In a paper in American Statistician, Michael Goodman makes the claim that double-headers are swept more often than they are split.[1] The Elias Sports Bureau, however, found that double headers are swept only 26.2% of the time
-
I understand the time frame plays into each one's research, but for it to be that big a difference? -
Most likely the questions have been phrased differently in each case. IOW, I think it's just a semantics issue.
If each game of the double header was a coin flip, you'd expect a sweep 50% of the time.
You'd also have to say the chance of a given team sweeping was only 25%.Comment -
look at the citation -- the quoted article looked at just the 1966 season.
seeing is is examining what amounts to probably 1-2% of all double headers i think it fair to pull the small sample size card.Comment -
Got to know if any assumptions are being made.Comment -
obviously most are split, not sure what the guys malfunction is but thats common sense and knowledgeComment -
Not true.
Year - Sweeps - Splits
* 2008 - 14-11
* 2007 - 12-10
* 2006 - 10-13
* 2005 - 7-11
* 2004 - 23-15
* 2003 - 20-8
* 2002 - 14-10
Last edited by Bluehorseshoe; 04-09-11, 10:01 PM.Comment -
In order for "most are split" to be true, most times the winner of the first game would have to lose the second game.
Why would you think that the winner of the first game--that is, a team for which you have some slight evidence of being better--would be less likely to win the second game?
In your "common sense," stop emphasizing the first word, work more on the second.Comment -
bust out with all of mlb history, not just one decade or so
I remember watching a game where an announcer claimed historically most doubleheaders are split
Unless he was just blatantly lying or misinformed, I am inclined to believe him before some random internet posters on a gambling forumComment -
I would be willing to bet that the 26% figure was in answer to a specific side sweeping. Probably on average a 26% chance of either a visit or home team sweep, making a sweep by one or the other a slight favourite.
I`ve also done some research in this area and it agrees with the sweep as the most likely outcome as well. If you think about it`s only common sense really. Team A and team B are playing two games. Team A wins the first. You know nothing else about these teams - although the sample size is a bit small
, the most likely case is that team A is the better team, and therefore a favourite to win game 2.
PLPComment -
You're inclined to believe ex-jocks who probably still think batting averages are good hitting metrics and have never done a statistical study in their lives, over "random" internet posters who aren't asking you to "believe" anything, but to instead think for yourself for once?
Think, warriorfan.
If the games were 50/50 propositions, what would the odds be of a sweep?
In what ways and for what reasons are the games not 50/50 propositions?
How would the fact that they aren't 50/50 propositions affect the odds of a sweep?
You don't need a database to get the answer, you just need to think it through.Comment -
A real doubleheader is an Ernie Banks "let's play two" where one game shortly follows the other, not the day-night doubleheaders you mostly see these days for weather related make up games.sbrComment -
Time frames would have to come into play in a major way in any sort of data on double headers, the unions, relocation and basically money changed everything on double headers.
They used to be much more common and the lineups would be completely up in the air, you literally would have to be out at the ball park to see who was playing and warming up to find out who was pitching in the second double header. .
Sundays would be huge for double headers but this started to fade out in the 50's. Here are some percentage I had from the hardball times, www.hardballtimes.com . Its amazing looking at these stats, double headers used to make up 33% of games in the NL 1956 and 26% in the AL in 1956, they make up less than 2.5% of games in either league now.
the percentage of double headers
1956 33.17% 26.54%
1957 26.49% 22.73%
1958 23.05% 26.17%
1959 19.35% 22.33%
1960 18.42% 24.96%
The percentage of double headers and the attendance of games. The attendance seems to have played a big part of the change, probably union pressure as well.
Year NL AL NL Att AL Att
1961 19.06% 29.84% 14,106 12,531
1962 22.17% 27.44% 13,990 12,380
1963 18.50% 26.73% 14,035 11,256
1964 19.95% 28.75% 14,834 11,345
1965 18.94% 28.13% 16,705 10,939
1966 17.80% 26.05% 18,561 12,614
1967 18.52% 29.14% 16,014 13,996
1968 17.47% 21.43% 14,496 13,938
1969 17.27% 18.91% 15,514 12,471
1970 15.04% 17.06% 17,160 12,420
1971 14.20% 18.43% 17,824 12,286
1972 15.48% 19.59% 16,699 12,313
1973 15.24% 15.84% 17,173 13,821
1974 14.81% 13.99% 17,467 13,409
1975 12.98% 17.01% 17,096 13,696
1976 14.81% 14.27% 17,140 15,158
1977 11.11% 15.56% 19,620 17,365
1978 12.77% 14.32% 20,800 18,152
1979 12.77% 9.57% 21,811 19,833
1980 9.46% 9.36% 21,710 19,338
1981 6.99% 6.67% 19,376 18,755
1982 7.00% 7.93% 22,127 20,335
1983 9.24% 8.63% 22,125 21,137
1984 7.62% 6.88% 21,402 21,130
1985 5.25% 5.30% 22,958 21,672
1986 4.54% 4.41% 23,048 22,198
1987 3.33% 3.00% 25,473 24,054
1988 4.13% 3.71% 25,283 25,199
1989 4.32% 4.59% 26,027 26,345
1990 5.35% 3.71% 25,197 26,772
1991 2.68% 2.82% 25,460 28,322
1992 3.91% 3.00% 24,806 28,006
1993 2.29% 2.47% 32,533 29,395
1994 2.49% 2.51% 32,139 30,367
1995 1.79% 2.18% 24,936 25,108
1996 3.53% 2.82% 26,789 26,230
1997 3.17% 4.77% 28,118 27,635
1998 3.24% 2.65% 29,605 28,372
1999 2.70% 2.38% 29,582 28,094
2000 2.08% 2.38% 30,608 27,970
2001 1.93% 1.85% 30,654 28,997
2002 1.85% 2.12% 28,614 27,313
2003 2.32% 2.11% 28,358 27,231
2004 3.17% 3.09% 31,062 28,948
2005 1.54% 1.41% 32,108 29,339
2006 2.32% 1.41% 32,107 30,393
2007 1.31% 2.38% 33,998 31,207
2008 1.70% 2.47% 34,068 30,459
2009 1.54% 2.29% 31,788 28,433Last edited by gregm; 04-18-11, 10:53 PM.Comment
