Tuesday, February 01, 2005

Saberskeptic

While legions of baseball fans have become enthralled with the science of sabermetrics, I for one continue to have my doubts. For those of you who thought baseball was simply a game ("you throw the ball, you hit the ball, you catch the ball. Sometimes you win, sometimes you lose and sometimes it rains. Think about that for awhile." -- Bull Durham) then think again and say hello to the arty science of Sabermetrics. While it sounds like those in the Star Wars universe have bid adieu to the standard unit of measurement, it is actually the mathematical analysis of baseball statistics made popular by Billy Beane (Oakland A's GM) and the book Moneyball. While Sabermetrics can provide some interesting insight, other times it just makes you scratch your head say "huh?" Case in point, a fairly new analysis being done at Baseball Musings whereby the defensive probability of a player making an out is being predicted via a little formula called "The Probabilistic Model of Range". Catchy ain't it? The system uses six factors to help determine this probability that run the gamut from vector (direction of hit) to what ballpark the hit was in. However I should warn you that for the average person to try and wrap their brains around it requires abundant quantities of caffeine and sugar (I recommend about three Red Bulls and a box of Krispy Kremes for the first hour or so).

Like many people I was fascinated at first glance. Mainly because the PMoR made the case that the Angels were the worst fielding team in baseball in 2004. Worse than the Royals and Expos? Yep, apparently quite a bit so. Heck, even that Triple A “team” the Diamondbacks fashioned from what appeared to be extra grounds-keepers and some college interns pulled from the IT department was sounder defensively than those bumbling, stumbling Angels. That caught my attention. But it was in the more recent analysis where the PSoR was applied to 2004 third basemen that I had to finally shake my head and say "no mas". The Angels very own Chone Figgins, the man who made every throw to first base an adventure, is the fourth highest ranked player at the hot corner defensively. I realize he was not a "full time" third baseman which could skew the numbers a bit but c’mon. Figgins had 21 errors in a little over 700 innings which ranks him somewhere between Gary Gaetti and Scott Spiezio in the pantheon of Halo third basemen. My most vivid memories of Figgins playing third in ‘04 is of sailing throws and diving stops by Darin Erstad. In fact, it could probably be argued that Erstad was as responsible for Figgins' "success" defensively last season as anyone. But Figgy, as much as I love the guy's work ethic and versatility, had 11 errors in 705 innings at third. He had the 21st highest error total in the league for third basemen despite being a part-time player and was ranked 81st in fielding percentage with .943. By comparison, the "shaky" defense of Dallas MacPherson (which was, allegedly, the only thing holding him in the minors much of last season) had zero errors and 1.000 fielding percentage in his brief 93 inning stint. Clearly D-Mac took care of business with his glove even if his hitting lagged which probably had a lot to do with the Angels confidence in the youngster for the coming season.

But any statistical model that puts Figgins among the elite defensive third baseman has to make you raise an eyebrow. Of course, that is the beauty of statistics. Very often they help you make a point while simultaneously helping someone else make a counter-point. Kind of an 'eye-of-the-beholder' deal. The problem with building more and more complex metrics with which to reduce baseball players to is that the numbers begin turning on themselves. For every Scott Rolen (.977 fielding % & 10 errors in 1228 innings played) who is rightfully placed near the top (#5 in PMoR), you have a Melvin Mora (.943 FP and 21 errors in 1210 INN) just six slots down and ahead of #12 Alex Rodriguez (.965 & 13 E in 1364.1 INN) and amazingly way ahead of #29 Mike Lowell (.982 FP & 7 E's in 1326 INN). And no, the irony of fighting statistics with statistics is not lost on me. In fact, that is kind of the point. Most baseball fans intuitively love statistics. We obsess over batting average, on-base percentage, strikeouts-to-walk ratios and beers per nine innings (well, maybe that one is just me). Statistics and baseball go together like the Simpson sisters and horrible contrived lip-syncable pop music. But maybe we need to put our collective slide-rules down and keep our analysis to fifteen or sixteen key statistical categories. And somebody tell Chone Figgins' agent to call Billy Beane pronto and demand a trade and huge raise to play third for Oakland instead of that slacker Eric Chavez who ranked a paltry ninth in PMoR rankings, five spots behind Figgy.

7 Comments:

Anonymous Anonymous said...

Get ready for every pinhead on the internet to call you a moron or just click their tongue in superiority at you. The Statitology Church is not keen on dissent. If you don't believe them - you are an infidel. Maybe Sabermetrics is based on the Koran.

--Tha Halofan

11:39 PM  
Blogger Richard said...

PMR quantifies range only, so Figgins' throwing difficulties don't enter into it.

But yes, anything that says Figgins and Melvin Mora have better range than A-Rod raises a few questions. Personally I’d rather stick with UZR, which is much the same concept, only with fewer of those blatant inconsistencies.

Defensive metrics are in their infancy, and none of them are particularly effective at this point. So you can’t read too much into any of them.

12:25 AM  
Blogger Rich said...

I had that thought of a potential backlash when I wrote that little piece. But to clarify just a bit, Sabermetrics is a tool, not a profession. As Richard pointed out in his comment, the defensive side is particular new and dicey. But anything that makes you look at players differently and opens up previously unnoticed potential talent is a useful tool. Very similar concepts to how people look at and rate companies prior to buying stocks. The numbers are interesting but never tell the whole story and can often be misleading.

Anyone who relies solely on one method of evaluation is foolish. Be it an ex-player like Ron Dibble (I PLAYED the game! I know! You didn't play so you don't know!) or a statistics guru who disregards attributes such as leadership, clutch hitting and being a team-player. Somewhere in the middle there is a balance and that is where the gifted GM's live.

8:22 AM  
Blogger Josh said...

I have to mostly agree with you there. I am glad that the Sabermetricians are making an effort to find one universal stat that tells you everything you need to know about all players (it is much like the physics quest to find the universal equation). Just don't say there is a thing such as clutch hitting around Richard. He is likely to go postal.

7:40 PM  
Blogger Rob said...

Piping up here -- it should further be noted that Figgy's "success" has come in a grand total of 2,031 chances. It's not exactly a small sample, as some claim, but on the other hand, it's not the 3,000 plus Beltre and Rolen got, either.

11:51 AM  
Anonymous Anonymous said...

Nice site. Check mine out if you can. global investing

1:14 AM  
Blogger Investment Center said...

Hey, you have a great blog here! I'm definitely going to bookmark you!

I have a basics investing
site/blog. It pretty much covers basics investing
related stuff.

Come and check it out if you get time :-)

9:11 AM  

Post a Comment

<< Home

Free Guestbook from Bravenet.com Free Guestbook from Bravenet.com