Some Notes on the OLIVER Batting Projections
Maybe for Jon Kabat-Zinn and his cadre of ubercalm friends, living in the present moment represents the greatest of goods. Indeed, speaking as someone who’s legitimately jealous of his own life, I can see the merits which lie therein.
However, I also know that, literally, one of life’s greatest joys — and, listen, people, I’m not effing around for even one second when I say this — is the joy derived from poring over pre-season baseball projections.
The whys and wherefores of this joy are a matter I’ve discussed in these pages with noted projectacators Sean Smith and Beloved Pole Dan Szymborski, so I’ll refrain from offering any poorly formed theories here.
Still, it makes sense to note that recent developments in the field add to the sense that projections are headed somewhere. Here I’m thinking specifically of the aforementioned Smith’s entrance into to the Mysterious Innards of Major League Baseball; of Tango’s continued attempts to assess the accuracy of projections via the annual Forecasters Challenge; of the Fan Projection project here at the site and its relative (if not rousing) success in Tango’s competition.
While I possess nothing like Tango’s facility with measuring the quality of projections, it’s totally within my skill set to provide idle commentary on some surprising individual cases.
In this post, I’d like to do that exact thing — with OLIVER’s batting projections, specifically.
The OLIVER system comes to us via the talent of Brian Cartwright and is hosted by The Hardball Times. It’s notable for a number of reasons, as follow:
1. It costs money.
Specifically, $14.95.
2. The interface is excellent.
It’s possible to filter by level, for example, which is helpful for identifying minor-league talent.
3. It projects players in terms of WAR.
And includes fielding projections.
4. It’s super extensive.
There are projections for over 2,500 batters. That seems like a lot.
5. There’s also a customizable fantasy guide.
To determine dollar values for auction leagues, for example.
I’m aware that the above sounds more or less like an advertisement. This isn’t my intention. And note, please, that I’m not qualified to speak to the accuracy of the projections (in particular, because OLIVER wasn’t one of the contestants in last year’s Challenge). If I’m enthusiastic, said enthusiam derives from (a) my natural excitement about the field of projection (for which the reader should adjust, if possible) and (b) THT’s presentation of the OLIVER projections, specifically.
Getting to the batting projections themselves, here are three notable items.
Regarding the Value of Catchers
Catchers are ranked way high by OLIVER — higher than I’d have thought. Joe Mauer‘s second overall, with a 6.1 WAR; Carlos Santana‘s 12th (4.5 WAR); Brian McCann‘s 13th (also at 4.5); Buster Posey‘s 17th (at 4.1); Jesus Montero‘s ranked as a catcher and is 26th (3.5); Matt Wieters (also 3.5) is right behind him. Tyler Flowers and Robinson Chirinos aren’t far behind, either.
That’s six catchers among the top-27 position players, and nine in the top 50.
There are four (or probably more) possible explanations for this:
1. We’re in a pretty excellent period for catchers.
2. Catchers who can hit are more valuable than we might assume.
3. The OLIVER projections for catchers are optimistic.
4. The OLIVER positional adjustments for catchers are high.
Again, I’m not qualified to comment with any depth on these matters.
Regarding the Value of Fielding
OLIVER appears to be quite aggressive with regards to projecting runs afield, which range from +30.4 runs (Atlanta prospect Andrelton Simmons) to -23.3 runs (Colorado’s Chris Nelson, rated as a shortstop).
By comparison, our current Fan projections — with 251 hitters projected — feature a range between +15.0 (Carl Crawford) and -12.0 (Andre Ethier). Shockingly, Andrelton Simmons has not met the requisite minimum votes.
The result of such aggressive fielding projections is that quite a few players with underwhelming offensive skills rate as major league-caliber players. Simmons, Jordy Mercer, Peter Bourjos, Drew Stubbs, Colin Curtis, Brandon Crawford, Cliff Pennington, and Zack Cozart all qualify as basically average-or-better major leaguers, mostly on the strength of their defensive numbers.
A Kinda, Sorta Prospect List
In terms purely of projected 2011 WAR, here’s your top-10 prospect list:
1. Jesus Montero, C, NYA
2. Robinson Chirinos, C, TBA
3. Derek Norris, C, WAS
4. Andrelton Simmons, SS, ATL
5. Mike Moustakas, 3B, KCA
6. Zack Cozart, SS, CIN
7. Jason Kipnis, 2B, CLE
8. William Myers, C, KCA
9. Matt Young, OF, ATL
10. Zelous Wheeler, SS, MIL
New Minnesota shortstop Tsuyoshi Nishioka’s projected 3.3 WAR would place him second on this list.
Are you taking Javier Vazquez in Round 2 again this year?
Who would ever do anything SO absurd?
Just ridiculous.
I thought it was Harden in the third round!
If Stanton puts up those numbers, he will be the NL MVP…. if you want to see those numbers, you’ll have to purchase the subscription… I highly recommend it.
Looking at Mauer’s WAR on FG, and using the WAR graph tool, it’s pretty clear FG drastically undervalues catchers.
MLB catchers have an average .311 wOBA, about .330 for the those with 81+ games started. There’s a lot of guys who catch but can’t hit. Compare to the average MLB 1b having a .357 wOBA, .377 for a starter. Therefor the few guys who can catch and hit well give their teams quite a bit of relative value.
I do need to use the regressed fielding values in the projections, as I use regressed for batting and pitching. I did calculate that Simmons had +30 runs in 2010, and scouts have given him excellent evaluations, but we can not yet assume that he’ll stay at that level going forward. On the other hand, Jordy Freakin’ Mercer was +11 in 56 pro games in 2008, +27 in 137 games in 2009, and +23 in 142 games in 2010, so that looks legit. He just needs to find a replacement level bat, or he’ll be Ronnie Cedeno with an even better glove.
Crawford’s numbers are absurdly low
Considering Wheeler is, at best, a fringe prospect, his inclusion makes me wonder about the entire list.
Oliver’s MLEs of Wheeler’s performance
Year LevelAge wOBA BA OB SA FR WAR
2008 A 21 .273 .218 .286 .322 +11 -0.3
2009 A+ 22 .311 .246 .327 .370 +4 +1.0
2010 AA 23 .320 .257 .342 .366 +7 +2.1
A 310-320 wOBA is slightly below average for a MLB SS, definitely below a starter, with a good but not great glove. But when his competition is Yuniesky Betancourt (.290 wOBA, -9 FR, -0.8 WAR) he should be a prospect.
A 1-2 WAR player usually only starts for a lower tier team, but is good enough to be on a MLB bench. But when 2 WAR is the 3rd best position player on a team (trailing only Fielder and Braun) that’s not good (for the team).
From what I have read, Wheeler’s weight is catching up to his projected wOBA and the question of whether he can stick at SS has expanded to whether he will become a bigger cult hero than Pablo Sandoval. If a 5-10, 230-pound SS sends Yuni to the bench, both questions answer themselves. But if he continues to eat himself into a corner, the answer is probably Dmitri Young.
“Eat himself into a corner” made me laugh out loud. I like the visual of him standing in a room, not allowed to be in the middle because of how fat he is. All while holding a first baseman’s mitt.
You forgot about Weeks (4ish), McGeehee (3ish), Hart (2.3) all better than 2.1
Weeks as well. Oliver has a playing time estimate, and THT has a team expert who constructs a depth chart. Oliver predicts 518 PA, brought down somewhat by Weeks injury shortened season in 2009, while the depth chart says 658 (but who wants to bet that Weeks stays healthy two years in a row). Throw in a -10 glove, and Weeks is 2 WAR +/- 0.5. If Betancourt does play SS, every infielder (Fielder, Weeks, Betancourt, McGehee) is brutal with the glove.
When is John Eric Hanson’s Slapshot coming to fangraphs?
I would not count on them before mid-March.
I have no idea how the process of getting them posted here will work. In particular, I don’t know if it’s reasonable to send a set of projections followed by updates. I’ll probably have my computer projections in mid-February (when I know where (almost) everyone is playing) and then I manually update for factors that I don’t have coded (generally things that I don’t have consistent data for) and playing time estimates. I never actually finish this process, but my one fantasy league has it’s auction on March 19 . . . so I should be 98% done and ready to share by then.
That’s amazing! I can’t wait.
I have my main league draft on the 12th of March. Any chance they’ll be up by then?
wow, even cliff pennington is average or better? lol
For a reality check, fangraphs has him as a top 5 shortstop in MLB.
Did Drew Stubbs not just have a 3+ WAR year with his bat alone?
Seems a bit harsh to lump him in with the all glove no bat guys, or was it a mistake?
That is the highest rating on Derrek Norris I’ve ever seen. Man can get on base, and the power should be back with a healed wrist, but for a guy who has never played above High A, I’m curious about how he piles up the WAR.
Honestly, how are Stubbs and Pennington on your list? They can actually hit..
How is Sunnyvale?
Brandon Crawford sighting!
The projected WAR is what the player is expected to produce in MLB, based on the MLEs of his past three seasons and his expected playing time (plus defense).
In the past three seasons in the minors, Derek Norris compiled MLE WARs of 1.8, 2.3 and 1.9. As he will be only 22 in 2011, he still has room to improve by the aging curve. His peak is expected to be around 255/370/490. Not much batting average, a good amount of walksn and 25-30 HRs. A .370 wOBA is hard to find in a catcher, so that level of production would make Norris worth 3 to 4 WAR – a solid starter and maybe All Star.
Drew Stubbs the past three seasons has had MLE wOBA’s of .300, .304 and .324, only last year’s approaching average for a MLB CFer, but he is tremendous with the glove, +12, +28 and +17 by my measure. Being 26 in 2011, Stubbs is expected to maintan this same level for the next few years.
Brandon Crawford’s UCLA numbers did not translate well, a .379 wOBA there producing a .263 MLE. The past two yers in the minors he has MLE wOBAs of .290 and .287, barely above replacement for a MLB shortstop, but in those seasons has been +21 and +22 with the glove, which puts him at 2 WAR.
“of the Fan Projection project here at the site and its relative (if not rousing) success in Tango’s competition.”
Not sure you can call finishing 10th out of 21 (in the Overall Results table) a “success” for teh Fan Projection project. Wouldn’t you expect the average projections of fans to be, well, average? I would define “success” for a projection system as some evidence that it is more useful than average.
I think 10th out of 21 is quite an accomplishment. Sure, that’s roughly average, but it’s roughly average scored against some of the best, most complex projection systems available. I’d be one to call that performance a rousing success.
From one point of view, the Fan Projections have a role similar to that of Marcel; they form a baseline behind which any rigorous projection system should not fall. If your projection system can’t beat Marcel (read: most simple computer projection) and the Fan Projections (read: simple aggregation of fan knowledge), then what information is your system contributing?
In theory, all good projection systems should beat Marcel, and should beat the Fans. That the Fans performed so well tell us that either the Fans know more than we give them credit for or that projection systems are not as advanced as we would like to believe.