The Joe Birmingham Mystery

Mr. Richmond has always attempted to craft a public image of SOM as embodying a vaunted statistical accuracy. Clearly their "Cadillac" re-created seasons show just how off their regular season statistical re-creations can be. But in the case of Joe Birmingham's 1906 season, SOM seems to have really dropped the ball.

His 3.8 M card is based on 41 ab's. His numbers, per SOM, were:
1906 CLE 10 41 5 13 2 1 0 6 1 0 2 0 .317 .333 .415 .748 3L 3.80M

Not sure why this card is in the set, in the first place, but the above stats are just wrong, at least according to Baseball Reference which gives the following stats for 1906:
1906 21 CLE AL 10 41 40 5 11 2 1 0 6 2 1 2 .275 .293 .375 .668

Oopsie!? I'm willing to take a wild guess that BBR is slightly more credible than SOM here. And 11 for 40 is very different than 13 for 41.

So, now that discrepancy has been made public, I fully anticipate that it will be summarily ignored by SOM and that any emails to blog boy on the topic will be disregarded entirely.

That said, why is this card in the set at all? It is the 1906 Gates Brown with defense.

Maybe Gates Brown is the 1968 Joe Birmingham without defense.

That's weird.

Birmingham's main claim to fame is the fact that running to second after a pickoff attempt he got hit in the head by a thrown ball and ultimately scored the only run in one of the greatest pitching games ever. (The detail about him getting beaned was provided by the winning pitcher himself, Addie Joss, who wrote a newspaper article about it.) One guy, reviewing "The Greatest Game Ever Pitched: Juan Marichal, Warren Spahn, and the Pitching Duel of the Century" by Jim Kaplan, describes the Joss game like this:

--------------------------

My own nomination for the best pitching duel wasn't even mentioned by Kaplan, and I don't know why. For one thing, apart from the two World Series games he covered, it was the only game that mattered in a pennant race, adding an element of urgency that was absent on July 2, 1963. I'm referring to the October 2, 1908 game between the pennant-chasing Chicago White Sox and Cleveland Naps, featuring future Hall of Famers Ed Walsh and Addie Joss. With a week left in the season, the Naps trailed the first-place Detroit Tigers by a half-game, with the White Sox one-and-a-half games out. The pressure was on both pitchers, and they were well-equipped. Joss came in sporting a 23-11 record and a 1.20 ERA. Walsh was even better, with a 39-14 record, including pitching both ends of a doubleheader just three days earlier.

Both pitchers were at the top of their game that day in Cleveland. The Naps scored an ugly run off Walsh in the third inning. Joe Birmingham led off with a single and Walsh picked him off, but Birmingham got in a rundown and the White Sox threw the ball away, allowing Birmingham to race to third base with nobody out. Walsh retired the next two batters with no damage, but an Ossee Schreckengost passed out enabled Birmingham to score. That was the only run Walsh surrendered; in eight magnificant innings, he yielded only four hits and struck out a season-high 15. It wasn't magnificent enough, because Joss pitched a perfect game.

Probably, the wrong stats are from an old data set, prior to researchers reviewing the newpaper boxes and so on.

the numbers work out if he had 42 PA (not 41), 41 AB (not 40), and 13 H (not 11) -- all the other numbers (extra-base hits and walks) can stay the same, and you get 317 333 415

so probably some researcher found out that one of his "hits" was, maybe, an error or FC, and another "hit" didn't occur at all (maybe a pinch-hitter got a hit, and Joe got credit)

I remember when Ty Cobb officially was credited with 11429 AB and 4191 hits -- numbers burned into my brain when I was a kid, but they were later corrected

A year or two ago, I emailed SOM to let them know that Jake Stenzel's numbers were slightly off, such that he now is credited with 600 plate appearances rather than the 597 reflected on his card. While the changes were minor compared to Birmingham, it affected his injury risk (which I would think would be much easier to fix from a programming standpoint). They acknowledged the discrepancy but wouldn't do anything to fix it.