Historic comparables (Volume 1)

In Bill James’s original Historical Baseball Abstract (published in 1985), he did something pretty cool. Well, actually he did a lot of pretty cool things, but one in particular has always stuck with us: When writing about old-time ballplayers, he identified a more recent player who was comparable. This was an exceptionally effective device at bringing a stat page to vibrant life, as well as seeing a familiar modern player in a new light.

But it’s been almost a quarter-century since James came out with that book, and we thought it would be fun to revisit that concept. We’ve come up with a list of players from ages past—generally speaking, pre-World War II—and found the modern players who most closely resemble the old-timers.

To help us in our quest, we created a similarity scores system to compare players across eras. For details, see the References and Resources section below, but essentially, we looked at six quantifiable factors:

{exp:list_maker}Games Played
Win Shares (WS)
Win Shares Above Bench (WSAB)
Win Shares Above Average (WSAA)
All-Star Win Shares (WSAS)
Position {/exp:list_maker}
Also, we considered defensive ability, using the defensive grades in James’s Win Shares book. But we should stress, we used this system only as a guide to narrow the field. More often than not, we chose a player that the system considered reasonably comparable, but who was not the No. 1 man on the list. More than just statistics, we’ve considered things like career patterns, physical similarities, personality traits, and other factors. The goal is to find not merely the most statistically-similar player, but the most similar player, period.

With that introduction out of the way, let’s dive in.


George Sisler and Don Mattingly

Player                 G      WS     WSAB    WSAA    WSAS     Pos    Def
George Sisler        2055     292     131      62      24      12     C-
Don Mattingly        1785     263     123      63      15      12     A-

Sim Score:  878

First, the obvious: Both Sisler and Mattingly were batting champs and MVPs who played first base. Both players were major stars in their 20s, Sisler from 23 to 29, Mattingly from 23 to 28. But both suffered rapid declines: Sisler missed all of his age-30 season with severe sinusitis, and when he returned, he was a shadow of his former self. Mattingly, meanwhile, suffered from chronic back problems, and his production collapsed after age 28.

The big difference, of course, is that Sisler breezed into the Hall of Fame, while Mattingly has almost no chance of being elected. Some of that, of course, is just due to context: Sisler played in an era when the average hitter batted .288, while the league average in Mattingly’s day was .261. But Sisler, at his best, really was a little bit better than Mattingly. For what it’s worth, Sisler was elected to the Hall of Merit, and Mattingly wasn’t.


Chick Hafey and Kevin Mitchell

Player                 G       WS    WSAB    WSAA    WSAS     Pos    Def
Chick Hafey          1283     186      94      54      15      48      C
Kevin Mitchell       1223     178      91      54      17      48      D

Sim Score:  970

Hafey is one of the many Hall of Famers from the 1920s/30s who were voted in by the Veterans Committee in the 1970s, and who have no business being there. But that isn’t Hafey’s fault, and though he wasn’t genuinely of Hall of Fame quality, he was one heck of a ballplayer.

Hafey falls short of normal Cooperstown standards not for any lack of peak production, but because of major issues with durability and career longevity. When in the lineup and at his best, Hafey was a top-tier cleanup hitter, delivering abundant power with a high average. All of this, of course, also perfectly describes Mitchell.

Hafey and Mitchell were both right-handed. Neither was especially tall (Hafey was listed at 6-foot-even, Mitchell at 5-foot-11), but both were very muscular. Both ran better than most guys of their size, though neither could be described as “fast.” Hafey was reputed to possess a terrific throwing arm, though I’ve always been a little skeptical of that, given that he was primarily deployed in left field. Mitchell, whom I watched a whole lot, most assuredly did possess a terrific throwing arm, when a young player; he was of course a third baseman (and even occasionally a shortstop!) before settling in left field.

And if all this isn’t enough similarity for ya, then how about this: Both Hafey and Mitchell had cousins who also played in the major leagues.


Using Recurrent Neural Networks to Predict Player Performance
Technology is rapidly advancing possibilities in decision-making.

Chuck Klein and Juan Gonzalez

Player                 G       WS    WSAB    WSAA    WSAS     Pos    Def
Chuck Klein          1753     238     111      56      10      48     C-
Juan Gonzalez        1689     234     104      49       8      48     D-

Sim Score:  956

In the eyes of the average fan, Klein and Gonzalez in their primes were both all-time greats. Juan Gone won two MVP awards; Klein had one trophy and two second-place finishes. Both men were home run champions and RBI machines. Both struggled with injuries, and both were finished as stars by their early 30s.

Interestingly, neither player won the MVP in his best season. Despite winning the Triple Crown in 1933, Klein was the MVP runner-up, and while Gonzalez’s best overall season was 1993, he finished fourth in the voting that year.

Another characteristic the two have in common? Neither was as good as their Triple Crown stats would have us believe. Yes, Klein was a career .320 hitter … but he did that in the high-offense 1930s, and more importantly, his best years came in Baker Bowl, a hitter’s paradise matched only by the Rockies’ ballparks before the humidor. While Klein had some gaudy numbers, he was really no better a hitter than contemporaries like Dolph Camilli and Bob Johnson. In other words, he was a very good player, but a marginal Hall of Famer.

As for Gonzalez, even if we target his best years (1991-2001), his OPS+ ranks just 13th in baseball. Like Klein, he benefited from playing in a hitters’ park in a hitters’ era. He didn’t walk much and had virtually no defensive value, and he didn’t deserve either of his MVP awards.


Edd Roush and Jim Edmonds

Player                 G       WS    WSAB    WSAA    WSAS     Pos    Def
Edd Roush            1967     314     166     103      35      48     A-
Jim Edmonds          1925     301     159     100      35      48     A+

Sim Score:  964

The differences between these two are obvious: Roush was a high-average contact hitter with moderate power, who rarely walked, while Edmonds walked a lot, struck out a lot, and hit for outstanding power. This dramatic difference in offensive profile had always blinded me from considering these two side-by-side, indeed until undertaking this exercise.

But while they aren’t quite peas-in-a-pod, the similarities between Roush and Edmonds are remarkable. Each batted left-handed and threw left-handed. Despite their divergent approaches at the plate, the bottom-line run production between the two was very close; Roush’s career OPS+ was 126, Edmonds’s 132, over careers of nearly exactly equal length. And, of course, each was among the most celebrated defensive center fielders of his day, and may be legitimately considered among the best of all time.

Roush is in the Hall of Fame; He was a Veterans Committee selection from the 1960s, and is among those they’re generally considered to have gotten right. He was elected to the Hall of Merit as well. How Edmonds will fare when he becomes eligible for Cooperstown is a very interesting question.


Al Lopez and Bob Boone

Player                 G       WS    WSAB    WSAA    WSAS     Pos    Def
Al Lopez             1950     173      46      -9       0     240      B
Bob Boone            2264     210      54     -13       1     240      B

Sim Score:  880

In 1945, Al Lopez set a new career record for games caught. He held that record for more than four decades, longer than any other player. And the man who broke the record? Bob Boone, in 1987.

They’re similar in more ways than that, though. While both had some decent offensive seasons, neither were very good hitters, and they finished with near-identical rate stats (.261/.326/.337 for Lopez, .254/.315/.346 for Boone). They were both pretty solid catchers, though neither was spectacular (James gives both of them B grades for defense). And of course, both had very long careers, 19 seasons apiece.

And when they were done playing, both Lopez and Boone became managers. It’s at this point that their paths diverge. Lopez won more than 1,400 games (never once finishing under .500) and made it into the Hall of Fame (as a manager). Boone, on the other hand, won just 371 games, and his teams never finished over .500.


Larry Gardner and Robin Ventura

Player                 G       WS    WSAB    WSAA    WSAS     Pos    Def
Larry Gardner        1923     258     117      57      10      84     B+
Robin Ventura        2079     272     120      56       7      84     A-

Sim Score:  940

Left-handed-hitting third basemen are a rarity, of course, and we’ve explored previously how these two rank high among the very best of this unusual breed.

Neither Gardner nor Ventura was quite a star hitter, but each was a sound, consistent producer with the bat: Gardner’s career OPS+ was 109, Ventura’s 114. Quite like Roush and Edmonds, each of these two delivered his better-than-average offense in the mode typical of his era, Gardner as a contact-focused line-drive hitter, Ventura as an uppercutting power hitter a bit prone to the strikeout.

Both Gardner and Ventura were considered among the best-fielding third basemen of their time: Ventura won six Gold Gloves, and had the award been around in Gardner’s day he surely would have bagged several as well. Both were quiet, steady “solid pro” types, whose all-around talent was geared more toward winning ball games than grabbing headlines. Both changed teams a couple of times in mid-career, helping to lower their profile within the lore of any particular franchise.

References & Resources
The similarity scores used above are calculated as follows:

Factor 1: 0.2 times the absolute value of the difference between each player’s career Games Played.

Factor 2: 1 times the absolute value of the difference between each player’s career Win Shares (WS)

Factor 3: 1.33 times the absolute value of the difference between each player’s career Win Shares Above Bench (WSAB).

Factor 4: 1.66 times the absolute value of the difference between each player’s career Win Shares Above Average (WSAA).

Factor 5: 3 times the absolute value of the difference between each player’s All-Star Win Shares (WSAS).

Factor 6: 1 times the absolute value of the difference between each player’s Position value. Position values are as follows (taken directly from Bill James’ original Similarity Scores system, which is also used at Baseball-Reference.com):

C: 240
1B: 12
2B: 132
3B: 84
SS: 168
OF: 48

Next, we simply add together all six factors and subtract from 1000. The higher the score, the higher the similarity.

Finally, a very brief explanation of the various forms of Win Shares used here: One Win Share equals 1/3 of a win. The methodology for calculating Win Shares is outlined in the book Win Shares, by Bill James and Jim Henzler. Win Shares Above Average (WSAA) compares a given player’s Win Shares to an average player, given the same amount of playing time. Win Shares Above Bench (WSAB) compares a given player’s Win Shares to a “bench” player (i.e. 0.7 times an average player); All-Star Win Shares (WSAS) compares the player to an “All-Star” (i.e. 1.5 times an average player). WSAA and WSAB can be either positive or negative; WSAS is counted only when it is positive.

Print This Post

Comments are closed.