PECOTAs are out

by lboros on Feb 2, 2008 10:39 AM EST in St. Louis Cardinals News

PECOTA is the best projection system there is. most major-league clubs either subscribe to it or have developed a similar system of their own. they're only projections, not real numbers --- no need to point that out, anyone. lots of players will over- or underperform these projections. but they're still useful; if they weren't, every team in the majors wouldn't be using them.

the PECOTA numbers are proprietary; you have to be a Baseball Prospectus subscriber to get access. i'm a subscriber, so i have the numbers, but it'd be extremely uncool --- tantamount to theft --- for me (or any other subscriber) to post them extensively here. having said that, BP is pretty generous about teases; they generally don't mind if you give away little bits of the data. hell, they give away little bits of it themselves. accordingly, here are a few teases, in no particular order:

PECOTA sees ankiel as a legit 30-homer hiter; that's his projection, 30 homers. only 10 national-league players got 30-homer projections, and two of them are cardinals (pujols and ankiel).
pujols projects to be the best hitter in baseball, and it's not close. he leads all hitters in projected batting average (.327), on-base percentage (.427), slugging (.577), and equivalent average (.335).
PECOTA has colby rasmus at .245 / .327 / .437 with 21 homers. edmonds' projection is .240 / .327 / .415 with 9 homers.
PECOTA likes the cards' exchange of 3bmen, at least from the offensive standpoint -- troy glaus: .258 / .364 / .472 with 21 homers. scott rolen: .263 / .328 / .430 with 13 homers.
the loss of eckstein is no loss, per PECOTA --- his projected line (.274 / .326 / .352) is almost identical to brendan ryan's (.268 / .322 / .348). unfortunately it's a lot better than cesar izturis' (.251 / .304 / .299).
PECOTA's projections for the cardinal rotation are pretty much useless, because most of the candidates are projected as swingmen --- ie, their projections assume a large number of relief innings, which artificially lowers their era. even wainwright is projected as making 14 relief appearances, along with 24 starts --- his 2006 stats (all compiled in relief) trick PECOTA into assigning him a split role. (his projected era is 4.14.) both clement and pineiro have decent era projections (4.54 and 4.35, respectively), but PECOTA assumes only 10 starts and fewer than 100 total innings from each pitcher; over 180 innings, we'd have to assume worse numbers. the system still likes anthony reyes, projecting him to the 2d-best era (4.32) and 2d-highest innings total (115) among st louis' pitchers.
wainwright's list of comparable pitchers includes john lackey and aaron harang.
rasmus' list of comparables includes carlos beltran. rick ankiel's list of comparables includes jose cruz.
projected era for braden looper: 4.82. jeff suppan: 4.84. matt morris: 5.02. jeff weaver: 5.06.

Tweet Comment 37 comments | 0 recs |

Do you like this story?

Comments

Projection Accuracy

I like the positivity coming from PECOTA, although the Izturis slugging percentage made me cringe!

Question for Larry or anyone else: Is there or has there ever been an effort to rate the various entitites providing projections? Other than crunching a lot of data, I would think it would be easy. I am curious how accurate these things are - I'm not skeptical, I would just want to know on an absolute level and compared to others out there, how accurate these projections are.

by Fred Head on Feb 2, 2008 11:32 AM EST reply actions

I've seen....

that the best projection systems are around 70% accurate and PECOTA is normally at the top. unfortunately I can't remember where I read this

by KoryWiu14 on Feb 2, 2008 12:01 PM EST up reply actions

BP actually compares projections

for various systems at the end of each year, looking at how they compare to real performances. That's one reason why Larry says that PECOTA is highly regarded: it consistently shows up at or near the top of these comparisons. Sure, there's a possibility that that's a self-licking ice cream cone and that BP chooses the metrics for the comparison so that their own product performs (or seems to perform) well. However, BP has a fair bit of integrity and doesn't load the dice in that way. It really is a good system, to all appearances.

That said, PECOTA doesn't do particularly well on players for whom there are no closely comparable guys to compare them to. An important part of their methodology is to look at each player's past few seasons and find the ten players in recent history who had a similar stretch of seasons at some point. (This includes matching size, handedness, etc., as well as stats.) They then see how those ten did in the following season and base the projection for the guy under consideration on the average of those performances. It works very well for "generic" players for whom there are many parallels in previous players' careers, for example most middle relievers, mediocre journeymen middle infielders (so that the Izturis projection is likely to be accurate, alas), and so on. It works less well for guys with unique career paths. The Ankiel prediction may therefore be far too optimisic, since he has few direct comparables, and more hopefully, the Barton one may be too conservative, for the same reasons.

by StanTheManFan on Feb 2, 2008 12:16 PM EST up reply actions

I will be astounded...

if Anthony Reyes has the second-most innings pitched among Cardinal starters...

I figger Wainwright, Looper and Pineiro will all pitch more innings than Reyes; and that's with the "worst-case" scenario that Clement, Mulder, et. al. don't pitch at ALL in '08.

Pitching's the key; IF Clement can contribute, the Cards have a shot at winning every day. (I don't see a "no-hoper" a la Kip Wells or Jason Marquis; but then, I thought DD could help Wells a year ago, so whaddaIknow????)

"In this game, don't nobody know nuthin' about nuthin'." -- attributed to Lawrence Peter "Yogi" Berra

by The Ol Goaler on Feb 2, 2008 11:57 AM EST reply actions

I'm

not saying this as doubting you in any way, but just out of curiosity, LB, how did you find out that every major league team subscribes to BP/PECOTA?

I am a subscriber, too. I think it's a great system. Last year PECOTA had Hanley Ramirez and Matt Holiday valued as first round picks for my league's scoring system and I got them in the 2nd and 3rd rounds, respectively. I would say their projections for those guys panned out pretty well.

by Toddius396 on Feb 2, 2008 12:16 PM EST reply actions

it's common knowledge

e.g., see this NYT article from two off-seasons ago:

For the growing number of baseball executives bent toward statistical analysis, a certain anticipation builds every off-season for the release of what is known simply as Pecota, Baseball Prospectus's überforecast of every player's performance the next season. Most front offices have an employee who consults it - particularly during the free-agent season.

No organization uses Pecota in a vacuum, instead incorporating it with other projection techniques like traditional scouting reports. . . .

"Everyone in baseball is in the guessing business," the Mariners executive Dan Evans said. "This makes it a little bit less of a guess."

http://www.nytimes.com/2005/11/13/sports/baseball/13score.html?_r=1&oref=slogin

by lboros on Feb 2, 2008 12:58 PM EST up reply actions

Youth vs. Experience, Part XCIV

So how many months of .251 will it take before La Russa shelves Izturis in favor of Ryan? I'm not looking forward to the flood of emails about La Russa's maddening preference for mediocre vets.

As I understand it, the rational behind Izturis was that he might rebound to his 2004 numbers. If he hasn't done that by mid-May, I hope TLR pulls the plug.

So says, The Dude

by Titus Pullo on Feb 2, 2008 12:29 PM EST reply actions

shelve Izturis in favor of Ryan

Dude, unfortunately I think you are wildly optimistic on the ETA of Izturis as a fixture on the bench. Ryan cannot spot start for Izturis when he is in Memphis. In order for Ryan to displace Izturis in the lineup, Ryan has to first displace Miles as the utility infielder. Miles may lack talent but he is fairly consistent in delivering to the best of his ability. Absent injury or an implosion by Izturis of the Junior Spivey magnitude, Izturis and Miles will stay on the 25 man which means there is no room for Ryan.

by jjray on Feb 2, 2008 1:06 PM EST up reply actions

If I had to guess...

I'd say that barring injury we won't see much of Brendan Ryan. I tend to be pretty optimistic about most of these things (the Edmonds and Rolen moves, the Clement signing) but for whatever reason I don't see Ryan getting much PT. It'd just be too easy to find another scrap-head veteran to plug in ahead of him.

Sad, Mr. Lebowski. Strong men also cry.

by airhad on Feb 2, 2008 1:44 PM EST up reply actions

oops

that should've been "scrap-heap" with a p.

by airhad on Feb 2, 2008 1:45 PM EST up reply actions

Looper

you have Looper's expected ERA. Is that with or without relief appearances artificially lowering his ERA?

Still looking for 1985 Regular Season games on DVD/VHS

by Hardcore Legend on Feb 2, 2008 1:07 PM EST reply actions

with

32 games, 14 games started

by azruavatar on Feb 2, 2008 2:45 PM EST up reply actions

Hmm, then that above comparision

may not benefit Looper.

Still looking for 1985 Regular Season games on DVD/VHS

by Hardcore Legend on Feb 2, 2008 3:59 PM EST up reply actions

i think

This shows that the Cardinals--while certainly not a good time--have made wiser decisions.

For example--the get a 4.80 something ERA from Braden Looper at a lot lower cost/years than Jeff Suppan. Same for a little better production from Pineiro as opposed to Morris.

Again, I understand they are all mediocre pitchers. I don't want mediocre pitchers. But if we are going to get stopgap/mediocre guys let's at least not overpay for them or overextend them--and I think Mozeliak has done well there.

If Rasmus can put up a +.750 OPS-then he should just play. To me that's a good enough number not to hurt you in CF.

Also, if Ryan can get in the .700 range---then he should play. Same for the second base situation. I actually like Ryan at second more than short.

SUBURBS: Where Americans cut down trees and then name streets after them.

by beanocook on Feb 2, 2008 1:14 PM EST reply actions

But Walt was the one that brought in

Looper and was overseeing things when Looper was moved to the rotation. Mo extended Pineiro.

Still looking for 1985 Regular Season games on DVD/VHS

by Hardcore Legend on Feb 2, 2008 1:24 PM EST up reply actions

but ryan

was brought up as a shortstop, and has a better fielding percentage there than he does at second, hell he basically learned how to play second in the majors.

I'd say if kennedy still doesn't perform to his pre-2007 days, we'll see ryan replace him or be called up as the reserve and miles will get 2nd.

Cardwash - Cardinal, Washington fan (Washington???? Yeah, I know)

by cardwash on Feb 2, 2008 2:06 PM EST up reply actions

I thought I read on the PD

that Ryan's position in hs/college was 2b and that he only switched to SS once he was in the organization. If so, it shouldn't be too much of a stretch for him to return to 2B (if the opportunity presents itself).

by ArkansasTravs on Feb 4, 2008 2:42 PM EST up reply actions

Suppan v. Looper

Suppan's pitching in a more hitter-friendly environment in front of an awful defense. PECOTA takes those things into consideration.

by Rob H on Feb 2, 2008 2:16 PM EST up reply actions

I don't have a subscription...

but did it have any projections for Brian Barton?

Miller sucks.

by Ankiels Missing Curveball on Feb 2, 2008 2:23 PM EST reply actions

.714 OPS

Cheeseburger in paradise.

by joker24 on Feb 2, 2008 2:45 PM EST up reply actions

PECOTA projects

Rasmus to be the 57th most valuable player in baseball over the next 5 years. Albert Pujols the second most valuable.

First?

Hanley Ramierz.

by azruavatar on Feb 2, 2008 3:01 PM EST reply actions

Defense not taken into account on upside

Problem being Pujols could probably play about as good a shortstop as Hanley...

Cheeseburger in paradise.

by joker24 on Feb 2, 2008 3:12 PM EST up reply actions

I was of the opinion that Hanley gets a bad wrap

at SS. He has the range, speed and good hands. He is still young and has to develop the rest of the package defensively.

Still looking for 1985 Regular Season games on DVD/VHS

by Hardcore Legend on Feb 2, 2008 3:46 PM EST up reply actions

especially playing for

an AL farm club like the marlins...

by SleepyCA on Feb 2, 2008 4:16 PM EST up reply actions

With all due respect,

Albert wasn't exactly Brooks Robinson when he came up; in fact, he was a below average fielder. It's actually possible for players to get better with both age and experience.

"The moment we want to believe something, we suddenly see all the arguments for it, and become blind to the arguments against it."

by cardsrul on Feb 2, 2008 11:20 PM EST up reply actions

Albert's Legacy

This is a little bit tangential but still in the spirit of projections. I'm just not sure it's diary worthy. With Albert being #2 over the next five years and probably around #2 for the last seven years, I love to ponder his place in history.

Thanks to Mr. James' Favorite Toy (so much fun, by the way), Albert should play about 7-8 more seasons. I'm sure the Favorite Toy isn't as accurate as Pecota, but it puts Albert's 14 year career at 2720/570/575/1736/1314/875 (H/2B/HR/RBI/BB/SO). And if his rates continue at .332/.420/.620, clearly, he's in elite company (top-20 all time in just about each of those save hits)

I'd like to know what records/milestones y'all would like (or hope) to see Albert finish his career holding? I know this is entirely premature since he's only played seven seasons, but few have played their first seven like Albert.

I'd love to see a Triple Crown season and for him to join the 600 2B/600 HR club (~16% chance according to FT) along with the exclusive Hank Aaron and Barry Bonds. As much as I'd love for Albert to get to 700+ homers, I can live without the scrutiny of PEDs speculation.

I also hope another consequence of Albert's progress is that it sheds light on what a truly great ballplayer Stan Musial was. It always irks me a little to hear about Ted Williams, DiMaggio and Mantle and not Stan (even though those three are no doubt deserving).

Sorry for the long post. Hope someone else has wondered about this, too.

by birdsonthebat on Feb 2, 2008 4:07 PM EST up reply actions

help...

Hey azru,

how did you generate that report? did you go thru and total VORPs from the player cards? If it's generatable or if i am missing it can u tell me how?

by RedbirdRay on Feb 3, 2008 10:41 AM EST up reply actions

If you're a BP member

you can download the weighted means projection spreadsheet from BP (see Nate's unfiltered post). The column "upside" is an attempt to quantify a players value over the next 5 years based on aging curves and skillsets and other proprietary calculations.

Rasmus has the 57th highest upside of all the position players in the system.

by azruavatar on Feb 4, 2008 1:18 PM EST up reply actions

Thanks

I kinda wondered if that was it. In the glossary, it doesn't describe it as anything of the like. I couldn't figure anything else would be it unless it was from individual player cards...

by RedbirdRay on Feb 4, 2008 2:03 PM EST up reply actions

D'Angelo Jimenez

He's listed as a DH, I believe for WAS, but that's a pretty rosy projection. After seeing it -- and the disappointing other options -- I'd like to seen him get first crack at the 2B job. Maybe an offensive/defensive platoon with Kennedy would be in order.

by sjoshi on Feb 2, 2008 3:16 PM EST reply actions

Another Miles casuality...

Jimenez will likely never be on the team. He'll start in Memphis and then eventually get released, I guess, if another major league team expresses interest.

by DiscoJer on Feb 2, 2008 3:22 PM EST up reply actions

Lousy Miles, Lousy Tony

You're right, he is Miles fodder, and it makes me livid.

His EqA and VORP are projected to be higher than Howie Kendrick's, dammit.

by sjoshi on Feb 2, 2008 3:46 PM EST up reply actions

Jimenez, however,

has disappointed pretty well wherever he has gone. He has played for seven major-league teams, only one of them (Cincinnati) for a full season, and they cut him loose the next year.

Here are a couple of samples of why:

2004: PECOTA projection: .271/.349/.404; real performance: .270/.364/.394 -- reasonable enough
2005: PECOTA projection: .275/.359/.413; real performance: .229/.319/.295 -- whoops
2006: not enough games for a real comparison, but such as it was, PECOTA projection: .255/.335/.383; real performance: .183/.333/.268 -- ecchhh!
2007: PECOTA projection: .238/.335/.365 (looks like PECOTA was starting to wise up...); real performance: .245/.379/.373 -- so he's back, sort of. Kind of. Maybe. Almost.

In other words, here we have an extremely inconsistent player who the prognosticators think has the potential to be useful. Except most of the time he isn't, and they never know why. Such a guy is probably worth taking a minor-league-contract chance on, given that there is a black hole at 2B otherwise. Believing any prediction for him, however, is a real stretch, and if you believe he's more likely to pan out than Miles or Kennedy based on an optimistic PECOTA, I have some beachfront property in Amarillo to sell you.

by StanTheManFan on Feb 2, 2008 10:18 PM EST up reply actions

The thing PECOTA was missing

on was the batting average and his power was inconsistent. The walk rates are consistent and if you look back at his BABIPs that explains some variance. He's not inconsistent because that's some inherent quality to his game (or at least I don't think we can come to that conclusion yet) but because there's been some random variation to his batting averages which are the least stable indicator of performance.

by azruavatar on Feb 3, 2008 1:05 AM EST up reply actions

Maybe we can't

conclude that there's something fundamental about his inconsistency, but his previous employers certainly reached that conclusion. 50 years ago, or maybe even 10 years ago, I would have been willing to accept the possibility that that was due to cigar-and-panama-hat scouting combined with lazy managing, but not any more. Major-league teams get the BABIP variations now and are willing to accept that some of them are irreproducible flukes. They still gave the guy his walking papers five times in the last five years.

Don't get me wrong -- I like the move to acquire him. Jimenez might turn out to be useful. He was a cheap enough gamble that any upside at all will make it look like a good investment. But I'm not holding my breath, and the notion that he quote-unquote "should" be preferred to Miles or Kennedy at this point because of a favorable PECOTA projection still looks pretty silly to me.

by StanTheManFan on Feb 3, 2008 1:21 AM EST up reply actions

I don't disagree

"But I'm not holding my breath, and the notion that he quote-unquote "should" be preferred to Miles or Kennedy at this point because of a favorable PECOTA projection still looks pretty silly to me."

What bothers me is that there are three players (Hoffpauir, Ryan, Jimenez) that all project better than Miles or Kennedy. I'll take the odds that one of them pans out, which just continues to make the Miles signing look silly.

by azruavatar on Feb 4, 2008 1:20 PM EST up reply actions

I was surprised...

...that PECOTA had Hoff as the 7th best Cardinals hitter in projected VORP. The projection is encouraging, too bad management probably won't play it out that way.

by roarke on Feb 4, 2008 8:55 AM EST reply actions