understand that I see no reason for you to continue to raise elementary concerns. It really isn't that hard to understand, and I look forward to some good discussions with you on this topic once you get a better background with this material.

ABCBaseball
04-05-07, 01:41 AM
See the following for my response.
Where do you get this stuff? Who said that anything would change significantly based on one at bat?Apparently you still are mistaking the comments about Betances and Chamberlain, with the comments about the 16 pitchers above (including Jackson), which is based on two different approaches. One relies more on a statistical model (the one with Jackson in it), the other based on scouts' (that's plural sources fyi) assessnments about the projectability of Betances and Chamberlain - that is why they were presented in two separate posts. Like I said in the post about Betances and Chamberlain, apparently some argue (you appear to fall in this category) that it is impossible for scouts to asess that a high school player has ace potential, or that a guy who has five innings in the minors like Miller of the Tigers, has ace potential. I happen to believe that scouts can make very good assessments that project several years into the future, and, that some scouts are better at this than others. You apparently think this is no more useful that reading tea leaves I guess.Sorry to inform you of this but ERA+ does not try to compare ERA's in the two major leagues. Since it doesn't, you once again show your lack of knowledge about these matters. Now, unfortunately, you need to be told that so that you can go back and get a bit more knowledgable about these matters before you can make meaningful criticisms. Please actually take a few days, try to work through the definitions, and then you ought to have a better appreciation for what these statistics are designed to do. It would seem that you are confusing my statistic MDERA and Baseball Prospectus' PERA for ERA or ERA+. If so, I suggest that you re-read what methodology I followed, or get ahold of a web page, or a book that walks you through BP's process for getting PERA stats. You will find that once you learn what they do, you will realize that such numbers are not absolute, or facts, or even necessarily accurate - they are merely estimates based on following a standardized process with the intent of describing the effectiveness of pitchers independent of park, league, and level. That means, that any given estimate might be significantly off, but in aggregate, the estimates are generally pretty accurate (here I am referring primarily to the PERA's and MDERA stats. So, in conjunction with scouting reports, and additional information, over time these estimates paint a fairly accurate picture of most pitchers. Your comments here, reflect your lack of appreciation for what MLE analysis is all about. If you would just take the time to seriously study this stuff we could have a much better discussion, and you would be less inclined to just assume that the process is useless. No method for estimating someting of interest is foolproof, no one would with any working knowledge of such processes would ever suggest otherwise. However, these models can go a long way in providing insight, particularly when looking at muliple measures over several years for a particular player, or, in larger aggregate estimates. However, one thing should be noted, when you have a pitcher getting an MLE estimate that is very strong, based on over 130 innings in any league, whether it is in the minors or majors, it is a major marker indicating strong potential. The idea is, that superior results as measured by this process is unlikely to occur unless there is underlying talent. That is, weak talent never produces exceptional results. However, strong talent, can produce weak performances. The idea is to identify those players who have talent.As I said before, there are no foolproof methods, there are only attempts to use scientific methods based in solid methods, to estimate a process of interest. In this case, we are referring to the process of measuring major league talent, before it gets to the majors.The misunderstanding is all yours, and I expect that you will educate yourself about what PERA means, how it is derived, and what it suggests when comparing pitchers in the major leagues and the minors with each other. Until you indicate you actually understand that I see no reason for you to continue to raise elementary concerns. It really isn't that hard to understand, and I look forward to some good discussions with you on this topic once you get a better background with this material.Nowhere did I attempt to say that PERA or ERA+ tries to compare ERA's based on what league the pitcher is in. They are both weighted stats against the league median and i don't need to be lectured on what they are. They are still unreliable in comparing pitchers in different league. Just because pitchers in say the 58% percentile will have similar ERA+'s DOES NOT MEAN that the 58% percentile pitchers in different leagues will be equal OR that PERA is anything more then a guesstimate that common sense should override sometimes.

The fact that PERA calibrates to a league where a 4.50 ERA is average still does not bridge the gap between Jackson and Johnson's stats. Also if you are going to hide behind BP for the PERA stat then why don't I see the kind of projections that you come out with on BP? Your projections are 10x more optimistic then the 90 percentile PECOTA's.

Hughes2.50
04-05-07, 01:56 AM
Nowhere did I attempt to say that PERA or ERA+ tries to compare ERA's based on what league the pitcher is in. They are both weighted stats against the league median and i don't need to be lectured on what they are. They are still unreliable in comparing pitchers in different league. Just because pitchers in say the 58% percentile will have similar ERA+'s DOES NOT MEAN that the 58% percentile pitchers in different leagues will be equal OR that PERA is anything more then a guesstimate that common sense should override sometimes.

The fact that PERA calibrates to a league where a 4.50 ERA is average still does not bridge the gap between Jackson and Johnson's stats. Also if you are going to hide behind BP for the PERA stat then why don't I see the kind of projections that you come out with on BP? Your projections are 10x more optimistic then the 90 percentile PECOTA's.You do need to be lectured, time and time again, you force anyone who actually understands this stuff to correct your innaccuracies. It isn't an insult to be told that you need to get schooled on a topic. If you were being honest with yourself, and weren't trying to pull a fast one, you'd admit you really don't understand this stuff, and you are learning as you go.<p>As for Pecota, my results and Pecota's correlate pretty well, compared to say Zips or First Inning. However, I believe that using the formula for defense independent statistics provides my model with some advantages and that is why I have posted the projections. Since we have sixteen pitchers listed above, and since some should pitch in the majors this year we ought to be able to begin seeing whether my MDERA numbers more accurately reflect actual ERA+ results than does Pecota's Pera statistic from last year or their projections for this year. My Career DERA is merely an algorithmically derived estimate based off last year's results. I won't provide more indepth analysis for free, it isn't worth it to me. People have enough information that they can redo my analysis if they want, like I said it isn't an entirely black box anymore.

ABCBaseball
04-05-07, 09:24 AM
Nm...

Metroidman
04-21-07, 08:52 PM
But you're using hindsight to look at Beckett and Ankiel's makeup. If you went back to 2000, not many people were badmouthing Ankiel's makeup: at most they were saying that he was a little wild. Baseball America loved him even more than they love Hughes. I do remember a couple of things being said about Beckett's makeup, but that makeup can go two ways: he's a bulldog, he never gives in, and also, he's an idiot, he isn't cerebral enough.

I think I misspoke re: variance. There is some list of comparables for Hughes. Some are spectacularly great, others are merely good, some are pretty bad. We don't necessarily know which of those tracks Hughes is on. You keep speaking in the absolutes about Hughes' makeup, but how do you know? Have you talked to him? Even if you have, isn't there some way it's not going to go that perfectly for him. If his makeup isn't as good as you say it is, or even if it is, and he still has growing pains, we're going to have to look at a greater list of comparables than saying: He's Dwight Gooden, no question. Given that, his projection is 3.8 or whatever. He then has some confidence band around that projection.

I would think that his 90th degree projection is extremely good, maybe even better than 2.95, and he has a 90th degree projection better than every pitcher in the league outside of Santana, Halladay, Liriano, Felix, and maybe 2 or 3 more. So his confidence band is also pretty high. But I don't think PECOTA's numbers are crazy, or even particularly wrong:

Some other projection models have him at:

ZIPS has him throwing 164 Innings, with 164 hits, 15 home runs, 58 walks, and 127 strikeouts, for a predicted ERA of 4.06, and a FIP ERA of 3.90.

First Inning predicts on the basis of AAA, and has him throwing 159 Innings, with 134 hits, 46 walks, 150 strikeouts, and 12 home runs, for an ERA of 2.86, and a FIP ERA of about 3.00.

I also think you're doing good work, as Reggievision says, but there's nothing wrong with a little bit of caution, which is what PECOTA and ZIPS are trying to do. First Inning has a list of comps making up their projection that you might not like: http://firstinning.com/players/Philip-Hughes-595/, at the bottom.

Beckett's line is in no way better than Hughes at the same age

Look at the walk rates and the Whip

Thats why Beckett gets destroyed at the ML level

pedromartinezfan
04-22-07, 12:01 PM
3.81 ERA, 8.48 K/9, 2.60 K/BB

His WHIP is 1.24, 13th best active if he qualified

= destroyed.

Betances#1
05-10-07, 09:48 PM
So, when push comes to shove, I'd say that Betances has better than a 50/50 chance to have a career ERA+ over 150 (maybe somewhat higher, if his command is really good)

This is very exciting, as I've been a huge Betances fan for a long, long time. He's a great kid.

You obviously know a lot more about this stuff than I do, but I am familiar with ERA+. I know that over 150 is pretty darn good!

So, if he has a 50% chance of having an ERA+ of over 150, how would you break down the remaining 50%? Obviously he's not guaranteed to have an ERA+ of over 150, so would you break it down something like this:

50% over 150
25% 130-150
20% 100-130
5% below 100

That's just off the top of my head and is show for demonstrative purposes. I have no idea if those numbers are close, but I figure they've got to be near what you say, since he's 50% to be above 150.

Anyhow, I can't wait to see this kid in the majors. What's your ETA on him? I think he's got to be on the fasttrack just like the franchise Hughes!

Philip Hughes Fan
05-10-07, 11:24 PM
So, when push comes to shove, I'd say that Betances has better than a 50/50 chance to have a career ERA+ over 150 (maybe somewhat higher, if his command is really good - he could end up like a lefthanded Randy Johnson for his career - minus the poor results to start his career)

So a teenager that's never thrown a pitch in A ball has a 50% chance of becoming the greatest pitcher ever? That may be the craziest thing I've ever read on the internet.

Betances#1
05-11-07, 12:01 AM
So a teenager that's never thrown a pitch in A ball has a 50% chance of becoming the greatest pitcher ever? That may be the craziest thing I've ever read on the internet.

I don't think Math is crazy. Hughes2.50 has done all the math, can't change the black and white.

It's still just a 50% chance. It's probably another 25% or so that he's only has a career ERA+ of 130-150 (that's my guess, I'm no mathematician!). Hopefully Hughes2.50 sees this thread and will weigh in.

Philip Hughes Fan
05-11-07, 12:08 AM
I don't think Math is crazy. Hughes2.50 has done all the math, can't change the black and white.

It's still just a 50% chance. It's probably another 25% or so that he's only has a career ERA+ of 130-150 (that's my guess, I'm no mathematician!). Hopefully Hughes2.50 sees this thread and will weigh in.

What he's doing is not math. Just because he uses numbers does not mean his method has any basis in reality.

Betances#1
05-11-07, 02:08 PM
What he's doing is not math. Just because he uses numbers does not mean his method has any basis in reality.

I think that's kind of harsh to say. I've reviewed his threads, and Hughes2.50 goes at great length into his methodology. I didn't know about this stuff, it's really quite eye-opening. I recommend you take a look!

Hopefully he'll see this thread and answer my question about Betances. Can't wait for the Sally league to start! What do you think a reasonable expectation for him is in that league? sub 1.50 era or so?

Hughes2.50
05-11-07, 07:35 PM
Betances#1 - Dellin Betances has tremendous talent. The comparables I've got are a more consistent JR Richard, or a young, but a right handed version of Randy Johnson with better control. Pretty heady comparisons. Those comparables, when joined with reports from the Yankees about what they are getting from him in their camps encouraged me to estimate that his performance level will be more consistent than either JR or Randy were early on. If so, he has an excellent chance to fullfill the estimated 150+ ERA+ for his career. There just aren't many pitchers with his combination of physical and athletic tools. Note that the estimate about his career ERA presupposes that his command and control will be better throughout the 21-26 period than JR and particularly Randy, and that his ceiling will be somewhere between those two. That would put him in the range of the 150 ERA+ for his career. Right now based on the little we have to go with in terms of innings beyond high school I'd say that once he makes the majors (probably in his 21st calender year) I'm estimating that the chances of him having an ERA+ higher than 120 is about 85% and above an ERA+ of 90 is about 98%, for his career. These calculations are completely dependent upon him actually reaching the majors (which I'd have to think are very likely, but we have to see how he does from here on out to know more about this)<p>I thought that Betances would have started out in the Sally league which has already been up and running for over a month now. Since he isn't there, and remains in extended spring training, I am unclear about what his status is as of now. And I am unclear about where he will be assigned once he is - I suppose Staten Island is a reasonable bet, although I do not currently know. Maybe someone with more information can respond about why Betances was left behind in EST. I look forward to seeing him pitch this year.<p>Before this year, I estimated his time of arrival in the majors to be 2009. Although I can't say that his not being in the Sally league indicates a setback, I will have to wait to see how he does to continue to estimate that he will arrive in 2009.

Betances#1
05-12-07, 12:28 PM
That's all awesome to hear, thank you! I think it's incredible that math can be used to predict careers so darn accurately. I imagine Betances is the exception rather than the rule, of course, since he has such utterly dominate stuff and makeup. I do have one question though -
If he's 98% to be above 90 ERA+ in his career, what percentage is he to not make the majors at all? Is it like 1% to be below a 90 ERA+ and 1% to not make the majors? If so, thats pretty awesome that he's almost guaranteed to be a useful major league player, and has a 50% chance of being a great pitcher. This team is really being rewarded for the tough times of the last few years. Another string of championships is definitely coming.

THE KID'S GOT TALENT!!! BE AFRAID MLB! HUGHES AND BETANCES ARE COMING!

Have you thought at all about getting a job with the Yankees? I know the Red Sox have guys that do sabermath. I'm not saying we copy them, but this kind of stuff definitely needs to be known by Cashman and Michaels and George, so they don't trade Betances for a middle reliever or something. I think they could use a guy like you! Or do you think they know what they have in Betances? Also, would he be the first major league player named Dellin?

Hughes2.50
05-12-07, 10:59 PM
That's all awesome to hear, thank you! I think it's incredible that math can be used to predict careers so darn accurately. I imagine Betances is the exception rather than the rule, of course, since he has such utterly dominate stuff and makeup. I do have one question though -
If he's 98% to be above 90 ERA+ in his career, what percentage is he to not make the majors at all? Is it like 1% to be below a 90 ERA+ and 1% to not make the majors? If so, thats pretty awesome that he's almost guaranteed to be a useful major league player, and has a 50% chance of being a great pitcher. This team is really being rewarded for the tough times of the last few years. Another string of championships is definitely coming.

THE KID'S GOT TALENT!!! BE AFRAID MLB! HUGHES AND BETANCES ARE COMING!

Have you thought at all about getting a job with the Yankees? I know the Red Sox have guys that do sabermath. I'm not saying we copy them, but this kind of stuff definitely needs to be known by Cashman and Michaels and George, so they don't trade Betances for a middle reliever or something. I think they could use a guy like you! Or do you think they know what they have in Betances? Also, would he be the first major league player named Dellin?I think it is more useful and accurate to think that math can be used to project, but not necessarily predict careers. That is, a systematic analysis can allow one to project the likely path that say a Betances might follow. But it is an entirely different thing to say that a career can be predicted with certainty. The margin of error, even with the best analytical tools is still substantial. The way that statisticians describe this problem is to say that certain models (statistical methods) under certain conditions (like projecting human performances) have a high margin of error due to uncertainty about the measures used, and, more seriously, missing variables (unmeasured or unknown, yet, crucially important variables which could account for differences in performance). What I mean is this, it is impossible to know if a person will fullfill their promise, without knowing many more things than is usually known to observers. <p>So, you asked about what the chances are that Betances doesn't even make the majors. Unfortunately it is very difficult to quantify that, but it is more likely that he won't make the majors today, than it might be in a year or two (after he continues to progress through the minor league gauntlet). <p>When I said that Betances has a 50/50 chance to post a career ERA+ better than 150, that was based on him actually making the majors and sticking around. We will know a lot more about how probable that will be as time goes on. However, at this early stage I can say that so far, Betances' profile, and his tracked performance as a nineteen year old bode very well indeed. As I said yesterday, I am somewhat in the dark about what Betances' situation is currently and why he is being held back for SS- Staten Island.<p>Obviously I would be more informed if I worked for the Yankees, and of course that would be a very attractive job for someone with my interests and loyalties.<p>However, I have to say that teams already have access to a great deal of information. The only reason that a team might want to hire someone with my skills and interests is if they were to feel that person could provide added value to their analytical approach.

soren
05-13-07, 12:02 AM
I feel like I'm taking crazy pills!

When I saw your other projections thread I Thought to myself "man, this guy is the Yankee version of Eric Van." After reading this thread, though, I realise something much more sinister is afoot. It would be one thing to have a system that takes small sample sizes and permutates them into guys being one of the best pitcher of all time. It's quite another thing to put any sort of faith in those projections. Without going into the entirely imposible predicition about Betances, I'll give you the obvious criticism; there is not a single 18 year old pitcher in all of the minor leauges who has a 98% chances of being a major league pitcher. For one thing, the injury attrition rate for 18 year old pitcher is undoubtedly higher than 2%. I'm sorry if this post deams overly rude, but your projection system is rediculous.

Hughes2.50
05-13-07, 12:28 AM
I feel like I'm taking crazy pills!

When I saw your other projections thread I Thought to myself "man, this guy is the Yankee version of Eric Van." After reading this thread, though, I realise something much more sinister is afoot. It would be one thing to have a system that takes small sample sizes and permutates them into guys being one of the best pitcher of all time. It's quite another thing to put any sort of faith in those projections. Without going into the entirely imposible predicition about Betances, I'll give you the obvious criticism; there is not a single 18 year old pitcher in all of the minor leauges who has a 98% chances of being a major league pitcher. For one thing, the injury attrition rate for 18 year old pitcher is undoubtedly higher than 2%. I'm sorry if this post deams overly rude, but your projection system is rediculous.Maybe you can start by rereading what people are saying. I've not said that any 18 or 19 year old pitcher (specifically Betances) has a 98% chance of being a major leaguer. That comment can be attributed to #Betances and I corrected that misperception on my subsequent reply.<p>I wouldn't conclude that you are taking crazy pills, but maybe glasses would help.
;)

gdn
05-13-07, 06:16 AM
Did I miss the formula in this thread or is it in another thread? I'm very curious to see it.

MaineSoxFan
05-13-07, 10:46 AM
Did I miss the formula in this thread or is it in another thread? I'm very curious to see it.

If you are referring to Hughes2.50's formula it is in the previous page of this thread.

Hughes2.50
05-13-07, 11:05 AM
If you are referring to Hughes2.50's formula it is in the previous page of this thread.I think he already asked that question earlier on this thread, and, was given an answer.<p>Edit: I guess I was wrong. He must have asked that question on another thread. Or maybe it was someone else.

gdn
05-13-07, 11:28 AM
I think he already asked that question earlier on this thread, and, was given an answer.Edit: I guess I was wrong. He must have asked that question on another thread. Or maybe it was someone else.I don't think it was me...



Anyway, I looked through the voluminous posts and found this. Am I correct in thinking that this is your formula?


CR+ is calculated by taking the total number of created runs, and dividing the number of runs that would have been generated by a league average run creator, playing instead of the evaluated player. The calculation depends upon obtaining the park factor (found on Baseballreference.com) and the league average for runs.

MaineSoxFan
05-13-07, 11:30 AM
I don't think it was me...



Anyway, I looked through the voluminous posts and found this. Am I correct in thinking that this is your formula?

If you are referring to pitching projections it is post #93 in this thread, not sure how to link to it.

gdn
05-13-07, 11:32 AM
If you are referring to pitching projections it is post #93 in this thread, not sure how to link to it.This?


I list the pitcher, his age, the team he pitches for his handedness, his results from last year on the major league equivalent ERA+ (league average standardized to ERA = 4.50) and then I calculate his career ERA by taking his raw walk and hit batter rate for the year and replace his MLE's with those numbers (Pitchers generally improve their command for the bulk of their career over the MLE values provided in any given year); then I calculate the difference in ERA+ between the two values and do a cubic transformation to normalize the distribution of the scores (I'll describe this in more detail if you are interested in the rationale).

MaineSoxFan
05-13-07, 12:44 PM
This?

That is my understanding

captainlad
05-13-07, 08:54 PM
I guess my problem with this whole thing is this...
What would you say about these pitchers - I'd like to get your predictions for them: (if you need other data on them, I can give you that.)

Pitcher #1, Age 21, LHP, Between the levels of AA, AAA and MLB has done this:
17-8, 32 starts, 183.2 IP, 146 H, 3.18 ERA, 76 BB, 210 K, 16 HR

Pitcher #2, Age 24, RHP, at A has done this:
13-4, 25 starts, 157.2 IP, 112 H, 2.74 ERA, 55 BB, 180 K, 9 HR

Pitcher #3, Age 20, LHP, at AA has done this:
14-9, 28 starts, 201 IP, 179 H, 3.22 ERA, 89 BB, 171 K, 18 HR

Pitcher #4, Age 22, RHP, at A+ and AA has done this:
9-2, 22 starts, 122 IP, 73 H, 1.40 ERA, 36 BB, 157 K, 7 HR

Pitcher #5, Age 22, RHP, at AA and AAA has done this:
6-9, 25 starts, 140.1 IP, 115 H, 3.46 ERA, 56 BB, 183 K, 15 HR

captainlad
05-13-07, 08:56 PM
The problem with projecting guys in the minors to have ERA+s that make them HOFs isn't the guys, but the method used to do so.

captainlad
05-13-07, 10:47 PM
Using FIP, which is (13*HR + 3*BB - 2*K) / IP . (In order to make this number look more like an ERA, you add 3.20 to the outcome.), I came to the result of this (in order)

3.29
2.70
4.00
2.26
3.19

Those are pretty low. What would those be in ERA+?

apolansk
05-14-07, 12:00 PM
Using FIP, which is (13*HR + 3*BB - 2*K) / IP . (In order to make this number look more like an ERA, you add 3.20 to the outcome.), I came to the result of this (in order)

3.29
2.70
4.00
2.26
3.19

Those are pretty low. What would those be in ERA+?

You would do 4.50(league average) / ERA * 100

Which would result in ERA+ for each of:

<table x:str="" style="border-collapse: collapse; width: 48pt;" border="0" cellpadding="0" cellspacing="0" width="64"><col style="width: 48pt;" width="64"> <tbody><tr style="height: 12.75pt;" height="17"> <td class="xl22" style="height: 12.75pt; width: 48pt;" x:num="136.77811550151975" align="right" height="17" width="64">137</td> </tr> <tr style="height: 12.75pt;" height="17"> <td class="xl22" style="height: 12.75pt;" x:num="166.66666666666666" align="right" height="17">167</td> </tr> <tr style="height: 12.75pt;" height="17"> <td class="xl22" style="height: 12.75pt;" x:num="112.5" align="right" height="17">113</td> </tr> <tr style="height: 12.75pt;" height="17"> <td class="xl22" style="height: 12.75pt;" x:num="199.11504424778764" align="right" height="17">199</td> </tr> <tr style="height: 12.75pt;" height="17"> <td class="xl22" style="height: 12.75pt;" x:num="141.06583072100312" align="right" height="17">141</td> </tr> </tbody></table>

captainlad
05-19-07, 04:59 AM
Can I please get Hughes2.50's response to this?
Thanks.

Hughes2.50
05-20-07, 03:30 PM
I guess my problem with this whole thing is this...
What would you say about these pitchers - I'd like to get your predictions for them: (if you need other data on them, I can give you that.)

Pitcher #1, Age 21, LHP, Between the levels of AA, AAA and MLB has done this:
17-8, 32 starts, 183.2 IP, 146 H, 3.18 ERA, 76 BB, 210 K, 16 HR

Pitcher #2, Age 24, RHP, at A has done this:
13-4, 25 starts, 157.2 IP, 112 H, 2.74 ERA, 55 BB, 180 K, 9 HR

Pitcher #3, Age 20, LHP, at AA has done this:
14-9, 28 starts, 201 IP, 179 H, 3.22 ERA, 89 BB, 171 K, 18 HR

Pitcher #4, Age 22, RHP, at A+ and AA has done this:
9-2, 22 starts, 122 IP, 73 H, 1.40 ERA, 36 BB, 157 K, 7 HR

Pitcher #5, Age 22, RHP, at AA and AAA has done this:
6-9, 25 starts, 140.1 IP, 115 H, 3.46 ERA, 56 BB, 183 K, 15 HR


Using FIP, which is (13*HR + 3*BB - 2*K) / IP . (In order to make this number look more like an ERA, you add 3.20 to the outcome.), I came to the result of this (in order)

3.29
2.70
4.00
2.26
3.19

Those are pretty low. What would those be in ERA+?On the previous page I posted the calculation that I use to determine the defense independent ERA (i.e., DERA). There you will see it differs slightly with the one you posted.<p>However, when you are trying to estimate the performance of a minor league pitcher, in terms of performance as translated into major league effectiveness, you need to know what the relevant mle adjustments would be for the variables, <b>before you calculate</b> the fips (or what I call DERA) statistic. So, your question needs to be supplemented with the information which would allow for an mle adjustment to be made to those posted stats.<p>Hope that helps.

majorwibi
05-20-07, 07:56 PM
On the previous page I posted the calculation that I use to determine the defense independent ERA (i.e., DERA). There you will see it differs slightly with the one you posted.However, when you are trying to estimate the performance of a minor league pitcher, in terms of performance as translated into major league effectiveness, you need to know what the relevant mle adjustments would be for the variables, before you calculate the fips (or what I call DERA) statistic. So, your question needs to be supplemented with the information which would allow for an mle adjustment to be made to those posted stats.
Hope that helps.

I'm curious to see what the projections for those guys posted above are. I've been able to find 4 of them and would be very interested to see how your system stacks up against some historical data given the current set of information presented.

captainlad
05-20-07, 09:16 PM
On the previous page I posted the calculation that I use to determine the defense independent ERA (i.e., DERA). There you will see it differs slightly with the one you posted.However, when you are trying to estimate the performance of a minor league pitcher, in terms of performance as translated into major league effectiveness, you need to know what the relevant mle adjustments would be for the variables, before you calculate the fips (or what I call DERA) statistic. So, your question needs to be supplemented with the information which would allow for an mle adjustment to be made to those posted stats.
Hope that helps.
So what are you saying? Do you want league averages for those players I listed? Or, what? Help me help you.

Hughes2.50
05-20-07, 11:35 PM
So what are you saying? Do you want league averages for those players I listed? Or, what? Help me help you.mle calculations require league and park effect adjustments. I need to know the league and the team (park) that the pitcher pitched in. It would be preferable if I had the year too, because league and park effects do change from year to year (sometimes more one year than another, but each year is at least slightly different).<p>Then you realize that we would be working with a one year estimate _ normally I want a trendline over several years. <p>Still, to do the calculations for any single year I need the information that I mentioned in the first paragraph.

Munson's 'Stash
05-21-07, 08:40 AM
...However, when you are trying to estimate the performance of a minor league pitcher, in terms of performance as translated into major league effectiveness, you need to know what the relevant mle adjustments would be for the variables, before you calculate the fips (or what I call DERA) statistic.



What adjustment? Are these the "scouting" adjustments you make? Wouldn't it be easier and more accurate to calculate the mle's (what I call MiLE's) first? Doesn't making adjustments to the numbers first allow for an unnecessary multiplier that leads to excessive deviation from the mean in the final results?

Maris*61
05-21-07, 12:57 PM
mle calculations require league and park effect adjustments. I need to know the league and the team (park) that the pitcher pitched in. It would be preferable if I had the year too, because league and park effects do change from year to year (sometimes more one year than another, but each year is at least slightly different).<p>Then you realize that we would be working with a one year estimate _ normally I want a trendline over several years. <p>Still, to do the calculations for any single year I need the information that I mentioned in the first paragraph.
Exactly. Peripheral DIPS plays a direct correlation to a pitcher's minor league isolated (k/bb ratio). Park effects are necessary to properly determine the coefficient of the denomenator used to calculate PNDIPS/ERA+.