PDA

View Full Version : A Close Analysis of Yankee Projections.



Munson's 'Stash
05-27-07, 11:20 PM
This thread is meant as a scientific analysis of poster Huges2.50’s projection system. He has not disclosed his full methodology, but he has provided most of the information here in the infamous “post 93.” (http://forums.nyyfans.com/showpost.php?p=3980460&postcount=93) Further information comes from subsequent posts in that thread, and quotes here are from that thread.

The point of this thread is not to attack a member of this community or to belittle his opinions. My goal is to better understand how he has come to his end results and to examine how his methods compare to the standard methods of projecting pitchers. It is meant as a critique of his method, so that everyone here might understand it and where applicable it might be improved.

It should be noted that “critique” is neither inherently positive or negative; I hope everyone who participates in this thread will be able to approach it with the attitude of scientific neutrality with which it was intended. From Here I will attempt to transcribe Hughes2.50’s (from now on referred to as H2.5) method from the original post #93. The original post is somewhat disjointed so it is my hope to provide some clarity by presenting the steps in order where possible. I have also edited some of his quotes as they appear here for spelling and punctuation. The section headings provided below are to help organize the entire enterprise (responses can be directed to or parsed by the appropriate section).

I apologize for the length, but I think it is necessary to be both thorough and complete so as to be fair to all concerned.


Section 1

The start of H2.5’s method is to combine Clay Dreslough’s Defense Independent Component Era (DICE) with the MLE’s provided by minor league splits.com. (http://www.minorleaguesplits.com/index.html)

I have some comments on this so far:

1. The Formula for DICE given in the post is incorrect, the actual formula for DICE is (13HR + 3(BB + HBP) –2K/ IP) + 3.00. I think this merely a typo, but if it is not, the formula listed would lead to very different results as the 3.0 runs are included in the numerator rather than modifying the final line to look more like an ERA.

2. I am curious as to the choice of Dreslough’s DIPs formula as opposed to FIPs (either TangoTiger’s version or TheHardballTimes version). There are subtle differences in the formulas, but the largest difference is that the latter two use a modifier of 3.20 instead of 3.00 to make the result of the base equation look like ERA. Obviously the use of Dreslough’s formula will give a lower number than FIPs (depending on how many HBP the pitcher has, but since this looks at top pitching prospects, I think its safe to assume that they will not have a great number of hit batsmen).

3. At this point I think it’s important to point out that the MLE’s are not predictive of future performance. They are intended to translate a specific performance at the minor league level into numbers at the ML level.

4. The most important point is that H2.5 has not specified in which order he combines the above statistics. I am assuming that he takes the MLE translation of the raw minor league counting stats (HR, BB, Ks) and then plugs those into the DIPs formula, but he hasn’t stated so explicitly. Any other combination of translations should produce skewed results.

Section 2

With these numbers in hand H2.5 generated a list of the top minor league pitchers (see post for a list starting with Adenhart and ending with Ohlendorf) who
…pitched in the minors last year, pitch for American League teams, and generated results better than league average ERA+ > 100.

The formula for ERA+ is 100 * (league average ERA/pitcher’s ERA). H2.5 sets the league average ERA at 4.50.

This is the first major problem in his method. He is arbitrarily setting the league average ERA at 4.50. ERA+ is meant to set a pitcher’s individual performance against the rest of his pitchers in his league. His source data at minor league splits does not provide league ERAs, or even park adjustments. Since different minor leagues tend to vary widely as to whether or not they favor offense or pitching, it is impossible to produce a meaningful statistic by choosing to set ERA at what is not quite a random number (see below).

To give an example of how this adjustment works let’s consider the single best year by ERA+ of the modern era (actually the second best season of all time, behind Tim Keefe in 1880), Pedro Martinez’s 2000 season. He had a 1.74 ERA vs. a 4.97 league average ERA for a 285 ERA+. If the league ERA is arbitrarily set at 4.50 he winds up with a 259 ERA+, which is still good, but drops him into a tie with Bob Gibson in 7th place among single season leaders, so the difference is significant.

Now 4.50 is not a bad guess to set the league average ERA at; the laERA that Pedro pitched against over his career is 4.49. However, the pitchers on this list are pitching in very different environments, and so far H2.5 has not accounted for league and park adjustments in his method. Using the actual minor league average ERAs, no matter how tedious to compile, would go along way towards accounting for this. Again, using Pedro Martinez as an example, the laERA has had a decent amount of variation over the course of his career. Using just his time in the AL, it varied between 4.42 and 5.07. Is there any reason to not expect the sum total of minor league performances to have just as wide (if not wider) variation?

Section 3

I must confess that here’s where H2.5 looses me. I had been able to follow what he was doing up until now, if not exactly why he was doing it. But here at the end he has numbers jumping in and out of various parts of his formula, so I’ll provide his text from the original post in two sections:


… 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)
This is where things get very problematic. H2.5 is extrapolating career numbers based on one season’s worth of minor league data. I can only assume that he’s using the player’s “raw walk and hit batter rates” in the DIPs formula shown in section 1 (along with his HR and K rates?).

I cannot find any reference on any sabermetric (or other) site that would suggest that these rates could be used to project an entire career at the MLB level. The stats as presented would indicate what skill set the player has (or had that particular year) but the rates themselves cannot be expected to simply continue a the same progression. This seems to be an extremely unadvised use of a small sample size of data.

From there he goes on:


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).


I think I speak for a great many people when I say ”Yes! We would very much like to have this explained in more detail!”

What is the purpose in calculating the different ERA+ values, and what exactly is the value in determining the difference between the single season MLE and the career number?

In all the readings and research I have done on baseball, I have never seen reference to a “cubic transformation.” It does not appear to be the basis of any statistical model on any sabermetric site. My background in mathematics is admittedly limited (although that has not hindered any other statistical reading or baseball research I have done), but I have never heard of “cubic transformations.” (Wikipedia doesn’t even have an entry on them (which is more than a little suspicious), and the Google results are mostly about obscure molecules or transitions in N-dimensional space.)

Which leads me to question H2.5’s use of them. His goal is to “normalize the distribution of the scores” but I can’t find any evidence that they either 1. actually do that, or 2. that it is relevant either statistically on a performance basis to “normalize” them in the first place.

Section 4

The above sections apparently end the mathematical part of the process. H2.5 continues, but it is not clear how exactly this continued commentary impacts his projections methodology. I hope he will provide those answers. I will quote him here and offer my response. notes: The list is not quoted, but can be found in the original post. The last part of the quoted text comes from the next sequential post in the series (#94).


I mentioned here somewhere that another pitcher than Hughes got the top spot based off last year’s results. Staying true to the model, I list Adenhart first above, even though 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.

I feel that when presenting a model it is correct to only reference the model within it’s own context. Adjustments made outside the model should be made after the fact and noted as such. A good example would be Nate Silver’s use of PECOTA, where he presents the original data and then points out what his system might be misinterpreting and why.

note: this passage also seems to indicate that there is data for more than one year. It would be very helpful if that data was added to what we have.


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.

This suggests that there is another component to H2.5’s method that does not use any stats whatsoever. The “scouts vs. stats” debate is often overplayed in my mind, but I feel that it stems from the fact that the middle ground between the two sides is compromised by one side looking at what did happen and the other looking for what has the potential to happen.

In any event, we do not know how H2.5 is altering his math to take scouting into account (if indeed he is at all). I do not think projection systems should ever do this. It is better to predict an anomaly (the best example I can think of is PECOTA suggesting that Dustin Pedroia is similar to Garry Sheffield) and then address it and learn from it, than to fudge the data or math in an attempt to make it go away. (Witness Einstein and his use of the cosmological constant. Most theories produce anomalies, and they are often the window to a deeper understanding of the theory and/or nature.)


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 that 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?

This is logically inconsistent and does not address the issue at hand. The point is not that it is impossible to project a player with less experience to have more talent and do better than one with more experience. The point is that statistics measure what has been experienced, and that mathematical models are unlikely to make that projection (without extra help). It is not the fault of the projection system, it is a tool with its own strengths and weakness. But that does not mean that this (or any) particular tool should be compromised in order to make it do everything. (I wouldn’t try to turn a table saw into a can opener.)


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.

This is also seems to indicate that the opinions of scouting are being shoehorned into the statistical model. Scouting is essentially an opinion. It can be an informed opinion, and a professional opinion, but it is still an opinion. I think I can speak for most of the sabermetrically inclined when I say that opinions have no place modifying the numbers within a mathematical system.

By way of example, consider Betances. Despite having “a once in a generation set of tools” he was not taken first overall, and he fell out of the first round due to a “slow start” in the spring. Why are the opinions of other scouts who prefer other players (who were picked earlier in the draft) not reflected in the listed results? Is it even possible?

Players often do not reach the peaks scouts see in them. This is not a fault of scouting (after all, the job of the scout is to find the possibility “where the return on investment is extraordinary”) but that does not mean that a projection system should be adjusted for such “wish casting.”

A look at how scouting is being used by sabermetricians can be found in TangoTiger’s “Wisdom of Crowds” fan database experiment. He combines a broad range of opinions to get a reading on something that is very hard to quantify (defense). The idea is that a broad range of opinions will reduce the possibility of outliers and small samples from corrupting the final data line. I am extremely doubtful that H2.5 has anything close to this level of input from professional scouts or other “non-public sources.” The most important point is that TangoTiger does not use these results to directly modify the defensive system he works on directly (Michael Lichtman’s UZR).

Conclusion

In order to judge how H2.5’s numbers stand up as a projection system, I would like to set my analysis of his system against what I think are axiomatic standards for any projection system.

1. The system must have a clearly defined methodology.

Hopefully I have helped somewhat with this, but obviously there are still questions that are up to H2.5 to answer. It should be noted that this does not mean that a system is necessarily “open source.” PECOTA’s inner workings are not available, but Nate Silver can at least explain in clear and concise language how it works. Problems that may arise from the methodology should be acknowledged as such. For example, using a simple weighted 3-2-1 system does not really work for rookies who don’t have any previous ML playing time, but that is understood.

I feel that it is important to note that he has said that his numbers require league and park adjustments (see here (http://forums.nyyfans.com/showpost.php?p=4142534&postcount=136)), but this is not described or accounted for in the original.

Obviously, there is more work to do.

2. The system must be complete and uniform.

By this I mean that the system has a complete data on all the players within its purview and the data is handled in the same manner for every projection. We should be able to compare any two (or more) like players (i.e. pitchers and position players). This does not rule out making judgments on the raw data set based on reasons outside the system (see the PECOTA example, above), but the data set itself should not be compromised.

The major issue with regards to this is what appears to be the influence of H2.5’s scouting opinions on his numbers. I would hope he can provide an explanation, as without one his results must be considered flawed and suspect.

3. The system must be testable.

Clearly, these are a (I daresay the) major problem with H2.5’s system. We currently only have a small list of player projections, and do not have any way to check against other players outside of the data set he has provided.

Testability does not mean that the system is expected to be 100% accurate, no system comes anywhere close to that standard. It does, however, mean that we should be able to track how players are doing against the system in order to see exactly how well it does. Without a complete roster of projections made it would stand to reason that the sample size is to small to declare any sort of success with the projections it makes that are accurate.

This also means that the system should be testable against nature. In the case of H2.5’s system we can compare his methods and projections against known mathematical processes and baseball populations.

The most basic mathematical truism in baseball is to beware of drawing large conclusions on a small sample size. Unfortunately this is exactly what he has done. He has taken pitchers with only a few professional (and no major league) seasons and extrapolated their entire careers. He has not provided a mathematical basis for making that leap of faith.

He has also projected a pair of 21 year olds to have career ERA+s greater than any other pitcher in history. While it is possible that we are about to bear witness to a new golden age of pitching, common sense would seem to caution against it, and he has not provided a suitable explanation as to why we should forgo that natural impulse.

Fin

I hope that this can set the groundwork for a more complete understanding the numbers and methods discussed, and I sincerely hope that if I have made any errors they will be presented and corrected here.

Edited to correct a few typos.

PaulieIsAwesome
05-28-07, 10:11 AM
Very good post. You've summarized and added to every question I've had about Hughes2.50's projection model. Hopefully we can get him in here, and have him answer some of these fundamental questions.

gdn
05-28-07, 10:34 AM
Fantastic post. His projection system is a little clearer to me now. Very well done.

AMarshal2
05-28-07, 11:13 AM
Fantastic post. His projection system is a little clearer to me now. Very well done.

How do you understand it more clearly now? I think what he has made clear (in case it wasn't clear from the start) is that his projections are bogus. Even if we ignore the problems laid out above, to adjust a MLE by X% in a favorable direction is nothing but arbitrary.

http://www.google.com/search?hl=en&q=%22cubically+transformed%22&btnG=Google+Search (http://www.google.com/search?hl=en&q=%22cubically+transformed%22&btnG=Google+Search)

Excellent post, btw.

captainlad
05-28-07, 07:10 PM
Very good post. You've summarized and added to every question I've had about Hughes2.50's projection model. Hopefully we can get him in here, and have him answer some of these fundamental questions.
Don't count on it.

But, agreed on all accounts - it's nice to see this laid out so carefully and explained and, well, picked over. I'm pretty sure we can all agree that any system that projects rookie league and single-A pitchers to be hall of famers as one that needs to be picked over.

gdn
05-29-07, 06:41 AM
How do you understand it more clearly now? I think what he has made clear (in case it wasn't clear from the start) is that his projections are bogus. Even if we ignore the problems laid out above, to adjust a MLE by X% in a favorable direction is nothing but arbitrary.

http://www.google.com/search?hl=en&q=%22cubically+transformed%22&btnG=Google+Search (http://www.google.com/search?hl=en&q=%22cubically+transformed%22&btnG=Google+Search)

Excellent post, btw.Well, before this thread, I had absolutely no clue how he was getting his projections and what he was using.

Now I somewhat understand what he's using.

So, yes, my understanding is a little clearer.

Edit: As to the first few hits from the Google search - that's just childish.

Hughes2.50
05-29-07, 08:33 AM
Well, before this thread, I had absolutely no clue how he was getting his projections and what he was using.

Now I somewhat understand what he's using.

So, yes, my understanding is a little clearer.

Edit: As to the first few hits from the Google search - that's just childish.As to the google search. And, they wonder why I don't respond.<p>A cubic transformation is merely a way to make poorly fitting data, more easily analyzed by applying a mathematical function to the data set. Although this is a pretty obscure term among people with some mathematical background, it isn't a mystery among people who have been specifically educated to analyze data sets.<p>gdn if you really don't understand something, drop me a pm and I will try to clear it up for you.

AMarshal2
05-29-07, 08:42 AM
Well, before this thread, I had absolutely no clue how he was getting his projections and what he was using.

Now I somewhat understand what he's using.

So, yes, my understanding is a little clearer.

Edit: As to the first few hits from the Google search - that's just childish.

Well, for me, I understand it less. At first I thought he was just toying with MLE's and adjusting them arbitrarily. I didn't realize there were so many other unanswered questions and problems with his projection (like the HOF level career ERA+'s, for instance).

Munson's 'Stash
05-29-07, 09:39 AM
As to the google search. And, they wonder why I don't respond.
A cubic transformation is merely a way to make poorly fitting data, more easily analyzed by applying a mathematical function to the data set. Although this is a pretty obscure term among people with some mathematical background, it isn't a mystery among people who have been specifically educated to analyze data sets.gdn if you really don't understand something, drop me a pm and I will try to clear it up for you.

I apologise if people thought that the google search is somehow out of bounds; my point was to show that this is not a mathematical tool or function that is commonly (or ever, for that matter) used by other baseball analysts. As I stated in the begining, the purpose here is to learn, and I don' t think a look for how his processes are used should be outside that.

If a you have "poorly fitting data" why are you trying to shoehorn it into predictions that do not match nature? If the data does not fit, why does that not point to a more fundamental problem with your methods?

ComeToNYRandy
05-29-07, 10:35 AM
Wait, so are you fitting a cubic polynomial to the data? Is that what you mean?

That might explain some of the more extreme projections, because while polynomials can be made to fit data pretty well, they tend to "explode" outside of the main data sample and could thus lead to some pretty extreme forecasts on inputs that are out of the normal range. Which could happen if we are taking small samples of minor league domination.

Or are you really doing some sort of complex Fourier-type transform? I found this paper (http://www.math.northwestern.edu/~george/preprints/sumtransf/sumtrans.pdf) (PDF), for example, that gives one cubic transform forumla (top of page 11 of the pdf). And to be fair to Hughes, a google search for "cubic transform" gives lots of legitimate hits.

To be fair to the rest of us, however, that looks like nonsense. And I have studied math, applied math, statistics, and econometrics for over 10 years. (I have a phd in econ from a good school, and have always done applied, data-driven work.) Even if you can't give us your precise formula (and I can totally understand you not wanting to, much as BP doesn't give out its formulas for PECOTA), perhaps you can explain intuitively what the cubic transform does?

Is it smoothing out data the way predicted values from a regression would (like fitting the data to a pre-defined functional form, such as a cubic polynomial)? That's something we can all understand, even if the formula/implementation is not something we're going to follow ourselves. Or is it something else entirely?

AMarshal2
05-29-07, 10:59 AM
I apologise if people thought that the google search is somehow out of bounds; my point was to show that this is not a mathematical tool or function that is commonly (or ever, for that matter) used by other baseball analysts. As I stated in the begining, the purpose here is to learn, and I don' t think a look for how his processes are used should be outside that.

If a you have "poorly fitting data" why are you trying to shoehorn it into predictions that do not match nature? If the data does not fit, why does that not point to a more fundamental problem with your methods?

I think that was directed at me. I linked a google search of "cubically transformed" to show that it didn't get any hits that were in any way useful. Of course, the first 3 or 4 hits are from the SoSH forum headers which are poking fun at H2.50. That's life on a message board. You say something silly, you end up in somebody elses sig.

I've been very critical of H2.50 because I don't think his work is credible and I think he's misleading a lot of people. I don't have a PHD in Econ (though I do have a BA), so if he can in any way show that I'm off base, I'd like to see how he formulates his projections. Until then I remain very skeptical.

ComeToNYRandy
05-29-07, 11:09 AM
I think that was directed at me. I linked a google search of "cubically transformed" to show that it didn't get any hits that were in any way useful. Of course, the first 3 or 4 hits are from the SoSH forum headers which are poking fun at H2.50. That's life on a message board. You say something silly, you end up in somebody elses sig.

I've been very critical of H2.50 because I don't think his work is credible and I think he's misleading a lot of people. I don't have a PHD in Econ, so if he can in any way show that I'm off base, I'd like to see how he formulates his projections. Until then I remain very skeptical.

So am I, but I'm trying to give him the benefit of the doubt. I only stated that I have a phd in econ because H2.50 keeps implying that if someone had studied enough, they'd know what he's talking about. Maybe that's true, but I have studied a lot of math and statistics, and I do not follow what he's saying. That's not to say that he's necessarily doing something wrong, but I would like at least an intuitive explanation of what the (now infamous) cubic transform is doing.

27 in 68
05-29-07, 11:13 AM
Good thread - I appreciate the discussion.

Whenever I see the word 'fit', in relation to numbers or stats, a red flag goes up. Data can be massaged to make existing trends more clearly visible, but drawing HOF conclusions from AA stats seems a leap. Once one engages in 'fitting' numbers, the possibility that one's own bias' leaks into the result is great.

Statistics, sufficiently tortured, will confess to anything.

Dow Jones
05-29-07, 11:25 AM
As to the google search. And, they wonder why I don't respond.
I'm pretty sure no one is wondering why you didn't respond more in depth to this thorough dismantling of your method.



A cubic transformation is merely a way to make poorly fitting data, more easily analyzed by applying a mathematical function to the data set.

Fine, but then why not explain why you need to make poorly fitting data more applicable to your method? Data is data. When you transform data you're altering it, and then it's not longer the data it was when you first analyzed it. So while your mathematical functions work better with transformed data set, you're no longer dealing with the real data.

Although this is a pretty obscure term among people with some mathematical background, it isn't a mystery among people who have been specifically educated to analyze data sets.

PH25 you are clearly an intelligent person with a reasonably strong knowledge of stats and data.....why would you go so far out of your way to discredit your own knowledge by predicting that most of our minor leaguers are going to be Koufax-esque and then either ignoring or shouting down anyone who disagrees?

Here you have a perfect opportunity to either defend your work against the plethora of academic questions that have been brought up in regards to it, and if you're not invested enough to defend it that way you can explain that the projections are not the most scientific and that you had a little fun analyzing the data and using different formulas to come to these projections. Instead you're focusing on the fact that cubic transformation is more common than Munson'sStache said it was, and that you shouldn't have to explain yourself because others occasionally make fun of you.

I don't post that much but I've read pretty much all of your posts and while I'll admit the projections can be fun to read about their basis in reality is something I think you're going to have to defend a lot more rigorously than your response above if you're really expecting people to take them seriously.

gdn
05-29-07, 12:06 PM
Just so we're clear - I didn't make my comment about "the first 3 or 4 links" to denigrate anyone here in this thread. It was merely an opinion that people on message boards can be childish at times.

Phobicsman
05-29-07, 12:35 PM
I think you guys are being a bit harsh on Hughes2.50.

He's gone to a lot of trouble to come up with some fantastic projections that say that we're going to have a kick-ass rotation for the next ten years that is going to have the Red Sox chewing our dust.

That's good news isn't it?

So, why argue with him?

Dow Jones
05-29-07, 12:59 PM
I think you guys are being a bit harsh on Hughes2.50.

He's gone to a lot of trouble to come up with some fantastic projections that say that we're going to have a kick-ass rotation for the next ten years that is going to have the Red Sox chewing our dust.

That's good news isn't it?

So, why argue with him?

I can't tell if you're being saracastic here or not.

For one, I find the projections to be so over the top that I have a hard time thinking of them in that much of a positive light because they're just so outrageous. That doesn't mean that I'm not excited about the prospect of these 3-4 guys coming through the system together and developing into a good if not great homegrown rotation in the near future. That's a wonderful prospect, but just because somebody makes up a flawed formula declaring that they'll be Koufax/Pedro/Clemens/RJ in their primes and first ballot hall of famers doesn't make it so.

That's why I'm arguing with him(can't speak for anybody else). If you were being sarcastic then please disregard my post. If you were not, I like projections to have a bit more basis in reality before I get too excited about them.

yankeebot
05-29-07, 01:05 PM
I can't tell if you're being saracastic here or not.

For one, I find the projections to be so over the top that I have a hard time thinking of them in that much of a positive light because they're just so outrageous. That doesn't mean that I'm not excited about the prospect of these 3-4 guys coming through the system together and developing into a good if not great homegrown rotation in the near future. That's a wonderful prospect, but just because somebody makes up a flawed formula declaring that they'll be Koufax/Pedro/Clemens/RJ in their primes and first ballot hall of famers doesn't make it so.

That's why I'm arguing with him(can't speak for anybody else). If you were being sarcastic then please disregard my post. If you were not, I like projections to have a bit more basis in reality before I get too excited about them. Phobicsman's been here over 2 years and has 16 posts. They all read similar to that one. Rah-rah.

Phobicsman
05-29-07, 03:47 PM
Phobicsman's been here over 2 years and has 16 posts. They all read similar to that one. Rah-rah.

Whats wrong with being 'rah rah' as you put it?

Better to be optimistic and trying to look on the bright side than coming up with sour cynical stuff all the time.

I just think Hughes2.50 has come up with some decent statistical analysis - don't see why you have to get abusive if someone wants to stick up for him.

captainlad
05-29-07, 04:38 PM
Because the point of objective statistical analysis is to allow models to be criticized, otherwise there would be no point to them.

yankeebot
05-29-07, 05:17 PM
Whats wrong with being 'rah rah' as you put it?

Better to be optimistic and trying to look on the bright side than coming up with sour cynical stuff all the time.

I just think Hughes2.50 has come up with some decent statistical analysis - don't see why you have to get abusive if someone wants to stick up for him. I didn't say there was anything wrong with it. Dow Jones wasn't sure whether or not to take your post as sarcasm. I was simply describing the kind of poster I perceive you to be. With the limited data at hand, I think I was spot on. I have no griipe with optimists.

Munson's 'Stash
05-29-07, 07:16 PM
I think you guys are being a bit harsh on Hughes2.50.

He's gone to a lot of trouble to come up with some fantastic projections that say that we're going to have a kick-ass rotation for the next ten years that is going to have the Red Sox chewing our dust.

That's good news isn't it?

So, why argue with him?

Not to pile on, but what if he's wrong?

The point of any projection system is to try to predict reality as accurately as possible. If your system causes you to over-value your own players due to excessively optomisitc projections what happens if they don't meet those expectations?

Consider a hypothetical example:

The Twins are willing to deal Johan Santana and come to the Yankees because they are one of the places he is willing to go and have talent the Twins are interested in (the Yankees young pitchers). Who are you willing to give up to get him, knowing that he is on the block and will sign an extension with the team he winds up with (so the Yankees will not not be able to just wait untill he hits FA)?

Santana has a career ERA+ of 143, tying him for 9th on the all time list. But H2.5 projects Hughes to outpitch him by a career ERA+ of 167, and has Humberto Sanchez just behind him at a career ERA+ of 142. He also thinks that Betances "has a better than 50-50 chance to have a career ERA+ of over 150." Similarly, Joba Chamberlain has the same chance to be over 130 by his estimation.

So who do you deal for Santana, if anyone? In the real world the Twins would undoubtedly want Hughes, and probably another of these pitching prospects. If H2.5 is correct Phil Huges is going to be the best pitcher in the history of baseball, and Humberto Sanchez is going to be more valuable to the Yankees than Santana (because of the number of years of arb control vs having near identical ERA+s). Similarly Betances and Chamberlain have good chances to be the second best pitcher in the history of baseball and in the top 30 of all time, respectively.

But let's consider a projection system with a better track record. PECOTA projects Santana to have 31.4 Warp over the next 5 years. It projects Hughes to produce 21.1 Warp and Sanchez to produce only 7.5 Warp in the same time period (and the low projection for Sanchez isn't out of line if think that his lack of durability may force him to the bullpen). (Betances and Chamberlain aren't even projected for so little actual data.)

Using these numbers instead the Yankees would come out ahead in a trade of Hughes and Sanchez for Santana (as measured by Warp). Looking closer, say the Yankees are willing to give up Santana for Sanchez, Bettances, and Chamberlain. By H2.5's analysis this is not a deal the Yankees should make (remember Sanchez = Santana, and both Bettances and Chamberlain have a chance to be as good or better than him). OTOH I really doubt there isn't an actual GM who wouldn't make that deal.

There are a lot more considerations that go into looking at a trade, obviously; but this should make it clear that improperly evaluating players (your own and others) leaves your team at an extreme disadvantage against teams with a better understanding of the reality of their player's potential. There is no reason that we as fans should be any less informed if we can help it, or embrace ignorance just because it makes us feel good.

yep
05-31-07, 01:12 PM
Excellent analysis by Munson's Stash. The long and short of it is that there is no "there" there when it comes to H2.5's methods.

There is no way to evaluate them, since no control data is shown and since there are a number of significant "black boxes" in his description. There's nothing inherently wrong with making up one's own projection system and then wishing to keep it secret, but when it produces such wildly optimistic results one has to expect that it is going to invite a certain amount of skepticism, if not outright ridicule.

The fact is that M.S. has done a really good job of not ridiculing or personally attacking H2.5, in spite of the fact that H2.5 seems hell-bent on making himself an easy target with the combination of dismissive arrogance, cryptic elitism, and offhand usage of terminology that sounds for all the world like big-word nonsense.

The bottom line is that, to an outside observer, H2.5's "projections" may as well be guesses, because there is simply no way to evaluate their validity. There is no way to test his methods and no historical track record.

JavyVazquezIsSick
06-01-07, 12:14 PM
Munson that was bang up work, and as soon as I have a spare day I'll actually read the whole thing.

Toaderly
06-02-07, 08:19 AM
Print it, stuff it into an envelope, and mail it to Cashman before he goes off on another panic move like Igawa.

Munson's 'Stash
06-03-07, 09:47 AM
I've been looking around the internet for examples of current studies of minor league pitchers that we might contrast to H2.5's methods. While I grant that it has not been an exhaustive search on my part the best one I have seen so far comes from the opposition, SoSH member (and former member here, IIRC) Eric Van:

His "Ultimate Pitching Prospect Stats Ranking" (http://sonsofsamhorn.net/index.php?showtopic=18340&st=0) thread in the SoSH minor league forums gives a very good list of all the current pitching prospects in the minors.

Note: for accurate reading, the first table in the thread is the current data. You have to go to post #19 to get the archive of the original post that started the discussion.

As a comparison to H2.5's compromised method, it does not include Philip Huges, as he's not likely to be sent down to the minors (except perhaps in a rehab capacity). It does give us a nice look at pitchers who are seen as legitimate prospects and how they stack up against the competition statistically and their age relative to the opposition.

Based on this info it would be nice to be able to compare how Yovani Gallardo and Will Inman did in H2.5's system, but I don't think he's likely to come forth with that data...

Hughes2.50
06-03-07, 12:36 PM
I've been looking around the internet for examples of current studies of minor league pitchers that we might contrast to H2.5's methods. While I grant that it has not been an exhaustive search on my part the best one I have seen so far comes from the opposition, SoSH member (and former member here, IIRC) Eric Van:

His "Ultimate Pitching Prospect Stats Ranking" (http://sonsofsamhorn.net/index.php?showtopic=18340&st=0) thread in the SoSH minor league forums gives a very good list of all the current pitching prospects in the minors.

Note: for accurate reading, the first table in the thread is the current data. You have to go to post #19 to get the archive of the original post that started the discussion.

As a comparison to H2.5's compromised method, it does not include Philip Huges, as he's not likely to be sent down to the minors (except perhaps in a rehab capacity). It does give us a nice look at pitchers who are seen as legitimate prospects and how they stack up against the competition statistically and their age relative to the opposition.

Based on this info it would be nice to be able to compare how Yovani Gallardo and Will Inman did in H2.5's system, but I don't think he's likely to come forth with that data...Although I can appreciate the work you went through typing up this Unabomber Manifestor, directed at me, I will not participate in a wild goose chase following around your wrong impressions. However, this last post allows for an opportunity to actually do what science intends to do, and that is - predict what a system will do. Given that, I would be more than happy to compare my system with Eric Van's. I'll start with a comparison of the two names you just mentioned and their performance this year so far (MDERA). <p>Gallardo's MDERA so far in 2007 is an incredible 222. So clearly, even Eric Van's formula which he himself says he pulled out of an orifice, and my system agree. Gallardo is having an impressive year - still we want to see how these numbers look at the end of the year. On the other hand, Inman's MDERA is a solid (especially for a pitcher as young as he is), if unspectacular 110. The MLE adjustments at Brevard didn't smile upon Gallardo last year either - so maybe the adjustments at Brevard might be too depressed. <p>In any event, Inman's numbers place him behind both Buchholz and Bowden on the MDERA statistic. How did I know you wanted to know that?<p>Btw, is Eric Van still on the employ of the Red Sox. If so, talk about compromised systems.
:D

Munson's 'Stash
06-03-07, 03:18 PM
Although I can appreciate the work you went through typing up this Unabomber Manifestor, directed at me, I will not participate in a wild goose chase following around your wrong impressions.

If I'm wrong, feel free to actually come up with some corrections. We' all love to hear them.


However, this last post allows for an opportunity to actually do what science intends to do, and that is - predict what a system will do.

Eric Van's system does not continue to making predictions as your's does though. EV's essentially shows what is going on now, correcting for factors that can obscure the data (age vs competition, etc). OTOH your 'system' goes on to make predictions for the future that you (so far) cannot support. Unfortunately, your number's agreement with Eric Van's don't really offer any validation to your work.



In any event, Inman's numbers place him behind both Buchholz and Bowden on the MDERA statistic. How did I know you wanted to know that?

I would imagine that since I've spent a pretty large word count trying to get you to release any "data" at all, it shouldn't be too hard for you to realize I'd like you to provide data for everyone on Eric Van's list on how they do running through your complete methodology (this would include their projected "Career Numbers" as well as how players you've done two years of computations for have had those guesses change).

The more data the better, but you've still provided so little that all your results don't really allow us to make any actual judgments.


Btw, is Eric Van still on the employ of the Red Sox. If so, talk about compromised systems.
:D

This is actually something you brought up earlier, which I think should be also be addressed; namely that sabermetric work done by people in the employ of MLB teams is somehow "compromised", or unobjective. There are several problems with this:

1. It has no basis no fact. Any good baseball analyst is going to be running their numbers in a rigorous manner, and not changing the numbers to fit with their team affiliation (see my post above relating to Phobicsman's question).

For a simple example, look at defensive stats: When MGL was employed by the Cardinals he didn't suddenly bump the numbers of Cardinals players (and I'm pretty sure that Walt Jockety didn't want him to). Now that he no longer works with the Cardinals and UZR numbers are available again, those of us who pay attention are delighted to be able to get them. The value in the work he did is still there.

2. Quite a bit of the best new work being done in sabermetrics is sponsored by individual MLB teams. They have the rescoures and incentive to invest in it. Even work they don't commission is often done with the hope of selling it to them (Dewan's Stats Inc and The Fielding Bible are good examples of this). I'm not sure what economic model you're addvocating, especially since...


3. It has been intimated that your own big reason for not releasing your methodology is that it might be "stolen." Just because your own work is not good enough to land you a job, it doesn't excuse you from making baseless accusations about the legitimate work of others, especially when you have 'black boxes' in your methods that seem to favor players by their team affiliation.


4. Even work that is private or proprietary has been vetted by other sabermetricians. Again, using UZR as an example, MGL's results and processes have been analyzed by others (Tango and Dolphin) who also have respected bodies of work (PECOTA is another good example). We're looking for that in what you write, which so far you have not provided.



Your system is "compromised" by nature of it's flawed assumptions and shoddy science, not by nature of who you happen to root for. Of course I could be wrong (as stated in the first post of this thread), but suspect that I'm going to be waiting a long time for you to actually prove anything.

Lucen
06-03-07, 04:05 PM
Although I can appreciate the work you went through typing up this Unabomber Manifestor, directed at me, I will not participate in a wild goose chase following around your wrong impressions.

Why is it that when someone wants to verify your methodology, every single time, no matter what their approach, you essentially accuse them of being on a witch hunt? You have not once, ever in your entire history of posting here regarding this projection system, faced down a valid criticism of your methods. What Munson is asking for here is completely legitimate and is written without the slightest hint of personal bias. What, of his questions, do you find so distasteful that you refuse to even acknowledge his concerns (and ours... meaning most of this community) about your work?

Seriously, explain to us what your problem is with running some blind historical data through your system to test its accuracy? What's so problematic about providing the projections for some of the other minor leaguers mentioned earlier in this thread? Why are you so hesitant to explain why a cubic transformation is important to your projections and how it affects them?

Why is it that several people who have credentials (likely far superior to your own) are still waiting for responses to their PMs? PMs which you promised would result in an open discussion of your methods for those with the proper credentials?

Whatever the reality of this situation, you are doing nothing but hurting your own credibilty by continuing to dismiss every dissenting opinion with either arrogant insults about their ability to comprehend your work or their lack of proper credentials to be privy to it in the first place. Sure, some people have attacked you personally with flame baiting. By all means, ignore them. But several people have attempted to offer you avenues to back up your claims and you've done nothing but scoff at them. All you have to do is approach these people seriously and respond to their criticisms and a lot of the heat surrounding your projection system would disappear.

And here's something else to consider. Finding flaws in your system isn't going to destroy your work, it can only improve it. No system is perfect, and none of the accepted projection systems in use today started off being as good as they are. All of them had to go through a testing period where peer review was used to improve them. If you are unwilling to go through this period, then most of us would ask that you simply do not use your projections as any kind of evidence while arguing in your posts here, as they will not have any validity.

Hughes2.50
06-03-07, 04:14 PM
If I'm wrong, feel free to actually come up with some corrections. We' all love to hear them.Oh, you are wrong. But just because you are doesn't mean I have to be sent on any of your wild goose chases.<p>I've told you before. If someone PM's me in good faith, I will answer questions about what I've written about so far in this forum.<p>What I haven't written about, or described in more detail, I hold in confidence, until, when and if I decide to provide it to a deliberative and formal process. A process that this form isn't. <p>Your presentation (this thread), doesn't follow scientific process. First, I haven't submitted my work for publication or formal analysis yet. Secondly when one does that, they don't do it in a public forum, where the lowest common denominator can squelch the deliberative process with false paths, or, worse yet, intentionally obstructive efforts.<p>Still, I have offered to respond to questions submitted about already described work by pm - if those requests are submitted in good faith. I don't consider your discussions to have met that mimimal standard. In part because you arrogantly have left up an avatar which I believe you intend to mock me with.

Hughes2.50
06-03-07, 04:21 PM
Why is it that several people who have credentials (likely far superior to your own) are still waiting for responses to their PMs? PMs which you promised would result in an open discussion of your methods for those with the proper credentials?Your comments are contrary to the truth in several areas. But lets stick with this simple statement you've made. The only person I have refused to answer is Stash by PM. Everyone else I've responded to, and can be verified by the administrators of this web site. So that is a lie.<p>Secondly, since Stash has himself admitted he isn't formally educated in the area of methods or statistics, and since he is the only one who I haven't responded to (and I said why that was in my last post to him), who are these other, unspecified people who have sent me PM's where I haven't responded? <p>You are mistaken, or, worse, your are making things up.

Lucen
06-03-07, 04:28 PM
Your comments are contrary to the truth in several areas. But lets stick with this simple statement you've made. The only person I have refused to answer is Stash by PM. Everyone else I've responded to, and can be verified by the administrators of this web site. So that is a lie.Secondly, since Stash has himself admitted he isn't formally educated in the area of methods or statistics, and since he is the only one who I haven't responded to (and I said why that was in my last post to him), who are these other, unspecified people who have sent me PM's where I haven't responded?
You are mistaken, or, worse, your are making things up.


Several posters have publically mentioned their credentials in threads you've been posting in, including this one, and said they've not received responses from you when they sent PMs.

For example...


Wait, so are you fitting a cubic polynomial to the data? Is that what you mean?

That might explain some of the more extreme projections, because while polynomials can be made to fit data pretty well, they tend to "explode" outside of the main data sample and could thus lead to some pretty extreme forecasts on inputs that are out of the normal range. Which could happen if we are taking small samples of minor league domination.

Or are you really doing some sort of complex Fourier-type transform? I found this paper (http://www.math.northwestern.edu/%7Egeorge/preprints/sumtransf/sumtrans.pdf) (PDF), for example, that gives one cubic transform forumla (top of page 11 of the pdf). And to be fair to Hughes, a google search for "cubic transform" gives lots of legitimate hits.

To be fair to the rest of us, however, that looks like nonsense. And I have studied math, applied math, statistics, and econometrics for over 10 years. (I have a phd in econ from a good school, and have always done applied, data-driven work.) Even if you can't give us your precise formula (and I can totally understand you not wanting to, much as BP doesn't give out its formulas for PECOTA), perhaps you can explain intuitively what the cubic transform does?

Is it smoothing out data the way predicted values from a regression would (like fitting the data to a pre-defined functional form, such as a cubic polynomial)? That's something we can all understand, even if the formula/implementation is not something we're going to follow ourselves. Or is it something else entirely?


So am I, but I'm trying to give him the benefit of the doubt. I only stated that I have a phd in econ because H2.50 keeps implying that if someone had studied enough, they'd know what he's talking about. Maybe that's true, but I have studied a lot of math and statistics, and I do not follow what he's saying. That's not to say that he's necessarily doing something wrong, but I would like at least an intuitive explanation of what the (now infamous) cubic transform is doing.



You haven't answered him here, and apparently you haven't done so in an PM either. So no, I'm not making things up and I'm not wrong, you're ignoring valid criticisms and questions from qualified individuals. He is just one example of several people who appear to be very well qualified to critique your work.

Edit: And this is a great example of what you've been doing. You ignore most of a post and pick out one little thing you take exception to as if it was the only thing said. This doesn't help your credibility.

Hughes2.50
06-03-07, 04:36 PM
Several posters have publically mentioned their credentials in threads you've been posting in, including this one, and said they've not received responses from you when they sent PMs.

For example...







You haven't answered him here, and apparently you haven't done so in an PM either. So no, I'm not making things up and I'm not wrong, you're ignoring valid criticisms and questions from qualified individuals. He is just one example of several people who appear to be very well qualified to critique your work.

Edit: And this is a great example of what you've been doing. You ignore most of a post and pick out one little thing you take exception to as if it was the only thing said. This doesn't help your credibility.You are misinformed. I received no PM from him. And why should I believe someone's credentials on an anonymous forum when they can't take the time to contact me by pm so that if they are serous I can verify who they claim to be. Why the reluctance to do that? Is it easier to take pot shots standing behind the veil of anonymity? Of course it is. Look, Stash could be Eric Van for all I know. Until, he or someone else actually puts themselves in a position to be taken seriously, I will continue to point out how this method used on this page is nothing more substantive than a Kangoroo Court.

Lucen
06-03-07, 04:41 PM
You are misinformed. I received no PM from him. And why should I believe someone's credentials on an anonymous forum when they can't take the time to contact me by pm so that if they are serous I can verify who they claim to be. Why the reluctance to do that? Is it easier to take pot shots standing behind the veil of anonymity? Of course it is. Look, Stash could be Eric Van for all I know. Until, he or someone else actually puts themselves in a position to be taken seriously, I will continue to point out how this method used on this page is nothing more substantive than a Kangoroo Court.

If you want to post the results of this system here and be taken seriously, you might want to put a little more effort in. ComeToNYRandy posted a polite and unbiased question here that, as far as I can tell, doesn't risk exposing the secrets of your system. Why not answer him? There is no damage done to you by doing so, and it would only serve to increase the validity of your claims. Heck, send him a PM and talk about it in private so he can come here and at least say that you've addressed his concerns to his satisfaction.

Your reluctance about responding to questions regarding your projection system is bordering on paranoia... If you're not willing to entertain anyone's questions about the system, can you at least stop using your results in your posts as if they prove anything?

As for Munson's post that started this thread, have you actually read it? He makes some very good points about your system and your posts surrounding it. If your only reason for not responding to them is that he's been a little mean to you in the past, you're really hurting your credibility here.

Hughes2.50
06-03-07, 04:46 PM
... If your only reason for not responding to them is that he's been a little mean to you in the past, you're really hurting your credibility here.Here, or at SOSH?<p>LOL

Hughes2.50
06-03-07, 04:48 PM
You haven't answered him here, and apparently you haven't done so in an PM either.Of course I haven't answered them in PM - there weren't any PM's to answer, contrary to the false impression you are valiantly trying to leave.

Lucen
06-03-07, 04:58 PM
Here, or at SOSH?LOL


What does SoSH have to do with anything? And you're hurting your credibility here... quite badly. Otherwise you wouldn't have so many people pointing out problems with your system and your posts.




Of course I haven't answered them in PM - there weren't any PM's to answer, contrary to the false impression you are valiantly trying to leave.

Try reading my whole post. I clearly said that if you wanted to lessen the constant criticism of you and your system, you might want to put a little effort in and PM him about his question. That way you can keep your discussions private but still address the concerns raised. Also, what possible reason do you have for refusing to answer that question publically? Seems perfectly legitimate and polite to me.

Or are you going to throw out that line about how this is the internet and you can't be sure he's actually a PhD, so he's not worth the time it would take to respond? If you're going to continue to hide behind that particular curtain, why offer the PM challange at all? It's the internet, so how are you ever going to be sure that ANYONE who approaches you is who they say they are. I could easily look up a professor from a college near here with a PhD in mathmatics, claim to be him and create an email address to back up that claim. If you're this paranoid about talking to anyone, will anyone PMing you ever get a response? I'm going to venture a guess and say no.

Hughes2.50
06-03-07, 05:07 PM
It's the internet, so how are you ever going to be sure that ANYONE who approaches you is who they say they are. I could easily look up a professor from a college near here with a PhD in mathmatics, claim to be him and create an email address to back up that claim. If you're this paranoid about talking to anyone, will anyone PMing you ever get a response? I'm going to venture a guess and say no.I've already responded to several PM's without asking for verification for who someone is. I've also described when I would ask for someone to verify who they are if they want some information, and under what context. If you can't imagine how something could be setup first using PM's (to maintain privacy) and then followed up in another fashion to verify someone, well, then you are probably someone who won't be able to get much information. Sorry.

Lucen
06-03-07, 05:16 PM
I've already responded to several PM's without asking for verification for who someone is. I've also described when I would ask for someone to verify who they are if they want some information, and under what context. If you can't imagine how something could be setup first using PM's (to maintain privacy) and then followed up in another fashion to verify someone, well, then you are probably someone who won't be able to get much information. Sorry.

Then why hasn't anyone stood up and backed you up, saying that you've taken the time to address their concerns? And again, what about ComeToNYRandy's question is keeping you from responding to him publically? If it's that he hasn't taken the time to PM you then jump through your hoops to prove to you he's worth your time, then you're just fueling the sentiment that you are arrogant and not worth taking seriously.

You seem to take offense at the mere possibility that someone disagrees with you or doubts you. You won't get anywhere in the academnic arena unless you change that, and quickly. There is just as much gained from proving a theory or system wrong as there is from proving it right. In fact, most research ends up in the realm of the former, not the latter. But you seem to refuse to acknowledge the possibility that anything could be wrong with your projection system.

That's not scientific, it's an excersise in ego stroking.

And again... If you refuse to face down criticism regarding your system, will you at least stop using your results in posts as if they prove anything?

Hughes2.50
06-03-07, 05:23 PM
And again... If you refuse to face down criticism regarding your system, will you at least stop using your results in posts as if they prove anything?Any question that is based on what I've already written in this forum, is certainly a valid area to question. Anyone who wants to get an answer to such questions can ask them so that the question evokes a simple response. I will not respond to interpretations, which most of this thread is actually all about. <p>If someone is really interested in getting answers beyond what I've written in this forum, they can PM me.<p>And, some posters here, have actually taken the time to understand how to do the number crunching. I'd advise you and your friends over at SOSH to do the same.

Lucen
06-03-07, 05:30 PM
Any question that is based on what I've already written in this forum, is certainly a valid area to question. Anyone who wants to get an answer to such questions can ask them so that the question evokes a simple response. I will not respond to interpretations, which most of this thread is actually all about. If someone is really interested in getting answers beyond what I've written in this forum, they can PM me.
And, some posters here, have actually taken the time to understand how to do the number crunching. I'd advise you and your friends over at SOSH to do the same.


Stop bringing SoSH into this, it has nothing to do with this. The fact that I post in both places does not have any bearing on what I'm saying here, what so ever. This is just another attempt to deflect valid criticism without actually facing it.

You've proven here that you will never actually respond to valid concerns people have with your work. You just dance around people's posts and either insult them or accuse them of alterior motives. I'm done trying to get a response from you in a civil discussion. But thank you for proving to anyone reading that you're not worth paying attention to, so long as you continue to operate this way.

Munson's Stash, thank you for the work you've done in detailing his system and pointing out the problem areas. I hope that it shows the rest of this community why they shouldn't put any weight into his projections and keeps people from getting overly excited about some of the team's prospects.

Hughes2.50
06-03-07, 05:35 PM
Stop bringing SoSH into this, it has nothing to do with this. The fact that I post in both places does not have any bearing on what I'm saying here, what so ever. This is just another attempt to deflect valid criticism without actually facing it.

You've proven here that you will never actually respond to valid concerns people have with your work. You just dance around people's posts and either insult them or accuse them of alterior motives. I'm done trying to get a response from you in a civil discussion. But thank you for proving to anyone reading that you're not worth paying attention to, so long as you continue to operate this way.

Munson's Stash, thank you for the work you've done in detailing his system and pointing out the problem areas. I hope that it shows the rest of this community why they shouldn't put any weight into his projections and keeps people from getting overly excited about some of the team's prospects.You've done a fine job of demonstrating how not to get answers.<p>I actually believe that you and the others posting here who have dual citizenship at SOSH aren't interested in answers. You appear to be interested in something quite different, or you would take simple, already described, steps to prove otherwise.<p>I think after seeing this, those who aren't here with (alterior?) ulterior motives will realize that simple, good faith requests will be answered.

Lucen
06-03-07, 05:39 PM
You've done a fine job of demonstrating how not to get answers.I actually believe that you and the others posting here who have dual citizenship at SOSH aren't interested in answers. You appear to be interested in something quite different, or you would take simple steps to prove otherwise.


Truth be told, I don't think you're for real. I believe you're either a Sox fan or a Mets fan here to make this community look bad. Fortunately, most of the people here seem to see right through you. My interest is mostly in exposing you for what I believe you to be. You've taken quite a few steps to help me along there, though. So thank you.

As I said in my last post, I'm done discussing this with you. So get the last word in if you like, but don't expect me to continue dancing in circles with you.

homer2931
06-03-07, 05:41 PM
H2.50
Why don't you run your system on the BA 2000 top 100 prospects list, and then let people here analyze it? Certianly giving the results of players from 7 years ago can't hurt.

Hughes2.50
06-03-07, 05:46 PM
H2.50
Why don't you run your system on the BA 2000 top 100 prospects list, and then let people here analyze it? Certianly giving the results of players from 7 years ago can't hurt.I will probably do that. Stash pointed out Eric Van's system, and my MDERA values can be compared to several of Eric Van numbers. In fact I posted two already. I'm not going to run those numbers right now, I want to see where this thread is going before I do that.

Hughes2.50
06-03-07, 05:52 PM
Truth be told, I don't think you're for real. I believe you're either a Sox fan or a Mets fan here to make this community look bad. Fortunately, most of the people here seem to see right through you. My interest is mostly in exposing you for what I believe you to be. You've taken quite a few steps to help me along there, though. So thank you.

As I said in my last post, I'm done discussing this with you. So get the last word in if you like, but don't expect me to continue dancing in circles with you.I assure you I am a diehard Yankee fan. Your attributions for who or what I am tend to be looked at unfavorably on this board. Your own stated motives reveal a great deal about what you are doing on this board.<p>Thank you for protecting Yankee fans. :P

Lucen
06-03-07, 06:03 PM
I assure you I am a diehard Yankee fan. Your attributions for who or what I am tend to be looked at unfavorably on this board. Your own stated motives reveal a great deal about what you are doing on this board.Thank you for protecting Yankee fans. :P


Until you PM me with credentials proving you are, in fact a Yankees fan, I will continue to operate under the assumption you are a plant.

((I know I said I wouldn't respond here again, but I couldn't resist. :) I'm done now, I promise.))

Munson's 'Stash
06-03-07, 06:12 PM
Oh, you are wrong. But just because you are doesn't mean I have to be sent on any of your wild goose chases.

Proving that you are, in fact, correct (or even have a point) is a "wild goose chase?"


What I haven't written about, or described in more detail, I hold in confidence, until, when and if I decide to provide it to a deliberative and formal process. A process that this form isn't.

Can you even describe what this process might be? Part of my point is that there is no formal process as you describe it. There is no reason for this forum to not serve this purpose.




Your presentation (this thread), doesn't follow scientific process. First, I haven't submitted my work for publication or formal analysis yet. Secondly when one does that, they don't do it in a public forum, where the lowest common denominator can squelch the deliberative process with false paths, or, worse yet, intentionally obstructive efforts.

This hasn't stopped you from quoting the results and suggesting that your research is valid. If the process is incomplete, and hasn't been vetted how can this be? Why are questions about your process out of bounds?

As far as the "lowest common denominator," the only person squelching discussion is you. The discussion in this thread has been very well informed and reasonable.



Still, I have offered to respond to questions submitted about already described work by pm - if those requests are submitted in good faith. I don't consider your discussions to have met that mimimal standard. In part because you arrogantly have left up an avatar which I believe you intend to mock me with.

It is not there to mock you, whatever you might think. My offer to change it if you are willing to explain your methods still stands.

Hughes2.50
06-03-07, 06:17 PM
Proving that you are, in fact, correct (or even have a point) is a "wild goose chase?"



Can you even describe what this process might be? Part of my point is that there is no formal process as you describe it. There is no reason for this forum to not serve this purpose.





This hasn't stopped you from quoting the results and suggesting that your research is valid. If the process is incomplete, and hasn't been vetted how can this be? Why are questions about your process out of bounds?

As far as the "lowest common denominator," the only person squelching discussion is you. The discussion in this thread has been very well informed and reasonable.




It is not there to mock you, whatever you might think. My offer to change it if you are willing to explain your methods still stands.Did it ever occur to your obtuse consciousness that a simple question would get an answer, if in fact I've written about it here? All I see on this thread are interpretations, presented as analysis. <p>A simple solution but one you haven't hit on yet is to ask me to answer simple questions, you know one step at a time kind of questions. Instead what I see are rhetorical flourishes intended to leave people with an impression without ever getting answers to many questions which could have been asked, and received answers for, without ever resorting to the dreaded PM.

yep
06-03-07, 08:17 PM
Did it ever occur to your obtuse consciousness that a simple question would get an answer, if in fact I've written about it here?...
Hughes2.50, everything Munson's 'Stash has posted in this thread has been reasonable and civil, and to my reading, perfectly simple and straightforward in terms of his questions and critiques. He has gone out of his way at every turn to keep things aboveboard and dispassionate.

Nevertheless, your responses are consistently vitriolic, hostile, even downright abusive. Nothing that 'Stash has said nor any aspect of the way he has said is even remotely "obtuse."

Surely, you must agree that some of the projections you have posted are out of step with what more mainstream prognosticators have offered. Surely you must understand that some people are going to be skeptical of a system that they can neither test nor measure, especially when it produces forecasts that seem too good to be true.

I know that a lot of people have attacked your projections and mocked you and your methods, but Munson's Stash has been very reasonable in expressing his skepticism. It does no favors to your credibility to lash out at him for it.

Munson's 'Stash has offered you a rare opportunity in the often cheap-shot world of web-forum disputes. He has presented a thorough and plainly-stated list of concerns and questions. You could, with a single post, respond to and dismantle all of them, and potentially shut up a lot of people even without revealing the mathematical particulars.

The fact that you choose not to respond directly to the meat and substance of his criticisms but instead to insult and belittle him and to pick at incidental phrases and to insinuate that he has stupidly failed to set up the questions in the correct format for you to respond suggests that your Emperor indeed has no clothes.

Munson's 'Stash
06-05-07, 01:12 PM
I'd also like to add that this is the Sabermetrics Forum, not ItL (or "HA"). I assume that the reason it is here is to answer and discuss sabermetrics with a more serious eye and rigorous approach.

WHIP
06-16-07, 08:56 AM
I'd also like to add that this is the Sabermetrics Forum, not ItL (or "HA"). I assume that the reason it is here is to answer and discuss sabermetrics with a more serious eye and rigorous approach.

I read most of what you wrote, dude, and it's impressive. But you wasted your time - it's obvious that this guy's a quack and not interested in any sort of legitimate analysis. He's like Duane Gish or Heber Jentzsch. Come up with a silly solution that's only tangentially relevant to reality and preach it as if it's gospel.

The fact that he arbitrarily designates a league average ERA at 4.50 should tell you all that you need to know - much like ESPN with their Player Ratings, he has no clue as to what he's doing. Any statistician or, well, critical thinker, should immediately see the problem with that approach.

The dialogue for which you're hoping will never come to pass, mainly because he's never been interested in having one.

Hughes2.50
06-16-07, 05:06 PM
I read most of what you wrote, dude, and it's impressive. But you wasted your time - it's obvious that this guy's a quack and not interested in any sort of legitimate analysis. He's like Duane Gish or Heber Jentzsch. Come up with a silly solution that's only tangentially relevant to reality and preach it as if it's gospel.

The fact that he arbitrarily designates a league average ERA at 4.50 should tell you all that you need to know - much like ESPN with their Player Ratings, he has no clue as to what he's doing. Any statistician or, well, critical thinker, should immediately see the problem with that approach.

The dialogue for which you're hoping will never come to pass, mainly because he's never been interested in having one.If you had any knowledge of how sabermetricians translate minor league equivalencies you'd know such a calibration is pretty standard fare. After the minor league adjustments have been made, it isn't unusual to set average major league pitching at a league average 4.50 (In MDERA terms that would be equal to 100). That is so straight forward you can even read about it publically on Baseball Prospectus website. Below are Pecota's explanation for their PERA a statistic similar to but not quite the same as my MDERA statistic, which I present in a similar fashion as an ERA+ statistic, while Pecota presents their PERA in ERA terms.<p>http://www.baseballprospectus.com/glossary/index.php?context=6&category=true


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 <i>PERA is calibrated to an ideal league with an average PERA of </i><b>4.50.</b>Italics and bold mine for emphasis. <p>Unfortunately your tone (bordering on violating the CS by calling me a quack), and your lack of fundamental knowledge about standard sabermetrics procedures, presents an unpalatable environment for me to respond within. As I told your idol Stash, all you have to do is ask simple questions and I'll give you answers. Today, I thought I'd correct your post because sometimes enough is enough.

WHIP
06-17-07, 10:18 AM
:lol:

Munson already explained why you're wrong. Not every minor league venue is equal and you're assuming that they are. We're not talking about PERA. We're talking about your assumption that all minor league ERAs average 4.50. He already illustrated how you run into inconsistencies when you make assumptions like this. That's about as fundamental a problem as you can possibly get in your little formula.

Why haven't you responded to him directly? Because his questions aren't simplistic enough? No, I know the real reason - it's because you cannot defend your sham of a system against his arguments. You know this, yet you continue to argue to the contrary until you're blue in the face. Like I said, Duane Gish. I can tell you a million times that there's an overwhelming amount of evidence for evolution, and I can cite all of this evidence to you, and you'll keep repeating the same tired slogan of "there are gaps in the fossil record." Different characters, same story.

Predictably you pretend that I'm the one who constructed an "unpalatable environment" when it's been that way from the start because of your input. Munson went to great lengths to explain what type of civil dialog he wanted, and you spit on his efforts, alleging that this was some sort of "witch hunt." So why should I treat you with any sort of civility? Everyone here already knows you're a disgrace. I'm merely articulating their beliefs.

Here, I'll ask you some simple questions if it'll make you feel any better:

How do you think you did on the SAT last month?

What colleges have you applied for? What's your safety school?

What's your opinion on the War in Iraq?

Hughes2.50
06-17-07, 10:38 AM
:lol:

Munson already explained why you're wrong. Not every minor league venue is equal and you're assuming that they are. We're not talking about PERA. We're talking about your assumption that all minor league ERAs average 4.50. He already illustrated how you run into inconsistencies when you make assumptions like this. That's about as fundamental a problem as you can possibly get in your little formula.

Why haven't you responded to him directly? Because his questions aren't simplistic enough? No, I know the real reason - it's because you cannot defend your sham of a system against his arguments. You know this, yet you continue to argue to the contrary until you're blue in the face. Like I said, Duane Gish. I can tell you a million times that there's an overwhelming amount of evidence for evolution, and I can cite all of this evidence to you, and you'll keep repeating the same tired slogan of "there are gaps in the fossil record." Different characters, same story.

Predictably you pretend that I'm the one who constructed an "unpalatable environment" when it's been that way from the start because of your input. Munson went to great lengths to explain what type of civil dialog he wanted, and you spit on his efforts, alleging that this was some sort of "witch hunt." So why should I treat you with any sort of civility? Everyone here already knows you're a disgrace. I'm merely articulating their beliefs.

Here, I'll ask you some simple questions if it'll make you feel any better:

How do you think you did on the SAT last month?

What colleges have you applied for? What's your safety school?

What's your opinion on the War in Iraq?While you are pretending to know what you are talking about, did you take the free tutorial suggestion to examine the calibration of ERA centering (noting how Baseball Prospectus is but one example of analysts who use this very same technique)?<p> Forget how SAT scores have been centered over the years, dilletante, did you in fact learn about how mle adjustments are made with a calibration for major league averages to equate with 4.50?<p>The problem with this thread is that you, and several others are acting as an echo chamber for your ignorance, all the while presenting a fatuous case that I won't engage in a 'serious' dialogue with you - and then arrogantly claiming I don't know what I am doing.

gdn
06-17-07, 10:55 AM
Let's keep it civil, folks :)

Hughes2.50
06-17-07, 10:59 AM
Let's keep it civil, folks :)Where were you when name-calling was happening (quack)?

gdn
06-17-07, 11:07 AM
Where were you when name-calling was happening (quack)?Quack? As in crazy? Or quack as in a duck?

I'm just sayin'; hate to see anyone ROd.

Edit: Nevermind, I got it.

Look, I'm just saying that if you feel people are name-calling, don't resort to the same thing. We're all guilty of it, yes, but sometimes it helps if someone not involved reminds you to step back. God knows how many times people have had to tell me that.

WHIP
06-17-07, 12:45 PM
While you are pretending to know what you are talking about, did you take the free tutorial suggestion to examine the calibration of ERA centering (noting how Baseball Prospectus is but one example of analysts who use this very same technique)?
Forget how SAT scores have been centered over the years, dilletante, did you in fact learn about how mle adjustments are made with a calibration for major league averages to equate with 4.50?The problem with this thread is that you, and several others are acting as an echo chamber for your ignorance, all the while presenting a fatuous case that I won't engage in a 'serious' dialogue with you - and then arrogantly claiming I don't know what I am doing.

I didn't mention SAT scores in general - I asked how you think you did on the exam last month. That's a simple enough question, and I'd really like you to answer it.

But your vocabulary does impress me. I bet you got at least a 550 on the Verbal.

Munson's 'Stash
09-03-07, 11:15 AM
I'm bumping this for Carbon Fiber (in the other thread).

I also plan on following up with a comparative analysis after the season ends of H2.5's predictions vs actual results, hopefully with an eye to understanding not only where he may have gone wrong, but why pitchers are so hard to project in the first place.

Carbon Fiber
09-04-07, 06:25 PM
I'm bumping this for Carbon Fiber (in the other thread).

I also plan on following up with a comparative analysis after the season ends of H2.5's predictions vs actual results, hopefully with an eye to understanding not only where he may have gone wrong, but why pitchers are so hard to project in the first place.

Thanks Munson's Stash ---

On another note, I found a site where I can download prior year statistics in delimited file format (i.e CSV) --- so when I get an afternoon I'm going to load them into a SQL table so I can start analyzing/querying away. I wish I could find a raw data feed of everything available (i.e. MLB must store that pitch by pitch data somewhere) --- the challenge with the basic query data on most sites is that I'd have to pull it down via scraping in order to get it into a useable, easily queryable format.

flymick24
10-20-07, 04:12 PM
there's really a simple solution to all of this

just ignore all of hughes2.50's posts until he proves that he can back up his projections with actual evidence that he's willing to share with the general public

oh wait, he doesn't like "witch hunts," right? darn....

ma-gaga
10-23-07, 12:06 AM
It's too bad. Munson was attempting to get some honest discussion going, and it was quashed.

Bouton_Howe
10-23-07, 12:37 PM
It's too bad. Munson was attempting to get some honest discussion going, and it was quashed.

H2.50, are we going to get a postseason review? Or are you waiting till spring training? Thanks!

CallOfTheCrow
10-24-07, 05:30 PM
How accurate is all this pecota, mle, numbers stuff? Everytime I see someone saying "so & so should do this because the mle says so" I don't even bother reading the rest of the post.

tangotiger
11-07-07, 02:55 PM
I'm very late to the party, but I just wanted to point out the difference between DICE, FIP, and that constant. Clay and I both use the 13/3/-2 scheme. The differences is whether you include IBB or not, and include HBP or not. Including or excluding these numbers will necessitate changing the constant (that 3.00 or 3.20 or whatnot) to calibrate to the league level. And also note that since 1993 or so, that constant has changed by a bit.

The short rule is: apply the league numbers to the IP, HR, BB, SO (and IBB, HBP if you have them) in the equation, compare to the league ERA, and you get your constant.

***

As for the "4.50" that is being discussed, it sounds like Hughes is saying that he does use the league ERA to compare against, and then adjusts it to a league ERA of 4.50. (That is, he indexes the ERA.) There's nothing wrong with this at all.

Munson however is saying that Hughes doesn't do this, that he actually takes the player's ERA and compares it to an ERA of 4.50, without looking at the league ERA at all, which of course would be a preposterous thing to do. But Hughes later comments in this thread makes it seem that he is doing that part the right way: ERA of 3.0 in a league of 6.0, means an indexed ERA of 2.25 (in a league of 4.50).

***

The question to ask of forecasts, and this is not just limited to Hughes but to all forecasts: what is the uncertainty level? You can forecast someone with a mean ERA+ of 150, but if your uncertainty level is +/- 75 95% of the time, then that doesn't really help, does it?

***

I have a general complaint with ERA+ in that it's the only stat that puts the context, the index, in the numerator, rather than denominator. If you figure SLG+, that'd be SLG divided by lgSLG. RC+ would be RC divided by lgRC, etc, etc. ERA on the other hand is lgERA divided by ERA. The problem with this approach is that it is not additive.

If you need convincing, it's highlighted in post 178 here:
http://www.baseball-reference.com/blog/suggestions#comment-48643

Giuseppe Franco
11-07-07, 05:15 PM
Leave. Hughes. Alone. Please!!!

hardrain
11-07-07, 05:26 PM
Sure is nice to see Tom Tango posting here.

tangotiger
11-07-07, 07:50 PM
Thanks Phil.

***

I have no idea what Franco is talking about.

Hughes2.50
11-07-07, 08:19 PM
I'm very late to the party, but I just wanted to point out the difference between DICE, FIP, and that constant. Clay and I both use the 13/3/-2 scheme. The differences is whether you include IBB or not, and include HBP or not. Including or excluding these numbers will necessitate changing the constant (that 3.00 or 3.20 or whatnot) to calibrate to the league level. And also note that since 1993 or so, that constant has changed by a bit.Hi Tom. It turns out that I have been using Clay's basic formula for the defense independent stats.


The short rule is: apply the league numbers to the IP, HR, BB, SO (and IBB, HBP if you have them) in the equation, compare to the league ERA, and you get your constant.

***

As for the "4.50" that is being discussed, it sounds like Hughes is saying that he does use the league ERA to compare against, and then adjusts it to a league ERA of 4.50. (That is, he indexes the ERA.) There's nothing wrong with this at all.That is correct, as you know that adjustement implies a standardization, which allows for comparisons among pitchers to a league average ERA of 4.50 and ERA+ of 100.


Munson however is saying that Hughes doesn't do this, that he actually takes the player's ERA and compares it to an ERA of 4.50, without looking at the league ERA at all, which of course would be a preposterous thing to do. But Hughes later comments in this thread makes it seem that he is doing that part the right way: ERA of 3.0 in a league of 6.0, means an indexed ERA of 2.25 (in a league of 4.50).

***I never said I did what Munson claimed. But then, he claimed so many things I never said, I realized that it was futile to even defend against one straw man.


The question to ask of forecasts, and this is not just limited to Hughes but to all forecasts: what is the uncertainty level? You can forecast someone with a mean ERA+ of 150, but if your uncertainty level is +/- 75 95% of the time, then that doesn't really help, does it?

***This of course is the crux of the problem, and where some critics have at least an intuitive appreciation for the problem. Yes, players who have never played in the majors or who are young, can still fail in so many ways along the way. Even players with some track record can still find their careers derailed or even cut short. Still the projections are useful because they are meant to project the next year, forward. Because we are only projecting one year forward with the help of MLE's it is not unreasonable to make an estimate and to expect that all things considered, the trajectory appears to cut a path where the estimate falls.


I have a general complaint with ERA+ in that it's the only stat that puts the context, the index, in the numerator, rather than denominator. If you figure SLG+, that'd be SLG divided by lgSLG. RC+ would be RC divided by lgRC, etc, etc. ERA on the other hand is lgERA divided by ERA. The problem with this approach is that it is not additive.I see your point, but I believe that consideration is offset by the fact that one can use ERA+ and CR+ to directly calculate winning percentage via the pythagorean theorem.

For example if we assume that Team A has an ERA+ of 140 (historically very high, but done in the past by a few Braves teams in the 90s), along with a CR+ of say 130 (what the Yankees of the last two years have averaged) - such a team would be expected to win 75+% of its games during the regular season.

Tom, as a test of this hypothesis (I've already done the analysis)) if you take Palmer's 2006 ESPN data for pitchers and multiply the ERA+/100 times the Runs Support/100 you will get a value that you can plug into the pyth theorem. That value is a robust predictor of winning percentages for pitchers.


If you need convincing, it's highlighted in post 178 here:
http://www.baseball-reference.com/blog/suggestions#comment-48643

tangotiger
11-07-07, 10:16 PM
I agree that it is irrelevant whether you put the ERA+ as 200 or 50, as long as you know what you are doing.

However, I've seen many people "average" an ERA+ of 150 and 75 to get an average that is not 1, when the average IS 1; it's not 112.5.

ERA+ of 150: ERA of 3.00 in a league of 4.50
ERA+ of 75: ERA of 6.00 in a league of 4.50

The average is obviously ERA of 4.50 in a league of 4.50. i.e., ERA+ of 100.

Hughes2.50
11-07-07, 10:25 PM
I agree that it is irrelevant whether you put the ERA+ as 200 or 50, as long as you know what you are doing.

However, I've seen many people "average" an ERA+ of 150 and 75 to get an average that is not 1, when the average IS 1; it's not 112.5.

ERA+ of 150: ERA of 3.00 in a league of 4.50
ERA+ of 75: ERA of 6.00 in a league of 4.50

The average is obviously ERA of 4.50 in a league of 4.50. i.e., ERA+ of 100. Yes. However, by keeping ERA+ as a it is, you can do the math manipulation I suggested in my last post, which yields very useful information about contributions to winning percentage.

tangotiger
11-07-07, 10:58 PM
Sure, but you can still achieve it by dividing.

The wrong ERA+ of 140 matches the right ERA+ of 71.4.

Instead of multiply 1.30 *1.40, you can do 1.30/.714, which gives you the same result.

I see no reason to show the "bigger is better" form, if smart people still fall into the trap I noted earlier.

Hughes2.50
11-08-07, 06:23 AM
Sure, but you can still achieve it by dividing.

The wrong ERA+ of 140 matches the right ERA+ of 71.4.

Instead of multiply 1.30 *1.40, you can do 1.30/.714, which gives you the same result.

I see no reason to show the "bigger is better" form, if smart people still fall into the trap I noted earlier.Tom since one of the major points of using ERA+ and CR+ is to estimate winning chances, you can easily see that if your form is used everyone would have to divide CR+ by ERA+ to begin to do the pythagorean calculation. Using the conventional statistic ERA+ instead, it doesn't matter what variable you multiply times the other (ERA+ * CR+ <br>or</b> CR+ times ERA+), however using your idea the calculation is contingent on CR+ being divided by ERA+. For example using your method you do get the correct answer by dividing CR+ by ERA+ with 1.30/.714 = 1.82, but not .714/1.30 = .55. While with the conventional use of ERA+ you get 1.4 * 1.3 = 1.82 and the opposite also leads to the same answer. <p> For the above reason and others the use of these statistics can yield useful information related to winning contributions, and doesn't require a wholesale change in how people perceive 'good' performances whether among pitchers or position players where contributions to winning are concerned.

tangotiger
11-08-07, 09:02 AM
Pythag is Runs Scored divided by Runs Allowed at its core.

This is analogous to RC+/ERA+ (if ERA+ is the correct form).

I have no worries about people dividing ERA+ by RC+. This will never happen. I do have worries about averaging the wrong ERA+. This has already happened, with knowledgeable people.

YMMV

Hughes2.50
11-08-07, 09:12 AM
Pythag is Runs Scored divided by Runs Allowed at its core.

This is analogous to RC+/ERA+ (if ERA+ is the correct form).

I have no worries about people dividing ERA+ by RC+. This will never happen. I do have worries about averaging the wrong ERA+. This has already happened, with knowledgeable people.

YMMVCorrect. But I think that intuitively it makes sense to say that a better pitcher is someone who scores higher on a dimension than a lesser pitcher. Manipulating the numbers gets you to the same point, but a quick estimate when looking at the expected winning percentage of a team is to find the CR+ and the ERA+ and then do as I suggested on the earlier post.<p>On the other hand, all things being equal if we are comparing two teams, or two pitchers on two teams we would first make the estimate for Team (Pitcher) A (say 1.3 * 1.4 = 1.82), then divide that number by an opposed team or pitcher B (say 1.1 * 1.2 = 1.32) giving the expected performance for team or pitcher A against team or pitcher B as 1.82/1.32 = 1.38 which leads to a pyth % for pitcher or team A playing pitcher or team B = 66%. <p>The above process doesn't require us to remember to change the current convention that better pitchers or teams with better pitchers have higher ERA+ scores than teams with lower ERA+ scores.<p>BTW, when knowledgeable people use ERA+ they can double check the accuracy of their calculations and whether the results are non-sensical by converting the ERA+ to the correct ERA and seeing if it is accurate. If people remember to do that then they won't make the mistake you mentioned above. However the problem isn't corrected if the ERA+ becomes a smaller number and not the higher number that it is now.

tangotiger
11-08-07, 10:03 AM
As I said, YMMV (your mileage may vary).

For 100+ years, people have had no problem with a pitcher's ERA mean "lower = better". I think it's a horrible choice to make ERA+ the opposite of that. Did you notice that at baseball-reference.com, Sean has OPS+ for pitchers... and guess what... he has the "lower = better". What happens if you want to do BB+ (walks per batter faced)? Are you going to make that also the "bigger is better" convention?

Personally, I find it gobbledygook to flip the numerator/denominator to force the "bigger = better". People understand that low walks by a pitcher is good and that low runs allowed is good. Show the ERA+ as 50, not as 200, if he allows runs at half the league rate. There's not a single person who says that the league scores at twice the rate that Pedro allows runs. But, you will hear them say he allows half as many runs as the league.

Anyway, this is my personal rant on the subject. There's no right or wrong answer. Unless of course you disagree with me :)

Hughes2.50
11-08-07, 12:26 PM
As I said, YMMV (your mileage may vary).

For 100+ years, people have had no problem with a pitcher's ERA mean "lower = better". I think it's a horrible choice to make ERA+ the opposite of that. Did you notice that at baseball-reference.com, Sean has OPS+ for pitchers... and guess what... he has the "lower = better". What happens if you want to do BB+ (walks per batter faced)? Are you going to make that also the "bigger is better" convention?

Personally, I find it gobbledygook to flip the numerator/denominator to force the "bigger = better". People understand that low walks by a pitcher is good and that low runs allowed is good. Show the ERA+ as 50, not as 200, if he allows runs at half the league rate. There's not a single person who says that the league scores at twice the rate that Pedro allows runs. But, you will hear them say he allows half as many runs as the league.

Anyway, this is my personal rant on the subject. There's no right or wrong answer. Unless of course you disagree with me :)All good points. I could always do a Tango Conversion when presenting those results. :D

Soriambi
11-09-07, 01:54 PM
Show the ERA+ as 50, not as 200, if he allows runs at half the league rate. There's not a single person who says that the league scores at twice the rate that Pedro allows runs. But, you will hear them say he allows half as many runs as the league.

Anyway, this is my personal rant on the subject. There's no right or wrong answer. Unless of course you disagree with me :)

That actually makes a ton of sense. I'd never even thought of it before. (I don't want to jump into the conversation because most of it is well beyond my comprehension math-wise, but I just wanted to point that out.)

Munson's 'Stash
11-10-07, 02:18 PM
As for the "4.50" that is being discussed, it sounds like Hughes is saying that he does use the league ERA to compare against, and then adjusts it to a league ERA of 4.50. (That is, he indexes the ERA.) There's nothing wrong with this at all.

Munson however is saying that Hughes doesn't do this, that he actually takes the player's ERA and compares it to an ERA of 4.50, without looking at the league ERA at all, which of course would be a preposterous thing to do. But Hughes later comments in this thread makes it seem that he is doing that part the right way: ERA of 3.0 in a league of 6.0, means an indexed ERA of 2.25 (in a league of 4.50).

I think what you're missing is the chronology. All I had to go on when compiling this was the original posts that I linked to and I don't believe that he had bothered to explain anything about indexing.

Furthermore, IIRC, in those posts he comments that he sets the league ERA to 4.50 because he doesn't have reliable splits or totals for each of the various minor leagues. My point of contention as related to his methodology was that this is pretty arbitrary; without looking at league wide performance and park effects the potential for the math to be influenced by conditions that could otherwise be accounted for seems pretty high; i.e. a guy who does well in a pitcher's league in a big home park should not be compared to a pitcher who has a similar performance in a smaller home park in a hitter's league.



I never said I did what Munson claimed. But then, he claimed so many things I never said, I realized that it was futile to even defend against one straw man.

How did I attack you? I attempted to get you to explain your methodology because your results fly in the face of common baseball sense. I analyzed what I could off of what little you "had released to the public" in this forum with hopes of keeping it as scientifically impartial as possible.

Just because someone questions you doesn't mean they're attacking you: when I wrote the initial post in this thread the google search for "cubic transformations" did not have a all the SoSH stuff that dominated the topic a week later. If I was wrong about something, you had several months to correct it. Instead you tried to have me banned from this site for violating community standards and demanded that I send you my academic credentials via PM. For the record, I only send those out to full professors.

And I'm still waiting for an explanation of how exactly cubic transformations fit into your grand scheme.



The question to ask of forecasts, and this is not just limited to Hughes but to all forecasts: what is the uncertainty level? You can forecast someone with a mean ERA+ of 150, but if your uncertainty level is +/- 75 95% of the time, then that doesn't really help, does it?

This is the crux of the problem and why I started this topic in the first place. Projections need to be more than just fan boy guesses, or the error bars need to be acknowledged at the beginning (if I'm not mistaken, this is a standard procedure in academia). AFAIK he is not running predictions for the entirety of MLB, or even the general population. His starting data set focuses on top prospects and the Yankees. Even if he were somehow correct on some of his predictions, you would still need to apply his results to a much broader population for the results to have value or be considered meaningful.

Frankly, I don't see how H2.5's methodology or results meet any kind of scientific standards, other than perhaps "exhibiting no errors in the arithmetic."

GIDP
11-11-07, 05:01 AM
i dont really understand any of this stuff but im curious if hughes2.5 still blieves this


Here are my projected ERA+ for each pitcher during his prime period (26-33 years old):

Hughes (175)/Wang (130)/Sanchez (140)/Joba (145)/Betances (185)

Billingsley (150)/Elbert (135)/Kershaw (175)/Kuo (145)/Morris(???)

Papelbon (140)/Lester (135)/Bowden (125)/Bucholz (125)/Matz (150)
http://forums.nyyfans.com/showthread.php?p=3791002#post3791002
how did you get the numbers? you still like them?

Carbon Fiber
11-11-07, 09:19 AM
i dont really understand any of this stuff but im curious if hughes2.5 still blieves this

http://forums.nyyfans.com/showthread.php?p=3791002#post3791002
how did you get the numbers? you still like them?

Cool - I never saw that. He said later in that thread that those are the average ERA+s during their 26-33 years, so if he's even close to right, the Yankees will be winning like 120+ games a year for five years running (as a RS fan, if that's close to true, I'm in for a tought stretch!). I don't have the baseball based expertise of many of you here (I'm on the math side), but it seems incredibly unlikely that those projections could come to be. I'd think that baseball projections would aim for the most likely result, but maybe (??) he is instead putting the ultimate ceiling/everything goes right projection to each of those guys.

Or maybe he's just having fun and doing a what-if kind of thing. It looks like that thread got heated quickly - maybe there was history between the posters.

Carbon Fiber
11-11-07, 09:30 AM
Instead of this back and forth with no projections, how about we come up with some contest for the best open (i.e. show your work) projection? Or something. It'll help pass time until P&C.

And this has got to be one of the more exciting baseball seasons coming up prospect wise - even as RS fan, I've gotta say that the Yankees have the best overall crop of SP in their farm then I ever remember seeing before. Math arguments aside, you guys have got to all be psyched to see them develop. Seeing hitters look completely overmatched by Joba this year was a site to behold for all fans of the game.

So, if I come up with an outsiders' (again, more math than baseball knowledge) projection formula, will anyone join in and have fun with it? Or is that best suited for another site?

EDIT: Awww, forget it... the objective data for prospects doesn't appear to be very comprehensive and if they haven't pitched much, there just isn't much there to discover correlations with any confidence --- I'm better at someone giving me a few gig of data and having at it. I think folks must use a lot of subjective factors in with the objective --- and I'm not really qualified for the subjective side of it. Bummer.