View Full Version : Quantrills PECOTA ERA
incarnadine
03-31-04, 03:19 PM
I bought the Prospectus 2004 book, and can't understand how Quantrill's projected ERA was over 4.00.
His last three years are 3.04, 2.70, 1.75. Career 3.66 ERA. How could they possibly project such a dramatic fall off the cliff?
AngelAstro
03-31-04, 03:36 PM
I forget exactly how PECOTA is determined. But knowing BP, it has to do with moving from the NL to the AL. Moving away from pitching haven Chavez-Ravine. Moving to a team with a terrible defense plus the fact that he doesn't strike out many. Aging a year.
WiffleWOOD
03-31-04, 04:17 PM
At the heart of PECOTA is the comparison of the player in question with as many as 100 like players throughout baseball history. The players are compared at the exact same age, and over a number of categories (all listed in the book) to find the most probable performance one can expect from a player of the same type.
since Quantrill is a reliever going on 35, I bet a large portion of the pessimistic projection is the relative success (or lack thereof) of pitchers of his quality at age 35.
KeepClaussen
03-31-04, 04:53 PM
Quantrils hit rate last year was around 24%, the 2 years before it had been around 30%. PEOTCA predicts a regression there. His HR rate was insanly low last year and it predicts he will not be able to maintain that either. Hes also moving out of the best pitchers park in baseball.
I think the projection is pretty good. Quantril doesnt K guys, His BB rate is good but not good enough to be more then slightly better then average unless he has some flukeyness like last season.
bakntime
03-31-04, 05:22 PM
Originally posted by AngelAstro
Moving to a team with a terrible defense plus the fact that he doesn't strike out many.
I still think this is something that sabermetrics (and many people) overrate terribly.
1) The Yankee defense isn't that terrible... Heck, it didn't seem to affect Kevin Brown yesterday, and on hard turf no less.
2) The defense doesn't play a big a part as it's made out to. I think it's used more as an excuse as to why pitchers are bad and a way to lessen good numbers by guys who play behind "good" defenses.
As long as a team isn't making lots of defensive mistakes, a good pitcher won't become a bad pitcher by going from the Dodgers, for example, to the "horridly" defensed Yankees.
WiffleWOOD
03-31-04, 05:25 PM
Originally posted by bakntime
I still think this is something that sabermetrics (and many people) overrate terribly.
1) The Yankee defense isn't that terrible... Heck, it didn't seem to affect Kevin Brown yesterday, and on hard turf no less.
2) The defense doesn't play a big a part as it's made out to. I think it's used more as an excuse as to why pitchers are bad and a way to lessen good numbers by guys who play behind "good" defenses.
As long as a team isn't making lots of defensive mistakes, a good pitcher won't become a bad pitcher by going from the Dodgers, for example, to the "horridly" defensed Yankees.
Sort of besides the point, but PECOTA does not use team defense in their projections, so this has nothing to do with Quantrill's numbers.
WiffleWOOD
03-31-04, 05:37 PM
also, it's reasons like those that bakntime brings up that I think DIPS (Defense Independent Pitching Statistics) are the most valuable tool of measuring a pitcher's performance. Stats like K/BB are much more indicative of an individual's performance than W/L, or even H.
http://www.baseballstuff.com/fraser/articles/dips.html
bakntime
03-31-04, 05:37 PM
Originally posted by WiffleWOOD
Sort of besides the point, but PECOTA does not use team defense in their projections, so this has nothing to do with Quantrill's numbers.
Yes, I was more commenting on that particular post about bad Yankee defense.
WiffleWOOD
03-31-04, 05:39 PM
Originally posted by bakntime
Yes, I was more commenting on that partcular post about bad Yankee defense.
i know, i was sort of responding to that one, indirectly, too
bakntime
03-31-04, 05:56 PM
Originally posted by WiffleWOOD
i know, i was sort of responding to that one, indirectly, too
:lol:
Sorry :D
also, it's reasons like those that bakntime brings up that I think DIPS (Defense Independent Pitching Statistics) are the most valuable tool of measuring a pitcher's performance. Stats like K/BB are much more indicative of an individual's performance than W/L, or even H.
I don't like DIPS. I think it's too exclusive. How can you take seriously a metric that doesn't count singles, doubles, triples, ground outs, pop ups, etc, etc?
That's where I take issue.
My contention is that a good pitcher might have a good DIPS simply because they're good, rather than it being due to the fact that DIPS is a "true" indication of how good a pitcher is. It's open to WAY too many anomalies for pitchers who allow the ball to be put in play too often. You can have a good K/BB ratio and still be a bad pitcher. Maybe it's rare, but it CAN happen, which is why you have to take ALL pitching stats together. There is no one single pitching metric that is good enough to stand on it's own, especially not DIPS.
incarnadine
03-31-04, 05:59 PM
I was aware of all the reasoning that people brought up, I was just shocked to see such a stark difference. And it is not as if Quantrill pitched three major league innings: he was in over 80 games and threw to a large sample size of batters.
Going from 1.75 to 4.30 is like moving from Safeco Field to a silo.
A silo on the moon, maintained in high vacuum.
AngelAstro
03-31-04, 08:12 PM
While defensive metrics surely are not perfected. Defensive Efficiency which is the balls put in play fielded by the defense shows that the Dodgers had the 5th best DE in the league at .7201, while the Yankees had the 4th worst at .6984. I think that we can all agree that allowing more base hits is going to be bad for a pitcher.
In any case, this point is moot because I was wrong in assuming that PECOTA took defense into account.
AngelAstro
03-31-04, 08:35 PM
Originally posted by WiffleWOOD
also, it's reasons like those that bakntime brings up that I think DIPS (Defense Independent Pitching Statistics) are the most valuable tool of measuring a pitcher's performance. Stats like K/BB are much more indicative of an individual's performance than W/L, or even H.
http://www.baseballstuff.com/fraser/articles/dips.html
I can agree with this; however, if you believe that K/BB is indicative of an individual's performance and you know that Quantrill doesn't strike many people out (K/BB is roughly 2:1), plus the fact that the Yankees defense is less efficient than the Dodgers defense, one would expect a decrease in performance.
Believe me, I hope this is not the case, but I wouldn't expect Quantrill to have similar numbers to 2003.
I think both it and ZIPS are nowhere near as telling as ERA+. I think looking at a pitchers ERA+ over his last several seasons is far more informative than comparing "similar" players. When I look at some of the "top ten most similar players" lists I have to laugh. They show players dissimilar to the player in question and to each other.
For Derek Jeter (for instance) they have Jackie Robinson, Ray Durham (!), Glenn Wright (!!), George Grantham (are you kidding me?), Edgardo Alfonso, Edgar Renteria, Carlos Baerga, Thurman Munson, Nomar Garciaparra, and Cecil Travis.
Derek Jeter is a SS witha career OPS+ of 122 and scored a phenomenol 926 runs in 1212 games
Durham was a 2B, has a career OPS+ of 103 and scored 888 in 1310 games. In case you're going to say Durham didn't have the guys behind him Derek did, remember Durham played on the same team as Frank Thomas when people were comparing him to Ted Williams.
Jackie was the closest, with an 862 comparison (whatever that means). Durham's comparison was 853. But Durham's comps don't include Jackie or Derek. How can that be? And how can anything that uses a list that says Jackie Robinson and Durham are the two most similar players in baseball history to Derek Jeter be accurate?
It can't. PECOTA is a joke.
Originally posted by markp
I think both it and ZIPS are nowhere near as telling as ERA+. I think looking at a pitchers ERA+ over his last several seasons is far more informative than comparing "similar" players. When I look at some of the "top ten most similar players" lists I have to laugh. They show players dissimilar to the player in question and to each other.
For Derek Jeter (for instance) they have Jackie Robinson, Ray Durham (!), Glenn Wright (!!), George Grantham (are you kidding me?), Edgardo Alfonso, Edgar Renteria, Carlos Baerga, Thurman Munson, Nomar Garciaparra, and Cecil Travis.
Derek Jeter is a SS witha career OPS+ of 122 and scored a phenomenol 926 runs in 1212 games
Durham was a 2B, has a career OPS+ of 103 and scored 888 in 1310 games. In case you're going to say Durham didn't have the guys behind him Derek did, remember Durham played on the same team as Frank Thomas when people were comparing him to Ted Williams.
Jackie was the closest, with an 862 comparison (whatever that means). Durham's comparison was 853. But Durham's comps don't include Jackie or Derek. How can that be? And how can anything that uses a list that says Jackie Robinson and Durham are the two most similar players in baseball history to Derek Jeter be accurate?
It can't. PECOTA is a joke.
I don't think Baseball Reference's similiarity scores (which I'm guessing is what you're talking about) have anything to do with PECOTA. PECOTA attempts to be projective.
Here's how they figure them out....
http://www.baseball-reference.com/about/similarity.shtml
Whether it's a valid list is certainly up for debate. I'm not going to do the math, but I think some of the "weirdness" may be due to the way they weigh for different positions.
justin32099
03-31-04, 11:58 PM
Originally posted by bakntime
The defense doesn't play a big a part as it's made out to. I think it's used more as an excuse as to why pitchers are bad and a way to lessen good numbers by guys who play behind "good" defenses.
Actually, with Quantrill's low strikeout rates, the defense is fairly important.
Powered by vBulletin™ Version 4.0.7 Copyright © 2013 vBulletin Solutions, Inc. All rights reserved.
0