Brothers in Harm

Matt Bruce

Consider these two batting lines, compiled through the end of June by a pair of teammates:

	AB	R	H	2B	3B	HR	RBI
Player A	320	43	77	15	1	22	57
Player B	308	47	90	15	0	21	60

A casual fan might think that each is having an outstanding year. Both are on pace to finish with 30 doubles, 40 home runs, and 120 RBI. Although their traditional "production" numbers are quite similar, one has 13 more hits in 12 fewer at-bats. Those numbers may seem small, but they conceal a much larger difference. Let's add some stats to bring you a fuller picture, shall we?

	AVG	OBP	SLG	BB	SO	GDP
Player A	.241	.256	.500	8	43	15
Player B	.292	.373	.545	35	86	5

Batting average and slugging percentage were both derivable from the first table, but actually seeing them computed is useful. The difference in on-base percentage makes the comparison all the more interesting, while the GDP total is something I threw in for fun.

By now, any faithful reader of Baseball Prospectus should know exactly who Player A is. I have my friend and former Howe Sportsdata colleague, Brian Joura, to thank for bringing Player B to my attention. If you've guessed who he is, you might assume that I got the GDP totals confused. Nope, Garret Anderson actually has 15 twin killings to Mo Vaughn's five.

One might ask how Anderson manages to keep his run and RBI totals comparable to Vaughn's despite being such a worse offensive player. Brian, an intelligent fan who has yet to buy completely into the stathead gospel, asks a different set of questions:

"Since when is OPS (why is it okay to add things with different denominators?) the most important thing in the world? Why is it more important than scoring runs?"

The easy, lazy answer to that question is to point out that OPS is a better indication of how many runs a team will produce than the runs and RBI totals of particular sluggers. For those who have not seen it (a category that until recently included myself), I point you to a classic article in the Stathead.com library on the effect of one player's good OBP on his team's runs scored.

As eye-opening as are Tom Fontaine's conclusions, this article alone won't be enough to turn your friends into stathead disciples. He describes his simulation well, but readers won't be able to duplicate it without a copy of his code or at least a really specific description.

Brian's critique is that he does not trust extreme, cooked-up statistics. While I suspect that an analogous simulation would reveal Mo Vaughn (at his 2000 numbers) to be far more useful than Garret Anderson (at his own 2000 numbers), one must actually perform the simulation before the empirical proof is there.

Brian continues: "If Anderson had the same exact RBI and runs scored but had a .356 OBP (instead of .256) what would the Baseball Prospectus people say then? My guess is that he wouldn't be the most overrated player in the game. The next question becomes, can he post these numbers over a sustained period or is he just lucky this season?"

The first answer is that, if Anderson were getting on at a .356 clip, the Angels' offense itself would be far better than it is. Troy Glaus and Ben Molina would get the RBI and the recognition that is not yet available to them. Anderson would not be called "overrated" because he would not be overrated.

As for luck, one thing to get out of the way early: I do not know whether the home runs he has hit have been the result of luck or skill. Mike Scioscia has told the Los Angeles Times that wants Anderson to worry less about home runs than about making contact, a development about which I have mixed emotions. Scioscia seems to have discovered a problem but found entirely the wrong solution for it.

Let us take a batter's actual outcome (out, walk, single, whatever) to be skill, but the situations in which he bats (number of outs and men on base) to be luck. Critics of the second part would suggest that a good hitter bats in a good spot in the lineup precisely because he is a good hitter, but there is too much circular reasoning here to evaluate that claim fairly.

As it happens, a certain big-name sports web site gives player splits by base situation. Since that site competes with our new corporate overlord, I am reluctant to provide a link, but here's a hint: It's the one in Atlanta, not the one in Bristol.

Even more fortuitous is that Mike Scioscia's Angels have been unusually consistent with their batting order, at least in the middle of their lineup. From number three to six, fans in Anaheim have gotten used to seeing Mo Vaughn, Tim Salmon, Garret Anderson, and Troy Glaus, in that order. Here are their breakdowns (just try to beat this chart, Paisley!):

VAUGHN, Mo
Runners	PA	Out	BB+HP	1B	2B	3B	HR	RBI
None	182	116	21	22	9	0	14	14
1	65	39	9	10	4	0	3	7
2	35	20	3	10	0	0	2	10
3	16	7	5	3	0	0	1	8
12	28	18	3	4	2	0	1	10
13	15	11	0	4	0	0	0	8
23	5	5	0	0	0	0	0	0
Full	3	2	0	1	0	0	0	3
TOTAL	349	218	41	54	15	0	21	60
SALMON, Tim
Runners	PA	Out	BB+HP	1B	2B	3B	HR	RBI
None	181	110	25	27	11	0	8	8
1	69	42	11	7	4	0	5	11
2	25	16	4	2	1	0	2	6
3	9	4	5	0	0	0	0	0
12	21	16	1	4	0	0	0	3
13	21	11	8	2	0	0	0	5
23	11	5	5	0	0	0	1	4
Full	4	2	0	1	0	0	1	6
TOTAL	341	206	59	43	16	0	17	43
ANDERSON, Garret
Runners	PA	Out	BB	1B	2B	3B	HR	RBI
None	180	146	2	16	4	1	11	11
1	64	37	2	12	8	0	5	11
2	25	15	2	5	1	0	2	8
3	2	1	0	0	1	0	0	1
12	28	22	0	4	1	0	1	7
13	9	8	0	0	0	0	1	5
23	6	3	2	0	0	0	1	5
Full	14	11	0	2	0	0	1	9
TOTAL	328	243	8	39	15	1	22	57
GLAUS, Troy
Runners	PA	Out	BB+HP	1B	2B	3B	HR	RBI
None	205	117	33	23	15	0	17	17
1	38	24	6	3	3	0	2	5
2 &/or 3	72	43	13	10	2	0	4	25
Full	9	5	0	3	1	0	0	6
TOTAL	324	189	52	39	21	0	23	53

The first thing you'll notice is that I don't have full data for Glaus, thanks to an unexpected glitch at the site I used. Even from the data we have it's clear that Anderson gets somewhat better chances than Vaughn, and significantly better chances than poor Glaus. This is because Anderson not only fails to get on base for Glaus but also kills rallies before Glaus can bat.

Last but not least, here again are the vital statistics for the players at the heart of the Anaheim lineup:

	AVG	OBP	SLG	Runs	RBI
Vaughn	.292	.373	.545	47	60
Salmon	.269	.393	.507	55	43
Anderson	.241	.256	.500	43	57
Glaus	.305	.416	.636	57	53

If Anderson did his job, then Glaus would be a superstar.

about the author

Matt Bruce didn't get upstaged by the kids at school who ate paste, either. Send tips on getting decades-old permanent marker stains off your tongue at mb@strikethree.com.