Double Header Sweeps

[ScottyDo]

So I watched yesterday's double-header, which we split, and it reminded me of a stat I heard a few years back: double-header sweeps happen something like 25% of the time.

 

That's odd to me, because doing some real basic math, if the teams are evenly matched, you should see a sweep (one side or the other) 50% of the time, and a split the other 50%. What's more, if you assume that, on average, teams are NOT mismatched (i.e. one team has a 60% chance of winning the game and the other has a 40% chance), then the odds of either team A sweeping or team B sweeping goes over 50%.

 

I found the stat again, and Elias says that 26.2% of double headers result in sweeps. Given the immense sample size of double headers, it's clear something is causing this disparity.

 

I only bring this up because there were a few people outraged that we didn't sweep the Tigers in yesterday's DH, which I think is ridiculous because a) the odds of winning those two games WITHOUT the above stat aren't astronomically high and B ) something is clearly happening to push the odds against what people were clammoring for.

 

Plus, Carlos Torres was pitching.

 

P.S. if you have any insight into why the disparity between expectation and reality is so huge, please illuminate us.

Edited August 4, 2010 by ScottyDo

[Balta1701]

Stats like that suggest to me that there is a serious emotional component involved...one team is pissed that they lost the first game so they refocus for the 2nd game to avoid the sweep, while the other team becomes complacent for the 2nd game.

[ScottyDo]

Stats like that suggest to me that there is a serious emotional component involved...one team is pissed that they lost the first game so they refocus for the 2nd game to avoid the sweep, while the other team becomes complacent for the 2nd game.

I mean, it must be something like that. I was trying to take lineups and coaching moves into consideration but I couldn't really think of much that would swing the odds that far. It's gotta be something psychological, which is funny cuz as stat-heads, we reaaaaaally hate confounds like that.

[Balta1701]

I mean, it must be something like that. I was trying to take lineups and coaching moves into consideration but I couldn't really think of much that would swing the odds that far. It's gotta be something psychological, which is funny cuz as stat-heads, we reaaaaaally hate confounds like that.

It could also be that teams tend to win the game where they throw their better pitcher out against the other team's worse pitcher.

[witesoxfan]

It could also be that teams tend to win the game where they throw their better pitcher out against the other team's worse pitcher.

 

I was going to post something similar to this with some statistics, but the numbers didn't make sense as there was far too much variation. But in yesterday's games, Buehrle went against Porcello, who has been destroyed by the Sox, and Torres went against Bonderman, who is solid against the Sox when he can prevent them from hitting multiple home runs.

[LittleHurt05]

It could also be that teams tend to win the game where they throw their better pitcher out against the other team's worse pitcher.

 

I think this could be the biggest reason. Two baseball teams are rarely actually evenly matched during any given game. How good the starting pitcher is has a huge effect on the outcome. Maybe the pitching matchups are usually split where team has a better SP in game 1, and the other team does in game 2?

[ScottyDo]

It could also be that teams tend to win the game where they throw their better pitcher out against the other team's worse pitcher.

The only thing I don't get about that, though, is why the odds of a DH are so much different than the odds of winning 2 consecutive games on different days. Same pitching matchups (theoretically), different results. Do teams pull up AAA pitchers for DH's way more often than I thought?

Edited August 4, 2010 by ScottyDo

[Balta1701]

The only thing I don't get about that, though, is why the odds of a DH are so much different than the odds of winning 2 consecutive games on different days. Same pitching matchups (theoretically), different results. Do teams pull up AAA pitchers for DH's way more often than I thought?

If they don't, another thing that can be happening is people going on short rest.

 

Teams could also be burning out their bullpens in an effort to make sure they win one game rather than being swept. You have a 3-2 lead in the 5th inning of game 1, your starter walks the bases loaded, suddenly you've won game 1 but you've blown through 5 pitchers.

[LittleHurt05]

This season there have been 9 doubleheaders, 3 of them were sweeps (33%).

 

2010 DOUBLEHEADERS

DATE TEAMS GAME 1 GAME 2

April 24 Florida at Colorado Fla, 4-1 Col, 8-1

April 27 LA Dodgers at NY Mets NYM, 4-0 NYM, 10-5

May 8 Baltimore at Minnesota Bal, 7-3 Min, 6-1

May 12 NY Yankees at Detroit Det, 2-0 NYY, 8-0

May 12 Philadelphia at Colorado Postponed Col, 4-3

May 15 Washington at Colorado Col, 6-2 Col, 4-3

June 10 San Diego at NY Mets SD, 4-2 NYM, 3-0

July 17 Detroit at Cleveland Cle, 4-3 Cle, 2-1

July 25 Toronto at Detroit Tor, 5-3 Det, 6-5

August 3 Chicago Sox at Detroit CHW, 12-2 Det, 7-1

September 6 Florida at Philadelphia 1:05 PM 7:05 PM

Edited August 4, 2010 by LittleHurt05

[Greg Hibbard]

Almost always in game 2 there are significant lineup differences, with at least 1-2 backups (particularly at catcher, where there is usually a big dropoff in talent) playing instead of the regulars.

 

I would add that occasionally it seems the winning manager in game 1 will take more chances with the lineup in game 2.

[iamshack]

Well, keep in mind that there are 3 possibly outcomes, not 2 here. So the odds are more like 33.3 % in a sweep.

 

Then you add in the conservative nature in the way the bullpens are usually managed - most managers will manage more safely in the first game to avoid running through their entire bullpen, and I think it leads to more splits.

 

Most managers seem to almost manage for a split.

[ScottyDo]

Well, keep in mind that there are 3 possibly outcomes, not 2 here. So the odds are more like 33.3 % in a sweep.

 

Then you add in the conservative nature in the way the bullpens are usually managed - most managers will manage more safely in the first game to avoid running through their entire bullpen, and I think it leads to more splits.

 

Most managers seem to almost manage for a split.

That's not entirely accurate. It kind of works like a Punnett square, if you can think back to high school biology. You've got team A and team B. You've got 4 possible outcomes: AA, AB, BA, BB. If the teams are evenly matched, then you've got 2 sweeps and 2 splits.

[ScottyDo]

The bullpen thing makes the most sense to me, thinking about it.

[BFirebird]

Well, keep in mind that there are 3 possibly outcomes, not 2 here. So the odds are more like 33.3 % in a sweep.

 

Then you add in the conservative nature in the way the bullpens are usually managed - most managers will manage more safely in the first game to avoid running through their entire bullpen, and I think it leads to more splits.

 

Most managers seem to almost manage for a split.

And also think about the Sox situation. They are going to play 14 games in 13 days. This isn't like any other sport where you more than likely are going to have a day off within the next couple of days. Of course you always want to try to win today, but you still have to be conscious about how the next 13 days will go.

[iamshack]

That's not entirely accurate. It kind of works like a Punnett square, if you can think back to high school biology. You've got team A and team B. You've got 4 possible outcomes: AA, AB, BA, BB. If the teams are evenly matched, then you've got 2 sweeps and 2 splits.

Oh you're right. I wrote out the possibilities and saw the 4 outcomes, but for some reason only cancelled out the second sweep outcome, since it was the same for the purposes of your argument as the first sweep outcome. For whatever reason, I didn't seem to think that was the case with the two splits.

 

 

[knightni]

That's not entirely accurate. It kind of works like a Punnett square, if you can think back to high school biology. You've got team A and team B. You've got 4 possible outcomes: AA, AB, BA, BB. If the teams are evenly matched, then you've got 2 sweeps and 2 splits.

For all practical purposes, the two different splits would equal the same sought after result.

 

25% sweep on either side, 50% chance of split.

[iamshack]

For all practical purposes, the two different splits would equal the same sought after result.

 

25% sweep on either side, 50% chance of split.

I think it's the case for the sweeps too though, Jer.

 

There are 4 possible outcomes.

 

1) Sox win game 1, lose game 2.

2)Tigers win game 1, lose game 2.

3) Sox win both games.

4) Tigers win both games.

 

For the purposes of this argument, the splits are the same, as are the sweeps. 2/4 = 1/2.

 

So it is really a 50% chance of a split or sweep.

[bighurt574]

That really is a huge statistical variation from what you'd expect. Put another way, a team that wins game 1 only wins game 2 25% of the time. You'd expect around 50%, maybe even a bit higher if one team is better than the other. I have a hard time believing mindset accounts for all of that, but I can't think of anything else. Maybe the team that wins game 1 is more likely to rest some regulars in game 2?

 

Is there any difference between day-night doubleheaders and where one game immediately follows the next?

Edited August 4, 2010 by bighurt574

[knightni]

I think it's the case for the sweeps too though, Jer.

 

There are 4 possible outcomes.

 

1) Sox win game 1, lose game 2.

2)Tigers win game 1, lose game 2.

3) Sox win both games.

4) Tigers win both games.

 

For the purposes of this argument, the splits are the same, as are the sweeps. 2/4 = 1/2.

 

So it is really a 50% chance of a split or sweep.

The sweeps produce two different results, though. A split is a split. But, if you want to split hairs (heh), I guess a game 1 win in a split or a game 2 win in a split, are two different things.

[iamshack]

The sweeps produce two different results, though. A split is a split. But, if you want to split hairs (heh), I guess a game 1 win in a split or a game 2 win in a split, are two different things.

He's not arguing for the purposes of which teams sweeps though. He's arguing that one would think there would be more sweeps (by either team) than the statistics show.

 

There can only be a sweep or a split, irregardless of which teams do which, right?

[southsider2k5]

The sweeps produce two different results, though. A split is a split. But, if you want to split hairs (heh), I guess a game 1 win in a split or a game 2 win in a split, are two different things.

 

But that is canceled out by the Sox winning game one and losing game two, being the same as the Tigers losing game one and winning game two.

[knightni]

He's not arguing for the purposes of which teams sweeps though. He's arguing that one would think there would be more sweeps (by either team) than the statistics show.

 

There can only be a sweep or a split, irregardless of which teams do which, right?

True. But splits are more likely.

 

Twenty five percent is a fair percentage because a sweep produces two different results, while a split is one result, two times.

[Greg Hibbard]

expectationally,

 

50% of the time a split should occur

50% of the time a sweep should occur

 

it doesn't matter who is doing the sweeping - 26.2% of the time it is occuring. Seems like an anomaly.

 

It can be argued that the anomaly is explained by the decisions regarding the second event being influenced by the outcome of the first event.

[knightni]

But that is canceled out by the Sox winning game one and losing game two, being the same as the Tigers losing game one and winning game two.

No, they aren't canceled. They are just duplicated. The other result would be Tigers winning game 1 and Sox winning game 2.

 

But both results are a same result, a split by two teams.

 

 

[iamshack]

The easiest way to understand it is just to write out all possible outcomes.

 

You will get 4, but for the purposes of this argument, the two different splits and the two different sweeps are the same result, which leaves you with only two possible results, a sweep or a split.