Wednesday, April 27, 2005

+/- in baseball?

Since there is already a statistic for scoring pitchers, I'm wondering if I could come up with a similar system for hitters. Pitcher game scores are predicated on an average game being 50 and a perfect game being 100. So what would an average game be for a batter? 1 for 4 with a walk? What's a perfect game? 5 for 5 with 4 home runs and 12 RBI? Hitting is only part of the concern for me. I'd also want to account for fielding and baserunning. The problem is that it would end up being too subjective. For example, I'd want to subtract points not only for errors, but for plays that could/should have been made but didn't count as an error. Strangely enough, baseball does not account for mental errors. Perhaps a 0-100 scale is too complicated, and a +/- system similar to hockey would work better. Of course, the real stat nerds have already come up with BRAR and FRAR, so an easy way to come up with a +/- is to just add those two stats together. That still doesn't account for baserunning, however, and the RAR stats are really only good for looking at many data. I just think it would be neat to come up with one tidy number that told you immediately how good a player's game was. In Tuesday's Reds/Cubs game, for example, Adam Dunn had a good game at the plate, going 2-4 with 1 run scored, an RBI single, a solo homer, and 2 walks. However, he also struck out once, left 3 men on base, was caught stealing, and made a hideous misplay in left that led to a single becoming a triple. So did he have a good game or not? Well, in the pitching game score stat, an each earned run a pitcher allows results in 4 points being subtracted from his score. So I'd have to think that each RBI would be worth 4 points for the batter. Then throw in a point for each total base, a point for a run scored, a point for a walk/HBP, a point for a SB, a point for advancing from 1st to 3rd on a single. Defense would be more subjective--you can't award points for every putout or assist, because that would unfairly favor certain fielders. I think the best way to go would be to award points to difficult defensive plays made with runners in scoring position. So like, robbing someone of a grand slam would be worth 16 points. Robbing someone of an RBI single 4 points. And so on. Now for the negatives: subtract a point for each out made. Subtract two for each out made with a runner in scoring position. Subtract a point for base directly related to a defensive gaffe--and subtract two for every run that scores as a result of the misplay. Subtract one for a caught stealing. Subtract two for each double play. Subtract one for getting thrown out at third as a baserunner (not a base-stealer). Hmm, that's all I can think of for now. Let's see what Dunn's score would be:

+ 8 (RBI x 4)
+ 5 (total bases)
+ 1 (run scored)
+ 1 (walk)
- 2 (outs made)
- 6 (outs made with runner in scoring position x 2)
- 2 (total bases allowed by defensive misplay)
- 1 (caught stealing)

and we get +4. That's not +4 runs--just +4, like in hockey. Let's check out Alex Rodriguez's +/- from Tuesday night, as A-Rod unequivocally had the game of the year so far (from what I could tell on the game log, Rodriguez didn't have a major impact on the game defensively--two runs scored on groundballs to third, but as far as I know, he didn't make any mistakes).

+40 (RBI x 4)
+13 (total bases)
+3 (runs scored)
-1 (outs made)

and we get +55. Now let's look at Kevin Millar from Tuesday night, who did not have a good game.

-5 (outs made)
-4 (outs made with runners in scoring position x 2)
-1 (bases allowed by error)

and we get -10. Hmm. This is starting to make some sense to me. Dunn's negatives are about equal to Millar's but since Dunn did so much more with the bat, he has an above-average game. A-Rod's game is obviously an outlier. I need to do one more--see if I can't get something to add up to zero. Let's try Brian Roberts--from a quick scan of the game log, he didn't even get a play in the field, so let's assume his defensive influence adds up to zero:

+4 (RBI)
+2 (total bases)
-4 (outs made)
-6 (outs made with runners in scoring position)

to arrive at -4. Hmm. Not bad. The interesting thing is that Dunn, Millar, and Roberts all left multiple men on with runners in scoring position. What did they do to make up for it? Dunn homered and drove in another run with a single. Roberts drove home a run with a double. Millar didn't do jack squat. Dunn's score would have been better, but he had a below-average game in the field and on the bases. If Dunn had performed better in those phases of the game, his overall +/- would have been as good as Millar's was bad.

Considering I was just flying off the seat of my pants there for the past hour or so, I'd say this is a pretty good result. Next time I watch a game (which could be tomorrow night's ESPN Wed. night baseball late game), I'm going to keep score using the +/- system as opposed to the traditional method, and I'll see what results I come up with. I'm hoping that all the scores will be evenly distributed around zero.

Stay tuned...


Post a Comment

<< Home