I think it might be fair to say that Joe DiMaggio has never been a favorite of the sabermetric community. DiMaggio was the hero of the New York sportswriters of the postwar generation (postWorld War II.) That crowd overlooked his human failings, romanticized his leadership qualities, and would insist with a straight face that DiMaggio never made a base running mistake in his career.
They also disliked numbers. It wasn’t that they rejected statistical analysis; there wasn’t any statistical analysis in their generation that was worthy of the name. They disliked statistics. It was fundamental to their world view that Joe DiMaggio was a far greater player than Ted Williams. If it was necessary to site statistics to prove this, they would site them. If it was necessary to dismiss statistics to prove this, they would dismiss them. Most often, it was necessary to dismiss them.
When statistical analysts came along, we tended to be Ted Williams men—or, even worse, Mickey Mantle men. To the DiMaggio sportswriters, suggesting that Mantle was a greater player than DiMaggio was sacrilege. We sacrileged away. They saw Mantle as an uppity young’un, and saw us as the irresponsible advocates of the uppity young’un. We just didn’t know how great Joe was, because we never saw him play.
Missing in all this is that Joe DiMaggio was in fact a very great player—offensively and defensively—and that, on a superficial level, the stats do in fact understate his impact. Among all the great players in history, I doubt there was any who was hurt as much by the parks he played in as was DiMaggio. DiMaggio in his career hit 148 home runs at home, 213 on the road. He hit .315 in his home parks (park—Yankee Stadium), vs. .333 on the road. He hit more doubles on the road. He drove in almost a hundred more runs on the road (818720).
In addition to this, DiMaggio lost three prime seasons to World War II, seasons in which he might very reasonably have put up the best numbers of his career. If DiMaggio hit 46 homers and drove in 167 runs when he was 22 years old—which he did—might he not very plausibly have hit 55 homers and driven in 180 when he was 28, 29 or 30? And that is without adjusting for the park. When DiMaggio hit 46 homers and drove in 167 runs in 1937, he hit 27 homers and drove in 87 runs on the road. Adjust that for the park, you get some big numbers.
I decided to recreate DiMaggio’s career batting records on the following assumptions:
1) DiMaggio plays not in Yankee Stadium, but in a "neutral" park, neither particularly helpful to his numbers nor particularly harmful to them, and
2) DiMaggio plays through the war.
This leaves several issues. We could project DiMaggio’s career numbers if he had played in St. Louis’ Sportsman’s Park, which was the best park in the league for him. In 127 career games in St. Louis DiMaggio hit .389 with 45 homers, 156 RBI, also 45 doubles, a slugging percentage of .759. I decided not to do that. We could also give him back some games he missed with injuries; since we’re operating on a counterfactual set of assumptions, we could at the same time throw in "what if he hadn’t missed so many games with injuries?" I decided not to do that.
I did, however—since this is essentially a proDiMaggio operation—give him a couple of other breaks. First, I decided that he would continue to play for the Yankees. It makes a difference; DiMaggio drove in phenomenal numbers of runs in part because he played for Yankee teams that were always first or second in the league in On Base Percentage. DiMaggio also had very high ratios of plate appearances per game. In his career he averaged 4.42 plate appearances per game, as opposed to 4.27 for Ted Williams, 4.18 for Willie Mays, 4.13 for Mickey Mantle, 4.23 for Hank Aaron, 4.20 for Stan Musial, and 2.86 for Aaron Ledesma. Compared to Mantle, Mays, Aaron or Musial, DiMaggio got an extra 30 plate appearances a year because the Yankees could sustain an offense.
In theory, if the Yankees had played in another park, their ability to sustain an offense might have been even greater, and DiMaggio might have had even more plate appearances per game. I decided not to give him even more, but I decided to allow him to keep the 4.42 plate appearances per game, which is more than any other comparable hitter got. Also, I decided to give him one extra year, on the speculative logic that, playing in another park, DiMaggio would have had better numbers, thus would have been under less pressure to retire in 1951, when his numbers took a sudden downturn. I let him play through 1952. This still gives DiMaggio only a 17year career, which is still a fairly short career. Willie Mays, missing a season to military service, still played in 22 seasons (19511973), Stan Musial the same (19411963), while guys like Cobb, Aaron, Yastrzemski and Brooks Robinson played 23 or more.
OK, now we have to decide what to give DiMaggio on a peratbat basis. We can start with DiMaggio’s career batting stats on the road:
G

AB

R

H

2B

3B

HR

RBI

SB

CS

BB

SO

BA

OBP

SLG

OPS

856

3461

742

1154

203

58

213

818

16

6

392

206

.333

.405

.610

1.015

We could use that to represent his neutralpark batting stats, and, actually, we will circle back to numbers very similar to those at the end of the process, but there are a bunch of other things to worry about. Let’s assume that an average park has a park factor of 1.00, and thus that the eight parks in the American League in DiMaggio’s time had a total park factor of 8.00. The Yankee Park Factor in DiMaggio’s years was usually around .86, which means that the other seven parks totaled up to about 7.14. That means that the average "road" park for DiMaggio had a Park Factor of about 1.02, meaning that it inflated offense by about 2%. We’re giving DiMaggio a logically indefensible little break by assuming that his "road" park is a neutral park. We have to factor his ability to hit in Yankee Stadium back into it.
We do that by dividing his Yankee Stadium performance by 7, and adding that back into his "road" performance, as if Yankee Stadium was his eighth road park. This produces the following batting line:
G

AB

R

H

2B

3B

HR

RBI

SB

CS

BB

SO

BA

OBP

SLG

OPS

982

3941

835

1305

230

68

234

921

18

6

449

229

.331

.404

.602

1.006

OK, so that’s parkneutral batting for Joe DiMaggio; his batting average drops to .331, but his OPS is still over 1.000. The only real players in major league history who have a career OPS of 1.000 are Babe Ruth, Ted Williams, Lou Gehrig, Barry Bonds, Jimmie Foxx, Albert Pujols, Hank Greenberg, Rogers Hornsby and Ed Sanicki.
At this point, however, we have DiPrived DiMaggio of his home field advantage. This is what DiMaggio should hit in a neutral road park (mostly. Technically, he still has a home field advantage in 1/8^{th} of his games.) We have to give him back his home field advantage.
In DiMaggio’s era the home team typically hit about 12 points higher than the road team. If DiMaggio was at home in half of his games, that would be 6 points. Since he already has some little remnant of his home field advantage embedded in the stats we showed above, we’ll improve him by five points.
We can improve his average by five points by removing 67 at bats from the batting line shown above. Since DiMaggio had 4.42 plate appearances per game, that requires that we also remove 15 games. Also, we’ll assume that of the 67 batting outs removed from DiMaggio’s line, 5 were strikeouts. That makes DiMaggio’s batting line as follows:
G

AB

R

H

2B

3B

HR

RBI

SB

CS

BB

SO

BA

OBP

SLG

OPS

967

3884

835

1305

230

68

234

921

18

6

449

224

.336

.410

.611

1.021

So that’s what we’ll allow DiMaggio to do per plate appearance, and don’t worry, I’m also tracking the stuff like Hit By Pitch and Sacrifice Bunts, I’m just not showing you that data.
We can break this down into eleven outcomes, which are as follows:
Strikeout
Hit Batsman
Walk
Home Run
Triple
Double
Single
Sacrifice Bunt
GIDP
Ground Ball Out
Fly Ball Out
My task is actually a little more complicated than this because I have to get RBI and Runs Scored right, as well. In order to do that, I have to make a certain number of the singles RBI singles and a certain number of them nonRBI singles, a certain number of them lead to runs scored and a certain number not lead to runs scored, a certain number of tworun homers and a certain number of grand slam homers, etc. That gets boring, so I’ll condense this a little bit.
Per 10,000 plate appearances, our parkneutral DiMaggio would have 513 strikeouts, 67 hit by pitch, 1,031 walks, 537 home runs, 156 triples, 528 doubles, 1,774 singles, 11 sacrifice bunts, 179 double play grounders, 2,512 other ground ball outs, and 2,692 fly ball outs.
Except that our reconstructed DiMaggio does not have exactly 10,000 career plate appearances. How many does he have?
DiMaggio had 7,673 plate appearances in real life. Let’s say that he played 150 games a year during the war (19431945) and that he played another 120 games in 1952; we’re giving him an extra 570 games. At 4.42 plate appearances per game, that’s another 2,519 plate appearances. That’s 10,192.
OK, well. . .we’ll adjust that and show you the totals later. My next task was to create a spreadsheet on which all of these events were listed one at a time—the right numbers of homers, the right number of doubles, the right number of strikeouts, etc. That was easy. Then I had to try to sort those so that they represent what Joe DiMaggio’s career might plausibly have looked like, if
1) He had not played in the toughest park in the league for a righthanded power hitter, and
2) He had been able to play through the years 19431945 without DiSruption.
Here’s what I came up with: