All the Record Books Are Wrong, Part 3
Continuing from Part 2 (at http://goo.gl/J6HRsk), Frank Williams brings it all home. His effort was a personal triumph, of course, but also reflected credit on the Society for American Baseball Research, which by its example nurtured such work. It seems beyond belief today, but the force of Williams’ conclusions compelled the three print encyclopedias–Macmillan’s Baseball Encyclopedia, Neft-Cohen’s Sports Encyclopedia: Baseball, and the Thorn-Palmer Total Baseball–as well as the Elias Sports Bureau and Baseball-Reference.com to conform their records, almost without exception, to his findings.
This discussion didn’t change any of the practices, but in later Spalding publications–How to Score a Baseball Game, J. M. Cummings (1913) and the 1917 Baseball Guide–John Heydler tried to set standard practices for his official scorers to follow. The American League published nothing on this subject during the period.
The existence and the consistent application of these practices during the 1901-19 era demonstrate that there should have been no changing of pitching won-lost records, except for out-and-out errors on the official sheets, by I.C.I. and Macmillan in 1969. Later editions of the Macmillan Baseball Encyclopedia came out under a new editor, Joe Reichler, and many changes made for the 1969 edition, particularly those of Hall of Famers, reverted to what they were supposed to be.
The 1982 edition states on page 2237, “It was decided that all pitching decisions during the period 1901-1949 shall stand as they are in the official records”–the same wording which has appeared in all editions but the first. The Neft-Cohen encyclopedia also tried to go back to the correct pitchers’ won-lost records for 1901-19. Yet the current editions of both books differ from each other, and neither agrees completely with the official records.
Macmillan did not switch all the pitchers’ won-lost decisions back to agree with the official sheets, and even when the switch was made, it was not always a complete job. A good example of this is Smoky Joe Wood’s pitching record. I was glad to see that Macmillan corrected his lifetime record to 116-57 from the 114-69 they had shown previously (11 of Wood’s 12 losses in the minors in 1908 had somehow found their way into his major-league totals), but the editor did not change Wood’s won-lost marks in relief or his number of saves. When I did Wood’s day-by-day record, I not only found an extra win for him, but also emended his relief totals. Macmillan should be showing Wood with a lifetime mark of 19-8 in relief with 11 saves, but instead, they are showing him at 15-9 with 17 saves.
When Wood’s 1911 record was changed back to 23-17 from the 21-17 cited in the 1969 edition, Larry Pape’s record should have been reduced by two victories, but it wasn’t. Even by doing that, however, Macmillan wouldn’t get Pape’s record straight because another win it has given to him should be transferred over to Ray Collins! Pape should be 10-8 in 1911, and Collins 11-12. If you add up all the Red Sox pitchers’ wins for 1911 in Macmillan, you will get 80; Boston won only 78 games. This is by no means an isolated instance of halfhearted, unreconciled tinkering.
The reason that Neft-Cohen doesn’t agree with the official sheets is that it is relying heavily upon American League won-lost decisions from the Spalding and Reach Guides for the period 1913-19, which are mostly unofficial records. Even some of the years prior to 1913 do not match up to the official sheets; the season of1915 provides a good example of some of these differences. Examining the Red-Sox won-lost marks, we see Rube Foster at 20-8 when he sould be 19-8; Babe Ruth at 18-6 when he was really 18-8; Wood at 14-5 rather than 15-5; Leonard at 14-7 (should be 15-7); Vean Gregg at 5-3 (correctly 3-2); Collins at 5-7 (correctly 4-7); Mays at 4-6 (correctly 6-5); and the aforementioned Ralph Comstock at 2-0 rather than 1-0. And this mess all arises from one team in one year.
Although the 1982 editions of both encyclopedias disagree on some of the yearly records of Cy Young, they are in accord when it comes to his grand totals of 511 wins and 313 defeats. The Hall of Fame Fact Book and the 1982 Macmillan agree completely on Young from 1890 to 1901, and I agree with them that this is his correct record for those 12 years. They also concur on the 1902 season in showing Young with a 32-10 record, but here I agree with Neft-Cohen, which gives Young a mark of 32-11.
I did Young’s 1902 season game by game and it is impossible to come up with any other record. Cy pitched in 45 games, of which 43 were starts, 41 of these complete, and two games were in relief where he had no record. His record in complete games was 32-9 (including a forfeit game of eight innings). The other two starts were incomplete games which were losses beyond a doubt. They are as follows:
May 2, 1902; at Boston
Baltimore 6 2 0 2 0 4 0 0 0 – 14
Boston 0 4 0 0 0 0 0 1 1 – 6 Young, 1 inn., LOST; Prentiss, 8 inn.
August 7, 1902; at St. Louis
Boston 1 0 0 0 0 0 0 1 0 – 4 Young, 1 inn., LOST; Sparks, 7 inn.
St. Louis 6 0 0 1 0 3 0 2 x – 12
There is no doubt after looking at these two games, in each of which Young allowed six runs, that he was 32-11 in 1902. Pitching practices both then and now would charge Young with these defeats. Interestingly, I obtained this information from the Macmillan reconstructed sheets which are housed in the Hall of Fame Library (remember, there were no official American League sheets for 1902). The 32-11 record that I.C.I. originally compiled in 1967 and which was printed in the first edition of Big Mac was correct; now Macmillan lists an “improved” record.
For 1903, both the 1982 Macmillan Encyclopedia and The Hall of Fame Fact Book show Young at 28-10, but this too is wrong. The Spalding and Reach Guides containing the American League’s official won-lost records show Young at 28-9. The previous Macmillan edition listed Young correctly at 28-9, and by changing this to 28-10 Macmillan has taken away the won-lost-percentage championship which is rightfully his.
Cy Young was in 40 games in 1903, of which 35 were starts–all but one complete–and five were in relief. He was 26-8 in complete games, and 2-1 in relief with two saves. The one start which did not affect his record is as follows:
April 20, 1903; at Boston, second game
Philadelphia 0 0 0 0 0 0 6 1 3 – 10
Boston 0 2 1 0 3 0 0 0 1 – 7 Young, 7 inn. (6 runs), Hughes, 2 inn., LOST
The two encyclopedias and The Hall of Fame Fact Book all agree on Cy Young’srecords for the years 1904-11, as do I with one crucial exception. For 1907, all three books show Young with a 22-15 record, but my research shows that he was actually 21-15.
In reviewing the Red Sox pitching staff day by day for 1907 from the official sheets, I discovered that an extra win had been marked on Cy Young’s personal sheet without a corresponding date. The extra win is sandwiched between the dates of May 24, when Young pitched a complete-game victory over St. Louis, and May 29, when Young made his next start and lost. He did not pitch in any games between those dates.
See the photo of this portion of Young’s official sheet, and note the peculiar placement of the extra “W.” It could be that the official scorer started to give Young a win on May 29, then realized his mistake and added the “L.” The writing was in ink and may have been difficult to erase. Or the scorer could have been thinking of the game of May 30, which Young finished for the victorious Ralph Glaze. In any event, the bottom of Young’s sheet showed a won-lost record of either 20-15 or 21-15, which was erased and changed to 22-15 (the handwriting was of the period). The scorer must have counted the “W”s without matching the dates. It could well be that he originally had Young at 21-15, but in rechecking counted the extra “W” and changed the total to 22-15.
In 1907 Young appeared in 43 games, completing 33 of 37 starts and relieving six times. In complete games, Young had a record of 18-13; in incomplete starts, 1-2; and in relief, 2-0 with a pair of saves. He pitched two more complete games which ended in ties, and one start in which he was not involved in the decision.
I checked out Young’s two saves and his incomplete start that the Red Sox won. They are as follows:
May 30, 1907; at Philadelphia–second game
Boston 0 0 0 0 0 1 0 0 5 – 6 Dinneen, 2/3 inn., Glaze, 7-1/3 inn., WON; Young, 1 inn.
Philadelphia 2 0 0 0 0 0 0 2 0 – 4
Glaze was batted for in the ninth, but was not removed for poor pitching (see Practice Two and the example of Marquard-Crandall). The official sheets gave this win to Glaze.
August 9, 1907; at Boston
Chicago 0 0 0 0 0 3 1 2 0 – 6
Boston 1 0 2 1 0 3 0 0 x – 7 Glaze, 6 inn., WON; Young, 3 inn.
Glaze was batted for in the sixth inning, but went out with the Red Sox ahead by at least 4-3. The official sheets give the win to Glaze. Macmillan shows Young with a relief record of 1-0 with three saves, but the correct figures are 2-0 with two saves, the wins coming on August 21 and August 28.
June 14, 1907; at Boston
St. Louis 3 0 0 0 0 0 0 0 0 – 3
Boston 0 3 0 0 1 0 0 0 x – 4 Young, 1 inn.; Winter, 8 inn., WON
Official sheets show Winter the victor. After proving to myself that Young was without a doubt 21-15 in 1907, I was led to wonder how the two encyclopedias ever reconciled the individual records of the Red Sox pitching staff to the team record that year of 59-90: if Young had one win too many, somebody had to have one win too few. My first hunch was that Glaze had been deprived of his win on May 30, described above, but no–both books are in accord with the official sheets in listing him at 9-13. Neft-Cohen and Macmillan each gave Winter one win too few . . . but this they balanced by giving Cy Morgan one too many!
The win which disappeared in order to balance Young’s belonged to Rube Kroh, who pitched a complete-game 2-1 victory over the Browns on August 18, 1907, yet is listed in Neft-Cohen and in Macmillan as 0-4 for the year. Whatever Macmillan and Neft-Cohen are basing their won-lost decisions on, it certainly is not the official record.
Another major difference between the official sheets and the reference books has to do with the lifetime record of Walter Johnson. The Hall of Fame Fact Book; Macmillan; Neft-Cohen-Deutsch; Turkin-Thompson–you name it, they all show The Big Train with a career record of 416-279. Yet my research proved him to have one more win and one fewer defeat.
I discovered the first error on his record while examining the 1912 season. This was the year in which Johnson became the first A. L. pitcher to win 16 straight games. Although this outstanding record was equaled later that same year by Joe Wood, and again by Lefty Grove in 1931 and Schoolboy Rowe in 1934, it remains unsurpassed in the American League 70 years later.
This great streak lasted from July 3 to August 23. It was well documented in the Spalding and Reach Guides and in all the newspapers of 1912. Ban Johnson fully accepted this winning streak and his references to it were amply quoted in the guides and the daily press. Included in the streak was a game played against Chicago on August 5.
At Chicago; 10 Innings
Wash. 0 0 1 0 0 2 1 3 0 1 – 8 Groom, 2-1/3 inn., Cashion, 5-1/3 inn., Johnson, 2·1/3 inn.
Chicago 1 0 6 0 0 0 0 0 0 0 – 7
Walter Johnson drove in the winning run in the tenth inning. The weekly listings in the Washington Post and the New York Times gave this win to him. The Reach and Spalding Guides gave this win to him. There did not exist a scoring practice in that period which would have given the win to anyone but him.
It was an open-and-shut case–except that the A. L. official sheets showed Jay Cashion as the winner, and thus left Johnson with only a 15-game winning streak!
This problem is unresolved today. There is no doubt in my mind that the clerk making out the official sheets made a mistake. Walter Johnson allowed no hits and no runs in 2-1/3 innings, plus drove in the winning run. This was a performance far superior to that of Cashion, who allowed three hits and two runs in 5-1/3 innings.
As mentioned earlier, during this period an A. L. official scorer recommended a pitcher for a win or a loss, and then Ban Johnson either agreed or changed the decision. The fact that Walter Johnson appears as the winner in the weekly newspaper listing proves that the official scorer recommended him for the victory. The fact that Ban Johnson accepted the 16-game winning streak proves he agreed with the official scorer. Walter Johnson’s record in 1912 should be changed from 32-12 to 33-12.
In 1917, AI Munro Elias published Walter Johnson’s pitching record from 1907-17 in a Washington newspaper. He showed Johnson’s record against every team for the 1912 season and added it all up to a 33-12 record. I have researched Johnson’s 1912 season day by day and my breakdown agrees completely with that of Elias.
The additional win for Johnson will give him a 9-1 mark against Chicago, which will tie the major-league record for most wins over one club in a season. This will also mean that he had the most wins on the road of any A. L. pitcher in 1912: his record was 17-4.
If Johnson is not given the victory, it will mean that his winning streak in 1912 was only 15 games, and he would no longer be tied for the A. L. high. I believe Johnson deserves the win because all the evidence is on his side. In the Cobb-Lajoie affair, the Baseball Records Committee ultimately left the batting title in the hands of Cobb, despite the obvious existence of a duplicated entry on his sheet, primarily on the basis that Ban Johnson had investigated the matter and had so ruled. Consistency as well as simple justice would dictate following Ban Johnson in the matter of Walter Johnson’s “hidden” win in 1912.
The other two errors in Johnson’s record occurred in the season of 1916. The Hall of Fame Fact Book, Macmillan, and Neft-Cohen all show him with a 25-20 mark, but the official sheets have Johnson at 24-19.
Anyone looking at the 1916 newspaper box scores of the games in which Johnson pitched and applying modern scoring practices would give him 20 losses. Under the practices of that period, however, only 19 losses were awarded to him. The game in question:
August 7, 1916; at St. Louis; 10 Innings
Wash. 0 0 0 0 0 0 2 0 0 – 2 Gallia, 7 inn., LOST; Ayers, 1-1/3 inn., Johnson, 1-1/3 inn.
St. Louis 1 0 1 0 0 0 0 0 1 – 3
Common Practice Number Nine applies perfectly to this game: the theory of charging the starting pitcher with the defeat if he was the one who allowed the most runs or could be held mainly responsible for the defeat. The official scorer must have felt this was the case and so charged Bert Gallia with the defeat, even though he did not pitch poorly. There are too many examples of this type of game between 1905 and 1916 to dismiss it as an error or a fluke.
The other problem on Johnson’s record in 1916 was whether he won 25 games, as shown in all the reference books, or only 24, as shown on the official sheets. The game in question was played at New York on June 26, 1916, and lasted 11 innings.
Wash. 0 3 1 3 0 0 1 0 0 0 1 – 9 Gallia, 3-1/3 inn., Harper, 3-2/3 inn., Johnson, 4 inn.
New York 0 1 1 3 0 0 2 0 1 0 0 – 8
The weekly newspaper listings in the New York Times show Johnson as the winner in this game. This would mean that the official scorer recommended him for the victory; but the official sheets show Harry Harper as the winner. Johnson allowed one run and three hits in four innings while striking out five. He was far better than Harper, who allowed five hits and three runs in three and two-thirds innings. Also, Johnson finished the game very strongly and was pitching when the lead run scored.
I cannot see Ban Johnson overruling the official scorer on a decision like this. Unlike the game of August 7, 1916 cited above, for which we have an abundance of examples to show that the official scorer was right in not giving the defeat to Johnson, here we have a situation in which the common practices of the period all point to a Johnson victory. He should have a 25-19 record in 1916, and accordingly a lifetime mark of 417-278.
I am not the first person to find errors on Johnson’s won-lost record. Up to the late 1950s, Johnson was universally shown with a lifetime log of 414-281, but then a researcher found that The Big Train had been charged with two defeats in 1911 which were really complete-game victories. This elevated his record that season from 23-15 to 25-13, and his lifetime totals from 414-281 to 416-279.
As we have seen from the 1910 Cobb-Lajoie situation, errors were made on the batting records, too–and something similar to Cobb’s “phantom” 2-for-4 game can be found in the 1913 record of Boston outfielder Duffy Lewis. His official sheet lists him as playing in a game on June 29 and collecting two hits (a single and a triple) in four at bats while scoring a run. The only problem was that the Red Sox did not play that day, so his 1913 totals have to be adjusted. Oddly, the rest of that famous outfield–Tris Speaker and Harry Hooper–also had their records botched that year, as Speaker was given an extra hit and Hooper an extra at bat. Small potatoes, perhaps, but it gives you an idea of how rampant scoring errors are in the years before 1920.
Even something impossible to overlook, like a record-setting winning streak, can be overlooked. Reviewing the season of 1891 in the newspaper accounts, I came across an 18-game winning streak by the Boston Nationals. All record books show Boston’s longest winning streak as 17 in 1897, but this is wrong. The 1891 Braves, who were then known as the Beaneaters, made one of the greatest comebacks in the history of baseball: on September 15, they were six and a half games behind Chicago, but then they started on the 18-game streak which resulted in their winning the pennant by three and a half games. The streak lasted from September 16 through October 2.
I hope that the Baseball Records Committee and the publishers of the various encyclopedias and record books can straighten out the statistics for 1876-1919. Since there are no official sheets before 1903 (N.L.) and 1905 (A.L.), I would think that the I.C.I.-recompiled batting records for the period 1876-1904 should be accepted as the most nearly correct. My personal goal is to finish my day-by-day record for all American League pitchers from 1901-19 and to compile a complete won-lost list for each pitcher based on the official sheets, except where errors are found. Included will be a won-lost record in relief along with the total saves.
In conclusion, I would say that with the identification of the common scoring practices of the 1901-19 era as they relate to pitchers’ won-lost decisions, the time is ripe to support the goals of this study and, at last, to clear up the baseball records mess.