Boxing Pandora

Today marks the thirtieth anniversary of the designated hitter. The occasion is celebrated with excessive tributes to Ron Bloomberg, who had the arbitrary brush with greatness of being the first DH. Here's yet another one:

After thirty years, the debate still rages as to whether or not the DH is good for baseball. In the last thirty years, there have been movements to abolish it and movements to make the rule universal throughout the majors. Today, baseball with its ever-more-nostalgic, revisionist history of the game now appears to be tended by those who are clamoring for the DH's demise. These are not just the purists-which I consider myself to be-but the media and the masses too are no against the "one-dimensional" player.

ESPN ran a poll on the issue last week and, though I did not see the final numbers, the anti-DH votes were leading 56% to 44%. So fans, or at least fans who have a propensity to vote on ESPN.com online polls, are against the rule.

ESPN also consulted their staff of experts (italics mine, not theirs) on the subject. It seems that the vote was two pro-DH and four anti-DH, if I read their responses correctly. Heck, four of them waffled more than Mike Dukakis on subject.

Here's my rundown from most pro-DH to most anti-DH in my estimation:

- Tony Gwynn: Gwynn is the most pro-DH of the six:

It's [the DH's] great for players who've had success in the game, but maybe can't take the grind of playing a full season at the end of their careers. It allows them to stay in the game -- and that's a good thing.

But still he admits:

Early in my career, I hated the designated hitter and thought baseball should get rid of it.

- Tom Candiotti: Candiotti, I put down as pro-ish. Here's why:

[B]ased on the way the game is played now I would like to see both leagues go to the DH. Let the hitters worry about driving in the runs, because the fans want to see scoring.

It's sort of a backhanded compliment, but I give it to the pro-DH camp.

Candiotti still displays curmudgeonly It-was-better-in-my-day-ism:

[T]he game I grew up with is gone...The emphasis in today's game is on offense rather than pitching and defense.

Oh, really? Is that why so many teams are trying to emulate the Rangers?

Pitching and defense are important, very important. They have improved greatly throughout baseball history and even in the twenty years since Candiotti made it to the majors. That baseball today can sustain 30 major-league pitching staffs, each consisting of about 12 pitchers at any given time is phenomenal. Sure, there are a lot of clunkers on almost any staff, but that a variety of pitchers with different handed-ness, styles, pitches, release points, and speeds are not only available on any given night but are used to the tune of between three or four a night (on average) is a great challenge for offenses. I cannot say with any degree of certainty that defenses are better on average than 20 years ago (at least without reams of research), but it seems logical to me given the historical trend in the sport. Pitching/defense today also has to overcome more difficult stadiums than in the bandbox days of the Seventies and Eighties.

Pitching and defense are as important as ever to winning but they are just one side to the coin with the other being occupied by batting and baserunning.

- Joe Morgan: Morgan, as always, is the least assertive. Maybe with this issue, it's to his credit. I put him down to con-ish. Joe sums up:

If it were my choice, I'd eliminate it [the DH], though it has served its purpose for the AL.

Here's another backhanded compliment, but still it's not enough for Joe.

Joe at his most negative on the topic goes a little bit like this:

I've never liked the DH because it makes a player one-dimensional.

But is a DH anymore one-dimensional than a pitcher who has no business hitting for himself-but I'm getting ahead of myself.

- Rob Neyer: Neyer unlike the rest starts out on a positive note:

I grew up with the DH.

I grew up with Hal McRae, the best DH (before Edgar Martinez, that is).

Then Neyer switches gears and comes out in against the DH. I put him in the con-ish camp:

So while it's been fun, and we'll always remember Hal McRae and Edgar Martinez fondly, 30 years is long enough.

Neyer's only argument against the DH is that "nobody needs help scoring runs any more" and "there are plenty of teams that don't care much whether their sluggers can actually play in the field without embarrassing themselves." Oh, and flannel unis are itchy, I think. OK, but I'm not sure that that gets to the root of the DH issue.

The last two arguments are the strongest anti-DH arguments and are also the most silly.

- Rob Dibble: Dibble blathers through a meaningless diatribe as only he could: via his own unique idiom that parallels English. Witness:

One of my former managers once was so caught up in a game that he inserted the pitcher into the leadoff spot and switched the replacement hitter/fielder into the old pitcher's slot in the batting order.

Ah, yeah, that's called a double-switch, and it's used to juggle batting orders late in the game. Maybe the 7-8-9 hitters were due up the next inning and didn't want to have to waste another pinch-hitter. If it's a close or tied ballgame, maybe your first three hitters get the run or runs you need and you win. It may be a gamble but it's a calculated one. You'd think Dibble would understand the strategy given he played most of his career in the then-DH-less NL and appeared in late innings.

Dibble then talks out of both sides of his mouth at once, and at each orifice espouses a silly notion as the basis for an argument:

If MLB officials are serious about speeding up games, they should realize that having weak-hitting pitchers bat three or four times a game would definitely move games along. It's time to dump the DH and make both leagues play with one set of rules.

The DH is about speeding up ballgames? I don't think so. The extra pressure on pitchers, the extra pitches expended, the extra relief pitchers used, etc.-don't you think that slows down a ballgame?

And why do both leagues have to play by the same rules? When I was a kid watching NL-only baseball in the suburbs of Philadelphia, I watched the AL on Saturday afternoon broadcasts with their different strike zones, different style of play, different umps (with those big chest protectors they held in front of themselves like shields), and a different rule, the DH. And I enjoyed it. The only issue about differences comes into play during interleague play time and the Series. My argument then is to eliminate the annoying and unfair interleague play and tinker with the Series rules.

- Jayson Stark: As far as I can tell Stark is the most anti-DH analyst on the list. He, as opposed to the slavering Dibble, at least formulates opinions based on fact. That they are the wrong opinions does serve to undermine his argument though. I'll take it point by point, but I won't reprint his entire point here; I'll just allude to it:

Point 1-It's absurd to play by different rules: Why? Again if this is solely about interleague play, by all means abolish the annoyance-interleague play that is. The DH was around for twenty years without interleague play. Did Stark hold a different opinion then? It seems so given his "intriguing" comment, so why is this a DH issue and not an interleague play issue?

Point 2-"The DH rule may have cost the Giants the World Series": Whoa, doggy! The Giants refused to use two decent bats as a DH in the World Series: Ramon Martinez and Damon Minor (Their leading DH during the season but not on the active World Series roster). Dusty Baker decided to use dreck like Feliz in game 7. He had used Dunston, Goodwin, and Shinjo prior to that.

The truth is that the Giants did not have much depth on their bench. They rebuilt their outfield during the year and ended up ignoring their bench. Their pinch-hitters had the second worst batting average in the NL (.196) and were last in OPS. Should the Giants lose because of a glaring weakness that they did nothing to fix? Well, yes. This is the World Series we are talking about after all. Oh and the decision to use Livan Hernandez in game 7 didn't help much either.

Point 3-A bastion for older players became a shelter for one-dimensional ones: No, it's used by different teams in different ways at different times. It can be used to ease an injured player back into the lineup, to load up extra bats in a lineup (e.g., Ray Durham on the A's last year), to give a younger player or a role player more exposure, to give a position player a rest from playing defense, and to give an older player an opportunity (e.g., Ellis Burks and Edgar Martinez). Teams vary their strategy over the course of the season as different conditions come about. To say that only one-dimensional players are used is over-generalizing and incorrect. Besides, I submit (again) that pitchers are one-dimensional players themselves.

Point 4-Fans come running for the smoky scent of runs: Stark argues that the DH generates no, or little, extra offense. But consider that AL DHs batted collectively .264 with a .788 OPS and average 23 homers per team (considering that they are used less than 162 games due to interleague play). The AL average was also .264 with an OPS 33 points lower (.755) and an average of less than 20 home runs per batting position. I don't have the numbers for the batting of NL pitchers but I would think it was considerable less effective.

Actually, I lied. I ran the numbers this morning. I included all pitchers who played no more than three non-pitching games (in order to filter out position players who pitched in lopsided ballgames and yet not overlook young pitchers used as pinch-runners, which happens occasionally). Here are the totals per league for 2002:

Lg   AB   R   H  2B 3B HR RBI SB CS  BB   SO IBB HBP  SH SF   BA  OBP SLUG  OPS
AL  281  14  38   5  0  0   9  1  0  12  121  0    0  25  1 .135 .170 .153 .323
NL 4864 276 709 128  7 23 287  1  2 176 1820  0   15 557 16 .146 .177 .189 .367

Stark's trying to say that those atrocious numbers did not affect the offense? C'mon. Given that AL pitchers fared ever so slightly worse than their NL counterparts with so much less exposure to batting, it makes a good argument for the universal implementation of the DH (the NL pitchers are able to hit the long ball once in a blue moon, however). Offensively, pitchers are just about at the lowest ebb possible for world-class athletes.

Point 5-Strategy and fancy ciphering: Look, I conquered double-switches playing APBA baseball at the age of ten. If I can do it, the average fan can, and maybe someday so will Bob Boone. It takes no baseball acumen to determine that when a .130 hitter is up and a man is on first with no outs in a tie ballgame, you should bunt. That's not strategy; it's rote. Bill James addressed exactly this issue in his original Historical Abstract and I will allude to James' work in my own take on the issue coming up next on the East Coast or following your local news on the West Coast.

- Mike: Let me begin by saying I am the sole person on the planet, apparently, who thinks that the system should continue just as it has, with DHs in the AL and pitchers hitting for themselves in the NL.

First, I begin my argument with James' analysis. The article to which I will refer is entitled "1973: DH Rule Increases Strategy" and comprises most of pages 257 to 261 in the 1988 paperback edition of The Bill James Historical Baseball Abstract-which I purchased used at $4.98 (the price tag's still there), probably the best baseball buy of my life (I also prefer the 1988 edition to last year's even with the addition of Win Shares).

James analyzes sacrifice hit and pinch hitter use as well as complete game totals. With each statistic, James calculates the average and standard deviation for all the teams in the league from 1968, five years before the advent of the DH, until 1986, the last year with data available at the time. He then compares the NL standard deviation to the AL. He finds in each case that the standard deviation increases in the AL in each case and is, on average, higher than the NL (it should also be pointed out that the AL had more teams than the NL from 1977-'86 which would tend to artificially reduce the standard deviation in the AL and yet they are still usually higher than the in NL).

James concludes that the DH has induced more strategy to the AL game:

What the DH rule actually does...is to eliminate from the game a series of forced, obvious moves, which involve in fact no option on the part of either manager, and thus no strategy. You've got a .113 hitter at the plate. A runner on first, and nobody out in the fourth, and you have to bunt don't you? Where's the strategy? With a DH up there at least you can do something. You're down four runs in the seventh with the pitcher leading off, and you have to pinch hit for him, don't you? What's strategic about that? The DH rule saves the pinch hitters, and thus in effect makes the roster larger. As such it creates, not eliminates, strategic options for American League managers.

God, I love that man. AL managers tend to use bunts and pinch hitters differently across the league. This creates an environment in which different strategies might be employed. Bunts are fewer and farther between in the AL due to the DH, but they are not uniformly fewer and farther between throughout the league. The same is true for pinch hitter use. AL managers are given the opportunity to make more personalized strategic moves.

Given that I watch a game to see something new as well as to see a good game of course. It's like going to the movies. How many times can you see the same Arnold Schwarzenegger film? (Maybe, that's not a good example given his track record.) I can understand why offbeat independent films are of more interest to the cinema cognoscenti and/or snobs (of which my wife accuses me of being a member). They at least contain something new or different. So, for me, the possibility that a manager may do something a little different to spur his team on is one thing that makes the game more entertaining. I realize that this may be an idiosyncratic point, but hey, I gotta be me.

I decided to re-run James data with numbers through 2002. Here's the sac. bunt data (for some reason my standard deviation across leagues is slightly different from his even though the average is the same; I'm not sure why-I let Access calculate mine):

Year	AL Avg	AL SD	NL Avg	NL SD	Ratio
1968	71.30	12.49	79.40	12.76	9.04%
1969	64.83	9.20	74.25	15.28	-31.02%
1970	67.58	20.82	68.25	15.63	34.50%
1971	73.67	15.88	76.50	15.20	8.52%
1972	74.42	11.20	72.08	17.15	-36.73%
1973	49.33	11.68	79.83	17.59	7.45%
1974	62.58	12.25	81.92	15.32	4.69%
1975	65.92	14.99	90.17	21.10	-2.78%
1976	68.17	15.51	81.25	20.42	-9.47%
1977	65.50	23.28	70.58	11.78	112.90%
1978	72.57	22.09	80.83	24.07	2.20%
1979	67.64	24.59	79.08	19.18	49.95%
1980	65.43	21.52	80.58	10.75	146.60%
1981	40.29	7.74	57.33	11.36	-3.07%
1982	54.43	21.46	81.50	13.37	140.39%
1983	45.71	11.17	76.75	7.84	139.09%
1984	44.71	12.32	67.42	14.99	23.93%
1985	46.29	19.26	75.08	15.91	96.33%
1986	46.14	17.17	72.42	17.73	51.99%
1987	45.14	11.75	68.58	11.95	49.33%
1988	49.43	12.71	78.17	16.75	19.95%
1989	51.93	15.67	74.92	11.90	89.99%
1990	48.79	15.93	73.00	14.55	63.89%
1991	52.36	10.76	74.25	15.38	-0.79%
1992	48.71	12.63	81.92	12.93	64.23%
1993	50.07	16.92	79.29	15.82	69.35%
1994	32.07	10.83	54.14	10.06	81.75%
1995	38.64	12.68	67.64	12.99	70.77%
1996	40.71	11.55	69.57	11.95	65.14%
1997	39.21	10.30	73.93	11.60	67.36%
1998	38.43	11.96	72.94	14.30	58.70%
1999	36.21	8.54	68.63	14.58	11.02%
2000	40.43	14.11	66.38	14.21	63.03%
2001	38.07	11.30	67.13	18.03	10.49%
2002	35.64	11.28	70.88	15.74	42.52%
Avg	52.35	14.39	73.90	14.86	36.65%

The ratio represents the AL standard deviation represented as a function of the AL average divided by the NL standard deviation represented as a function of the NL average. It's a ratio of what I believe is called Coefficient of Variation in the Statistics world. [For those concrete types: Ratio=(AL SD/AL Avg) / (NL SD/NL Avg)] Basically, a positive ratio means that the AL managers are using bunts in different ways throughout the league and a negative means that they are not. (James does assume that because the DH is not, generally-remember interleague play-, used in the NL, the strategic use of the bunt has not varied greatly over the last 35 years. This may or may not be true, but it does seem logical given the similarity across time in the NL bunt averages and standard deviations.)

So the ratio tells us that as James indicated, the AL and NL used bunts similarly until 1972, and then the DH suppressed bunting in the AL almost uniformly for about five years. But since 1977, the AL has been employing the bunt in a more varied fashion than the NL.

I decided to change James' last two parameters. The number of pinch hitters used per team is interesting but a) it's hard to derive from the statistical record (or I would have to go back through my old TSN Baseball Guides and I'm too lazy) and b) I think that there is a different metric that tells a more complete story. What we are interested in are the number of changes to the set lineup either in the batting order or defensively in the field. Pinch hitters only tell part of that story. There are replacements that play the field first and therefore, are never registered as pinch hitters. This is especially true of the dreaded double-switch. So I thought looking at the way that a manager uses his players in the lineup would be better expressed as a function of the number of players used per game. I derived this from the number of total games played for all players per team divided by the number of games played by the team. Again I took the average and the standard deviation across leagues:

Number of players used per game:
Year	AL Avg	AL SD	NL Avg	NL SD	Ratio
1968	12.93	0.66	12.29	0.32	97.36%
1969	13.12	0.69	12.68	0.48	38.55%
1970	13.09	0.57	12.95	0.46	23.15%
1971	12.64	0.62	12.63	0.35	76.05%
1972	12.79	0.53	12.53	0.47	8.95%
1973	12.89	0.69	12.92	0.48	43.57%
1974	13.07	0.53	13.26	0.54	-0.89%
1975	13.12	0.69	13.08	0.38	82.36%
1976	12.99	0.44	13.01	0.50	-11.48%
1977	13.13	0.65	13.42	0.39	68.00%
1978	13.03	0.79	13.17	0.48	65.55%
1979	13.10	0.59	13.17	0.37	62.41%
1980	13.38	0.55	13.40	0.52	6.21%
1981	13.40	0.55	13.69	0.54	3.28%
1982	13.24	0.53	13.40	0.47	14.36%
1983	13.30	0.68	13.45	0.45	51.83%
1984	13.64	0.60	13.49	0.47	26.62%
1985	13.66	0.54	13.46	0.40	35.22%
1986	13.40	0.47	13.75	0.48	0.96%
1987	13.50	0.57	13.74	0.41	42.90%
1988	13.35	0.41	13.35	0.44	-6.01%
1989	13.51	0.48	13.62	0.42	13.26%
1990	13.85	0.51	13.79	0.38	31.45%
1991	14.08	0.65	13.97	0.39	65.40%
1992	13.73	0.50	14.00	0.46	10.73%
1993	13.78	0.47	13.91	0.36	32.52%
1994	13.67	0.39	13.87	0.25	57.60%
1995	14.07	0.39	14.20	0.35	11.14%
1996	14.06	0.38	14.19	0.32	21.13%
1997	12.91	0.64	14.19	0.33	114.65%
1998	12.73	0.51	13.92	0.43	30.08%
1999	12.74	0.50	14.19	0.43	30.70%
2000	11.30	0.41	14.11	0.33	55.01%
2001	11.01	0.26	14.16	0.31	11.14%
2002	11.30	0.52	14.23	0.27	137.85%
Avg	13.13	0.54	13.52	0.41	35.28%

The same pattern develops as with bunts except that the AL had more variation among their teams as far as the number of players used even before the DH. I'm not quite sure why. (Note that the NL has more pitching changes but that "noise" is eliminated by comparing variation across leagues: if all teams make similar pitching changes, it won't affect the standard deviation.)

Lastly, instead of complete games which are dwindling into oblivion as we speak, I instead used number of pitchers per game, which I feel gets to the same basic concept, but is more pertinent to today's game.

Number of pitchers used per game:
Year	AL Avg	AL SD	NL Avg	NL SD	Ratio
1968	2.57	0.27	2.38	0.25	-0.78%
1969	2.71	0.25	2.49	0.31	-24.78%
1970	2.74	0.18	2.58	0.21	-19.13%
1971	2.55	0.26	2.43	0.15	62.80%
1972	2.47	0.27	2.44	0.18	44.87%
1973	2.16	0.21	2.58	0.17	47.16%
1974	2.16	0.14	2.64	0.31	-45.08%
1975	2.19	0.20	2.60	0.22	10.72%
1976	2.22	0.20	2.60	0.29	-20.73%
1977	2.31	0.21	2.78	0.28	-10.77%
1978	2.22	0.21	2.62	0.22	9.35%
1979	2.34	0.13	2.73	0.17	-15.72%
1980	2.39	0.22	2.77	0.17	48.15%
1981	2.44	0.25	2.93	0.29	6.38%
1982	2.46	0.15	2.81	0.19	-11.07%
1983	2.44	0.19	2.80	0.20	11.28%
1984	2.53	0.24	2.80	0.19	37.71%
1985	2.65	0.25	2.83	0.19	37.64%
1986	2.66	0.18	2.95	0.20	-1.07%
1987	2.73	0.22	3.08	0.20	26.01%
1988	2.65	0.16	2.85	0.21	-15.34%
1989	2.81	0.16	2.95	0.21	-21.83%
1990	2.97	0.16	3.07	0.19	-13.69%
1991	3.11	0.23	3.16	0.12	97.30%
1992	3.06	0.23	3.25	0.20	21.01%
1993	3.19	0.21	3.35	0.26	-14.86%
1994	3.26	0.25	3.39	0.25	3.81%
1995	3.31	0.22	3.59	0.29	-17.98%
1996	3.36	0.22	3.52	0.19	24.19%
1997	3.46	0.21	3.54	0.21	2.62%
1998	3.49	0.18	3.43	0.21	-19.33%
1999	3.52	0.21	3.59	0.21	0.26%
2000	3.52	0.22	3.57	0.20	10.49%
2001	3.51	0.21	3.73	0.20	8.73%
2002	3.45	0.32	3.78	0.17	100.60%
Avg	2.79	0.21	2.99	0.22	5.30%

I'm not sure if this supports James' theories or not. It fluctuates wildly but on average the AL has slightly more variability. Well, two out of three aint bad to quote Mr. Loaf.

James concludes:

I'm not an advocate of the Designated Hitter Rule; I'm only an advocate of seeing the truth and telling the truth. What the truth comes down to here is, a question of in what does strategy reside? Does strategy exist in the act of bunting? If so, the Designated Hitter Rule has reduced strategy. But if strategy exists in the decision about when a bunt should be used, then the DH rule has increased the differences of opinion which exist about that question, and thus has increased strategy. But if strategy is an argument, then I would argue that there is more of a difference of opinion, not less, in the American League.

I am inclined to agree with him.

Look, the purist in me hates the DH, too. But is that hatred logically based or is it just a knee-jerk reaction to a change in the age-old game? It's inelegant, but so is a Roger Clemens bunt attempt. So how do I reconcile this apparent problem of cognitive dissonance of a purist-me-who accepts the DH?

Here's an analogy that may help-it helps me. In Phsyics, there are instances in which a particle is seen as a wave and a wave is seen as a particle. Things have substance or do not dependent on the situation. And it works. Maybe the DH has merit even in a purist's game.

I say keep the status quo. Why? Because the NL is never going to adopt the DH with the derisive view that the media, the fans, and the analysts have of the rule.

The AL is more problematic. If one cannot accept James' theories, then there are practical issues preventing the demise of the DH. First, the players' union will not sit idly by as fourteen starting jobs, i.e., one DH per AL team, are irradiated.

Even if the union is bought out with expanded rosters (doubtful) or salary increases across the board (still more doubtful), there's the problem of what to do with the ex-DH's. So Edgar Martinez, Ellis Burks, and their ilk are dismissed. It's highly unlikely that these players can take the field defensively.

Well, maybe we have no problem with that. They will be missed but progress, er, rather regress, is in motion. How about pitchers batting for themselves? Given one spring training they should be able to catch up with their NL counterparts. OK.

But what about AL managers getting used to the change in strategy? I guess an offseason of Strat-O-Matic will cure them. Hey, Bob Boone switched leagues, didn't he?

There are a number of practical issues. They opened Pandora's box and it's going to be tricky to get all of the baseball "ills", that the DH created back in that box. They are not insurmountable problems but they are all trade-offs. What happens the first time a star AL pitcher is injured when he is at bat? When a Lou Piniella fails to make a proper double-switch or the like? When an Edgar Martinez is forced to play the field and is injured?

If the AL fans really want to see more .130 hitters, more power to them. Just don't say that it improves play, induces strategy, or eliminates one-dimensional players.

Do it because it's a matter of preference. That's really the only excuse.