Friday, July 11, 2008

A new 2007-08 streakiness study

If you're sick of clutch hitting studies, you're probably *really* sick of streakiness studies. But here's a new one from "Eric Karros" of "Sons of Steve Garvey."

He found that after winning one or two consecutive games, teams were slightly more likely to win the next. But after winning three, four, or five consecutive games, they won the subsequent game far less often than you'd expect from their record.

Bottom line: like many other previous studies, no evidence of "momentum" or a "hot hand."

Some of this might be the pitching rotation: teams that win four in a row are likely to have done it with their top four starters, and their fifth starter is a worse pitcher, so more likely to lose. But that doesn't explain the five-in-a-row effect.

Hat tip: Carl Bialik

Labels: , ,


At Saturday, July 12, 2008 10:36:00 PM, Anonymous Anonymous said...

Similar findings apply for NHL hockey:

At Saturday, July 12, 2008 10:42:00 PM, Anonymous Anonymous said...

On the above link, the end of the URL may be cut off. There should be a ".pdf" at the end, coming after "streaks".

At Sunday, July 13, 2008 11:15:00 PM, Blogger birtelcom said...

What may also be causing the decline below .500 after a series of wins may be home game wins followed by road losses. Home teams have had over the year an approximately .540 to .460 win percentage advantage over road teams. Winning streaks will thus tend to happen more often at home, but as a series of home game wins goes on, the likelihood that the next game for the streaking team will be a road game increases -- home stands only last so long. I think these factors could be expected to cause a perceptible increase in the likelihood of a loss after a win or series of wins if one uses a large enough sample.

At Sunday, July 13, 2008 11:17:00 PM, Blogger Phil Birnbaum said...

Birtelcom: Yup, that makes sense.

At Monday, July 14, 2008 12:17:00 PM, Blogger Tangotiger said...

I agree, that there are many biases here that would cause the effect, and not the "preceding streak" being the leading cause.

Baseball is played in 3-game or 4-game sets, and so, after 3 straight wins, it's possible that the 4th game is played against a different opponent. 4-game streaks must happen fairly infrequently that it wouldn't pass the stat sig test after just 2 years.

The home/road is an effect, as is the pitchers and closers. That is, if you win 3 straight games, you may have used your closer twice and so is not available.

Really, there is a host of such biases to account for.

At Monday, July 14, 2008 5:28:00 PM, Blogger Eric Karros said...

Hi Phil et al. This is "Eric Karros" from Sons of Steve Garvey. Thank you for reading our blog and my post.

I just want to say that I pretty much agree with Carl Bialik, birtelcom, and tangotiger: that 5th starters, starting a road trip, and facing a different opponent are all valid reasons not directly related to either luck or the quality of the team as to why a winning streak could end. The structure of the analysis allows for too much one-sided systematic error is retained. And even though the exact opposite phenomenon also occur (top of rotations are stronger, opposing teams have weak 5th starters, teams come home after road trips or start a new series against a crappy team), these impacts won't be captured as strongly given the way the analysis is set up.

I still think winning streaks are waaaay overrated though. You'll have to trust me on this one.


Post a Comment

<< Home