It was on this day in 1989 that Pete Rose was banned for life from Baseball, proving that this Rose by any other name might still gamble on anything that had a score.
My mind was drawn to baseball by the incredible game on Wednesday in which the Rangers beat Baltimore by the score of 30 to 3. Pretty good, but not quite the record.
The Chicago Colts of 1897 were a somewhat mediocre team, winning only 59 games out of the 138 games scheduled (ok, there were two ties, it happened back then); but on a steamy June 29, they put “ducks on the pond” as Dizzy Dean would say, and then brought them home for dinner. In that game they smashed the Louisville Colonels by the football-like score of 36-7 to pull just 14 games under .500.
In the game this year, the surprising thing to many people is that in their nine innings of batting, the Rangers put up five goose eggs. All the runs were scored in only four of the innings, five in the fourth, nine in the sixth, ten in the eighth, and then, with the game surely in hand in the top of the ninth, they let up and only scored six runs.
Baseball fans, pundits and reporters and such leap to explain that, in baseball, runs come in bunches. Here was a typical comment from a reader of a blog that commented on the “unusual” distribution of runs:
· “This is less surprising than you might think. A number of studies have shown that not only do runs most-often score in clumps but that the winning team most-often scores more runs in one inning than its opponent scores in the whole game!”
Ok, I’m a baseball fan, and I’ve heard both those ideas pitched around, but on this day I found a moment to check them out, and as the title of the blog states, “It ain’t necessarily so!”
I ran through the box score of every game that occurred that day, 16 games including the memorial stinker that we mention at the outset (which was actually only one of two games played by the same two teams that day). Here is what I found….
Subject one, Do Runs most often occur with more than one run per inning?
In the 16 games there were 284 half-innings (one game went ten innings and six teams didn’t have to bat in the bottom of the ninth). The distribution of runs scored? Well, in two hundred of the half-innings there was no score, that’s just over 70% and makes the five run-less innings by the Rangers seem more typical. But what about the “runs most often score in clumps”? NOPE, not so. Each run total became less likely as the score went higher. Of all the innings in which a team scored (84), there were 33 (about 39% of the run scoring innings) in which they only scored one run. Two runs were only scored 28 times (33%) . There were 13 half-innings with three runs scored, and 4 with four runs scored. That accounts for 78 of the 84 run producing team at-bats. The remaining six came from 2 five run innings, and one each of six, seven, nine and ten. The distribution looks typically like an exponential decay (picture). NO CLUMPING PATTERN. And if I add in the results from Thursday's games, the 224 half-innings of that day brings us to a total of 508 team at-bats. Here is the break out
Runs scored ___Frequency ___Percentage
___0____________ 359________ 70.6
___1_____________ 70 ________13.7
___2_____________ 42_________ 8.3
___3_____________ 19 _________3.7
___4______________ 7_________ 1.4
___5 or more ______11 _________2.2
and even of just the 149 half-innings in which a run was scored, the percentage of one run innings is up to 47%, and over 75% of the run scoring half-innings are two or less.
The next time the guy beside you in the bleachers looks over and explains the big inning as a result of the “runs in clumps” theory, buy him a bag of peanuts, and explain… “Peanuts come in bunches (ask someone from Georgia), but not runs, No Clumps.”
So what about the idea that the winning team scores more runs in a single inning than the losing team scores in the whole game? In the Rangers/Orioles double hitter it was true in the slaughter, but not in the 9-7 nightcap. It was true in the Yankees victory, but not in the Royals 7-6 victory over the Chi-Sox (it is however, almost always a statistical anomaly when the Royals manage a win) and not in the Tigers 11-8 loss to the Indians. The Twins had 7 in the first inning of their 8-4 win, so they did it, and the Devil Rays score both runs of a 2-1 win in the same inning so they also did it and the Athletics won 4-1 with a four-run seventh inning. So far that makes five supporting and three opposing.
But Arizona and Cincinnati both managed only to tie their opposition in the big inning, so they bring the record to 5-5. Pittsburg posted one for the hypothesis with a big victory, but Houston and San Diego both fell short; bang, we are at 6-7. In the three remaining games, The Dodgers killed the Phils and brought the thesis back to 7-7, but in the last two games of the day, both the Cubs and the Cardinals fell short. Total for the Day, seven for the theory, nine against. Ok, that’s pretty good but not “most often.”
I just wondered, so I checked the games from Thursday night. For that night, eight of the twelve winning teams DID score more in a single inning than the opposition. So for two days back to back the results show fifteen to thirteen in favor of the “one inning beats the loser” idea. Then I looked at Tuesday's games, 7 supporting, 8 opposing. That brings the three day total to 22 to 21 in favor. It looks about like a coin flip to me, certainly way short of the fai-safe bar bet I always heard it to be. I suppose I should look at a few hundred more games… maybe tomorrow.
· I I didn’t hear an explanation for this fact, which seems unusual to me. It was posted in response to a Freakonomics Blog about the game: “The other interesting baseball related fact about the game is that about half of those 30 runs came from the 7,8 and 9 hitters; typically the least productive batters in the lineup. “
— Posted by josh