Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
The fine people at BP have come up with yet another new statistical gauge, which they call PECOTA. This, as everyone should know, is short for Player Empirical Comparison and Optimization Test Algorithm, and it is a way for measuring a player's breakout -- and collapse -- potential. Preaching to the choir here, I know, but BP is hands down the most creative baseball engine around, which is all the more remarkable when you consider how stat-heavy they are. Here's the article.
Another fascinating study by Baseball Prospectus | 10 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
Coach - Thursday, January 23 2003 @ 09:36 PM EST (#98397) #
Author Nate Silver's not taking the whole exercise of predicting "breakout" seasons too seriously, calling it "a grownup's version of pin-the-tail-on-the-donkey", but trust BP to make it more scientific. A Peter Gammons column is referenced in the article, so for convenience, here are the 20 names he mentioned:

1. Carlos Zambrano, RHP, Cubs
2. Nick Johnson, 1B, Yankees
3. Erubiel Durazo, 1B, A's
4. Jeremy Giambi, 1B-DH, Red Sox
5. Jake Peavy, RHP, Padres
6. Bobby Kielty, OF, Twins.
7. Carlos Lee, OF, White Sox.
8. Brandon Duckworth, RHP, Phillies.
9. Adrian Beltre, 3B, Dodgers
10. Aubrey Huff, 1B, Devil Rays
11. Casey Fossum, LHP, Red Sox
12. Kyle Lohse, RHP, Twins
13. Preston Wilson, OF, Rockies.
14. Jay Gibbons, RF, Orioles
15. Tony Armas, RHP, Expos
16. Toby Hall, C, Devil Rays
17. Corey Patterson, CF, Cubs
18. Joe Crede, 3B, White Sox
19. Danys Baez, RHP, Indians
20. Kevin Mench, OF, Rangers

PG's breakout list reflects how many of his front-office connections named a player. I'd like to see the PECOTA numbers on Josh Phelps, Orlando Hudson, Chris Woodward and the other young Jays not included, just to see how they compare to my own purely subjective forecasts, which are based more on situational changes and hunches than anything else. (For example, I like Joe Crede hitting 7th in a strong lineup in his first full season, but don't buy into the Jeremy Giambi hype because of hard-to-quantify character issues.)

PECOTA is another fine BP evaluation tool, especially if you're preparing for a fantasy draft or auction. I'm sure it's useful, but will it persuade you to pony up the 40 bucks?
Gitz - Thursday, January 23 2003 @ 10:05 PM EST (#98398) #
Phelps is a gaping omission on that list, no question. Orlando Hudson doesn't have much breakout potential this year, however; he's just another rookie. And how many years in a row will Beltre show up on lists like this?!
_Sean - Thursday, January 23 2003 @ 10:13 PM EST (#98399) #
I'm glad to see that research grounds my high opinion of Carlos Zambrano as an underrated gem in the pitching-rich Cubs organization--there is of course no question that Mark Prior is the crown jewel-- but the Cubs do have several intriguing young pitchers who served me well in my intial forays into playing Out of the Park 4 (see also Cruz, Juan)

I'm also high on Kevin Mench of the Rangers. Coach, forget I said that when our fantasy leagues draft. -)

Myself, I'm leaning towards spending the $40. When you aggregate the investment over 12-months' time, it seems fairly modest in price.
_Gwyn - Friday, January 24 2003 @ 01:01 AM EST (#98400) #
Phelps is a gaping omission on that list, no question.

I do not think Phelps belongs on this list, he had an OPS of .925 last year. The only way he can 'breakout' from that is by becoming Barry Bonds!

I think Gammons list and certainly the BP numbers are looking to identify players who are going to be significantly better than last year. Phelps was so good last year, he cannot be expected to be markedly better, just maintaining the numbers he put up last year would make him one of the top ten hitters in the league.
Gitz - Friday, January 24 2003 @ 01:32 AM EST (#98401) #
If you say that Phelps shouldn't be on the list because of his .925 OPS, which was, recall, done in only 265 at-bats, then neither should Crede, Giambi, Huff, Lee, et al. How much significantly better can these guys get? Giambi had a .919 OPS, and whatever disdain you may think I have for him (and you're right), even if he stays in the lineup all year he won't out-perform that by much. Crede, though he had an OPS of 100 points lower than Phelps, can't be expected to go bananas on the league, either, since he had 65 fewer at-bats than Phelps. Huff had a nice year (.884 OPS), and realistically how much better will he be this year? Carlos Lee has been remarkably consistent the last four years, though last year he did significantly improve his walk rate and he'll be 27 in 2003. Those are all good things, and he could be due for a consolidation year. But I'm a pessimist by nature, so prove it to me, Carlos.

I think Coach makes an excellent point in that 1) PG is showing where he has contacts and 2) obviously those contacts don't extend far into the Blue Jays realm, because if Huff and Giambi and Crede are on that list, then Phelps should be, too.
Coach - Friday, January 24 2003 @ 08:52 AM EST (#98402) #
I don't care about the young Jays being left off Gammons' list, which (as Gitz points out) reflects only the opinions of whoever returns his calls. It's the PECOTA scores I'm curious about. I'm sorry if I confused the issue by copying the PG Top 20, but they're the only players referred to in Silver's piece. The whole point is whether a statistical approach can be better than the current methods at predicting "surprise" seasons. (It can't be much worse). I'm wondering how PECOTA will support, or contradict, Gammons' contacts and my own prejudices, when the next installment looks at the seven pitchers listed. Fossum's the only one I'm very high on.

Obviously Phelps isn't likely to improve on his 2002 by 20%, so his PECOTA number for "breakout" will be relatively low, but I'd like to know his "collapse" probablility, compared to Crede's, for one. Youngsters with talent who took longer than Josh to adjust (Nick Johnson and Corey Patterson are on PG's list, but there are many others) will logically score higher on breakout potential; they also shouldn't "collapse" off mediocre debuts unless they lose substantial playing time. This isn't some magic one-stop player-rater, it's a new tool that is intended to be better than random guesses at telling us where we might strike a gusher, and this is an initial field test. You still should multiply PECOTA by the YOGI factor (Your Own Gut Instinct) before deciding to pay $33 for Nick Johnson or pass on Preston Wilson at $19.

The latest BP yardstick gains some credibility with me by identifying Kevin Mench as the best breakout candidate of the 13 hitters on the Gammons list, but even more by cautioning against Little G, who I've always liked on paper but not on (or especially off) the field. Just to clarify on the subject of BP's value, the only thing preventing me from subscribing is poverty.
_Mick - Friday, January 24 2003 @ 09:40 AM EST (#98403) #
Mench is often compared to fellow former large-headed, bald Ranger OF Pete Incaviglia, but from what I've seen of him locally he appears to have a much higher upside. Maybe if Scott reads this he can chime in with the official Ranger take.

As for PECOTA ...
Since I have only skimmed the piece, and as all good journalists are, am more than willing to comment on it anyway, can I assume that the name of the metric is a reverent bow to former KC utilityman Bill Pecota (not to be confused with Biff Pocoroba, of course)?

And if that's the case, can someone computer Pecota's PECOTA so we can start using that as the stat-head's rebuttal to the overused gawfersaken "Mendoza Line"?
_rodent - Friday, January 24 2003 @ 03:47 PM EST (#98404) #
Don Preceptor:

If I click on the Jays' logo, why doesn't take me to a glossary of formulas and terms? I'm just guessing, but I suspect not all aficionados are statheads. Certainly many fans of pre-sabr yore had a clue, and likely a significant percentage of them didn't go roto. It would be nice to know enough of the languge to order a beaver tail. And educational opportunities are always welcome.

El Roedor
_Chuck Van Den C - Friday, January 24 2003 @ 06:55 PM EST (#98405) #
Rodent, please don't confuse the plethora of modern baseball metrics as having anything to do with "going roto". Fantasy leagues, save for a precious few, still have more to do with tallying counter stats and computing the odd ratio than actually attempting to measure performance.

For some self-schooling, try:

Baseball Prospectus, check out the Statistics and Glossary links

Baseball Primer, check out the Stats link

Baseball Reference, check out the Glossaries link
Coach - Saturday, January 25 2003 @ 09:40 AM EST (#98406) #
Rodent, the glossary is a good idea. I know Craig B has mentioned the need for a FAQ on Primer, and we'll link to it if and when it gets published. Chuck's right; there's no more VORP and EQA in fantasy ball than there was in Casey Stengel's head. Most leagues use the simplest of traditional stats; the challenge is getting productive guys at the lowest possible cost.

I'm of a certain vintage, pre-SABR, and fascinated by the revolution, but I still use the goosebumps approach (Frankie Rodriguez and Eric Hinske, among others, made my coachey-sense tingle first time I saw them) then check the stats to see if there's a tangible reason to be so excited. Usually, there is.

BB will be stathead-dominated, and is already overrun by lawyers, because what we're striving for is an intelligent discussion -- the Blue Jays for the thinking fan -- so codgers like you and me will have to face facts. If these threads resemble the conversations J.P. is having with Keith Law, or Theo Epstein with Bill James, I'm a happy guy. But I hope there will also be room for George Carlin and Yogi, because you can observe a lot by watching, even without an Internet connection and a spreadsheet.
Another fascinating study by Baseball Prospectus | 10 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.