November 21, 2007
When Four Aces Don't Matter
The 1940 White Sox
Since 1901, the major leagues have sported 2166 single team seasons, 2182 if you count the Federal League. In that time, the top 15 or so pitchers each season are generally thought of as aces, with adjusted ERA (ERA+, or ERA normalized for parks and leagues, with 100 being average) serving as the easy guide to an ace. After adding a few thresholds--22 GS/130 IP (to weed out some super-relievers from the '70s and a few rookie sensations called up in July), no fewer than nine wins, and a WHIP under 1.4 to help weed out flukes and pitchers who don't go deep enough into ballgames to get regular decisions--I got down to 1478 of these ace-level seasons, which seemed about right. It's not a perfect way of identifying aces, but since an ace is as much about perception as it is the performance, ERA+ with some qualifiers serves as a decent enough proxy. There are worse methods.
By this standard, some teams have multiple aces in a year. Forty-nine clubs had three aces (or ace-level performers, if you prefer), and five others had at least four. Here are those five, with their ERA+ listed next to their names:
Four of these pitching staffs were part of dominant teams which you might immediately be familiar with. The 1940 White Sox, on the other hand, didn't even reach 10 games over .500 until September 12, and finished the year 82-72, tied for fourth place and eight games out. It would be easy to say that Chicago's mediocre offense or Detroit's having three of the top five pitchers in the AL that year would explain it all, but it still doesn't change that the Cubs, Giants, and Braves were great teams, and the White Sox decidedly weren't. What accounts for having four aces and no chance at the pennant?
Oddly enough, the 1940 White Sox weren't even the best pitching staff in their league. The pennant-winning Tigers--led by Bobo Newsom, Tommy Bridges, and Schoolboy Rowe, all three among top five pitchers in the AL--achieved a staff-wide ERA+ of 120, an outstanding figure, and one that's better than the White Sox's mark of 118. The Tigers' 120 is in the top 50 all-time in the AL, while 118 doesn't quite make that list; it's still very good. Although the Tigers' raw numbers don't look like much (4.01 ERA), their ballpark pitching factor was 109, or somewhere in between Citizens Bank Park and Coors in 2005. In a league scoring 4.97 runs per game (this year's AL scored 4.90 per game), 4.01 in a launching pad is remarkable.
On the raw stats side, the Indians allowed the fewest runs per game that year with 4.11, due to an astounding 3.43 R/G at home (with a ballpark factor of 96). It's fair to say that 320 innings of Bob Feller (27-11, 2.62) helped them a bit.
In contrast, Comiskey Park was a neutral environment that year, with a ballpark factor of 101. This does not explain the following, though:
R/G, Home R/G, Away Offense 4.58, 6th in AL 4.93, 5th in AL Pitching 4.65, 3rd in AL 4.01, 1st in AL (by .31)
Despite these disparities, the White Sox were 41-36 both at home and on the road. Their 735 runs scored and the 672 runs allowed make for an 83-71 Pythagorean record, which is basically what they got with their 82-72 finish.
It wasn't the other pitchers that dragged this team down. Of the two other starters, Bill Dietrich had an ERA+ of 110, and Jack Knott came in at 97; "closer" Clint Brown had a 120, and mop-up reliever Pete Appleton was at 79. Getting a 29-21 record out of your B-squad certainly seems like it should be good enough for a team with a quartet who could all be found in the top 11 slots in the AL for ERA+ (minimum 100 IP) to contend strongly.
Ironically, Sox skipper Jimmie Dykes was a push-button manager when it came to his pitchers; the staff's 83 complete games led the league, and in the 72 other games (counting one tie), only 91 total relief appearances were made. It wasn't pitcher fatigue that caused to Sox to give up extra runs, though, as the ninth inning was the lowest-scoring inning for opponents both home and away.
The four aces weren't just pounding the doormats, either, as their ERAs were mainly from beating good teams. In line with a common practice at the time, Jimmie Dykes leveraged his starters, using his southpaws Smith and Lee for 15 of the 22 games against Cleveland and New York, and generally keeping them away from Boston. Smith also started nine times against the Senators and Athletics, with the Sox winning all of those games. Although Knott was 4-1 against the Tigers, he and Dietrich tended to pitch against the weaker teams, while Lyons was the Sunday starter and Rigney pitched against everyone. This was fairly sensible, and it generally worked out, as the best pitchers came out with the best ERAs while facing the best teams. Lee, in particular, was exceptional, giving up four runs or fewer in 21 of his 27 starts, most of them against the Yankees and Indians. For his part, Rigney delivered four runs or fewer in all eight of his starts against the Tigers.
The team was certainly undaunted by the Tigers' aces, as they were 13-9 head-to-head against Detroit, going 4-2 when Bridges started, and 2-3 when Newsom started--keeping in mind that Bobo went 21-5 that year--and 13-9 overall. Add all of these ingredients together, and what you think you've got cooking is a pennant winner. So what's missing?
The Apparent Explanations
The White Sox Paradox--that a team with this good of a front four and a credible back four not only weren't in contention, but matched their Pythagorean record in doing so--apparently boils down to two things:
Dutch Leonard, the veteran Senators knuckleballer, had a fine season, but not the wins to show for it: a 3.49 ERA, but 14 wins against a league-leading 19 losses. This was not out of line for Leonard or the Senators of the '40s; he was a good pitcher on bad teams pretty much his entire career. Like Tim Wakefield versus the Devil Rays, however, Leonard turned into Superman against the White Sox, going 5-1 against the; the White Sox were 11-5 against the Senators when Leonard wasn't pitching, so it wasn't the lineup that was giving them fits. What follows is the record for all pitchers who started against a contender six or more times, giving the starter's team's W-L rather than the pitcher's personal record. I also included other noteworthy W-L records for a more complete picture:
Name vs.DET vs.CLE vs.NY-A vs.BOS-A vs.CHI-A Tommy Bridges (DET) x 5-0 3-0 2-4 Bob Feller (CLE) 3-4 x 4-4 4-1 4-1 Al Milnar (CLE) x 2-4 4-2 Al Smith (CLE) x 4-3 5-0 Mel Harder (CLE) x 5-1 Red Ruffing (NY-A) 3-3 x 3-0 Marius Russo (NY-A) x 5-1 Johnny Rigney (CHI-A) 4-4 2-4 x Thornton Lee (CHI-A) 3-4 4-4 x Eddie Smith (CHI-A) 2-6 3-4 x Elden Auker (STL-A) 5-2 2-4 Vern Kennedy (STL-A) 0-3 2-4 Bob Harris (STL-A) 0-4 2-4 Johnny Niggeling (STL-A) 2-4 1-4 Dutch Leonard (WAS) 0-3 6-1 2-4 5-1 Ken Chase (WAS) 3-4 2-4 4-2 Sid Hudson (WAS) 1-4 3-3 Johnny Babich (PHI-A) 5-1 Nels Potter (PHI-A) 3-3 Potter Vaughan (PHI-A) 0-6
Not too many pitchers "owned" a contender: Auker did with the Indians, three different pitchers could make that claim against the Yankees, three against the Red Sox, etc. Five guys owned the White Sox, including three Indians. Perhaps understandably, the White Sox went only 6-16 against the Indians that year, with the lowlight being Feller's Opening Day no-hitter. None of the other contenders owed as much to so few as the White Sox with their 6-22 record against those five pitchers and a 76-50 record against everybody else.
These were all good pitchers, so it wasn't ignominious to lose to them, but to be owned by most of one team's rotation would break just about anyone. Cleveland exceeded their Pythagorean record by four games, as they were somewhat of the same skill level on both offense and pitching (96 OPS+, 116 ERA+) as the White Sox (92 OPS+, 118 ERA+). How they got to second place was basically a product of their having the whammy on the White Sox. Had even a few of those 5-4 games gone the other way, we'd remember less about the controversies involving Ossie Vitt's relationship with his players and more about Jimmie Dykes's great group of pitchers.
Which brings me to the other point: the White Sox played an astonishing amount of one-run ballgames, going 31-32 in them. Historically, 38 teams since 1901 have played 63 or more one-run ballgames, but only three of them are found prior to 1940; 14 of those teams were in the '60s, and no team has played that many since 1993. While that statistic is no help by itself--after all, the White Sox fared reasonably in those contests--it does point to something else strange at work, borne out in the following table of how many times each contender scored or allowed X number of runs in 1940:
Team 0 1 2 3 4 5 6 7 8 9 10+ Offense DET 4 8 16 17 16 23 16 12 8 10 24 CLE 4 23 19 23 18 17 11 13 8 4 14 NY 7 14 17 21 18 16 12 11 8 8 22 BOS 9 15 11 16 20 14 12 15 11 7 24 CHI 7 19 20 14 23 15 16 10 7 8 15 Pitching DET 10 17 28 15 18 10 11 14 12 7 12 CLE 13 22 23 18 19 20 11 7 3 6 12 NY 10 19 13 31 21 12 17 6 8 5 12 BOS 4 16 13 14 24 16 15 15 9 12 16 CHI 9 17 15 34 20 17 7 12 5 2 16
The offenses fall into two natural groups: the Tigers, Yankees, and Red Sox, all with impressive blowout capabilities, and the noticeably weaker Indians and White Sox. What's more interesting is the pitching distribution. The White Sox were fairly standard in this group in shutouts or one-run ballgames, achieving no real difference from the Tigers or Yankees, with the latter yielding only one run less than the White Sox did over the season, 671 to 672. Yet the Tigers--who gave up 717 runs that year and played in an extreme hitting environment--held their opponents to two runs in 28 games, which would have won ballgames even with the mediocre White Sox attack. The Indians were no slouches: in a whopping 58 games, opponents scored two runs or fewer against Feller and friends.
In contrast, the White Sox pitching staff was not quite so dominant, yielding three or four runs over one-third of the time. While that's still a good mark, Chicago was also prone to giving up double digits, which they couldn't get away with. The Indians allowed seven or more runs 28 times, the Yankees 31 times, and the White Sox 35 times. The Indians had a weak offense, but their pitching staff made sure they rarely needed much,while the Yankees were decidedly capable of big days on offense. In contrast, the White Sox had a small offense but needed big days that they simply could not provide. With an offense like Chicago's, the difference in a pitching staff consistently allowing 2-3 runs a game versus 3-4 runs is large, particularly with as many one-run ballgames as they played. If the pitching had given more two-run ballgames to the offense but had more blowouts--keeping the same amount of runs allowed but distributing them less evenly--the White Sox would likely have won more often. Detroit allowed 45 more runs, but allowed them in a way far more conducive to winning with a small offense, which they didn't really need to do, but it helped anyway. (In case you're wondering if those low-run ballgames were an "away from Coors" effect, 39 of the 55 games where they allowed two runs or fewer were at home.)
The White Sox, in essence, had New York's pitching but Cleveland's hitting, which was the worst of both worlds; when combined with Cleveland's head-to-head domination of the White Sox, 82-72 sounds about right.
Such was the lot of the 1940 White Sox, an odd team gifted with one of the five deepest starting staffs of all time, and yet not dominant enough for the middle-of-the-pack offense to win with. As strange as it may be to look at this five-team list and wonder how this one club didn't even contend, the other numbers back up their final record. As great as the four aces were, the one guy they couldn't beat was Pythagoras, who with the White Sox's was probably smiling in 1940--and playing for the Indians.
Brandon Isleib is a contributor to Baseball Prospectus. He can be reached by clicking here.