Tuesday, July 01, 2014

Do Walk-Off Wins Create Momentum?

Last night, the Tigers were down 4-1 in the bottom of the ninth and many of you had probably gone to bed for the night or were doing something else thinking the game had been lost.  After all, there wasn't much chance against untouchable Athletics closer Sean Doolittle.  I was only half listening myself at that time, instead looking up the results of one of my fantasy teams.

I knew the Tigers were threatening though and Doolittle wasn't his usual dominating self having just walked slumping outfielder Austin Jackson (just Doolittle's second walk of the season) to load the bases with one out.  So, I looked up at the television screen to see Rajai Davis at the plate and Ian Kinsler on deck.  It never occurred to me that Davis might be the hero and I was mostly waiting to see what Kinsler would do. Then Davis shocked us all by connecting on a hanging slider for a line drive home run to left field, a walk-off grand slam.  

It was the most improbable and exhilarating victory for the Tigers so far this year, but what does it mean for the future?  Does the walk-off win make it more likely that the Tigers will win tonight?

It is widely believed that a walk-off win or sudden victory creates a boost for a team that carries over to the next game more often than not. I wanted to see if this was true (I actually did this study last year after the Red Sox walked off the Tigers in Game one of the American League Championship Series).  So, I went to the retrosheet database and found all walk-off wins between 1995 and 2012. There were 3,769 of these sudden victories during the period which comes out to about seven per team for a season. My goal was to see if walk off winners had a tendency win the next game after their walk offs.

Twenty-eight walk offs fell out of the sample because they occurred in the final game of the season and thus were not followed by another regular season game. That gave me 3,741 games with which to work. I discovered that teams won 52.6% of the games immediately following walk off wins. That's more than 50% so at first glance it seems that there is a bit of a tendency for teams to win games following walk offs.

Before jumping to conclusions though, there are a couple of important factors to consider. First, walk off wins only happen at home so chances are (85% of the time to be exact) that the next game would also be at home. Since teams win 53.9% of their home games, you would expect them to have a good winning percentage in games after walk offs even without a carryover effect. Also, teams with good winning percentages tend to have more walk off wins. For both of these reasons, one would expect a win the next game after a walk off more than 50% of the time even if walk-off wins had no influence on future games.

I calculated the expected winning percentage in games after sudden victories as follows: For each walk off, I calculated the winning percentage of the specific team for that year and site. For example, the 2012 Tigers won 61.7% of their home games. Thus, they would have a probability of .617 of winning a home game the day after a walk off assuming no carryover effect. I did this for each of the 3,741 walk offs and then calculated the average probability to be .528. This means, that assuming no carryover effect, we would expect 52.8% of the games after walk offs to have resulted in wins.

Since the expected winning percentage (52.8%) for games after walk offs was almost exactly the same as the actual winning percentage (52.6%), I can conclude that, in general, a walk-off win has no effect over the result of the following game. As cautious as I am about accepting intangibles, this result is still surprising to me. I was not expecting a dramatic effect but I thought that sudden victories would have a small influence over ensuing games.

Another question is whether walk-off losses create negative momentum.  This is, of course, relevant to the Tigers-Athletics scenario since the walk-off winner is facing a a walk-off loser in game two of the series tonight.

As with walk off wins, there are there are two factors to consider. First, walk-off losses happen on the road and are followed by road games 86% of the time. Since teams win just 46.1% of their road games, you would expect them to have a low winning percentage in games after sudden losses even without a carryover effect. Also, teams with low winning percentages tend to have more walk off losses. For both of these reasons, one would expect a win in the next game after a walk off loss less than 50% of the time even if walk-off losses had no influence.

The expected winning percentage in games after sudden losses is calculated the same way as for walk-off wins.  It turns out that a team would be expected to win 45.7% of the time after a walk-off loss if there is no carryover effect.  Since, this is almost the same as he actual winning percentage of 46.4%, I can conclude that,  in general, a walk off loss has no affect on the result of the following game. 

Based on these analyses, the Tigers thrilling victory should not affect their performance in game two any more than any other result. Tigers former manager Jim Leyland's theory was that "Momentum is only as good as the next game's starting pitcher".  So, let's hope that emerging stalwart Rick Porcello is on his game tonight versus the Athletics.  

1 comment:

  1. AnonymousJuly 01, 2014

    Well perhaps there still are some benefits to miscellaneous successes breeding future success, but to the extent that it does not then baseball would be nothing more than a game of probability events colliding with each other. The same can be said for the spin of a roulette wheel, and despite one that may have superstitious beliefs it is not more likely for the next spin to be black if the previous 10 spins just happened to be red.

    Every game is a new game with a new set of random rolls with the odds being represented by what a player is truly worth yet tricky to objectively define. if a player can find a way to hone their mental focus to shore up the odds, then more power to them and they all hopefully strive to do that anyway.

    Other than that, the name of the game is getting the best odds potential players in place for each given situation and trying to build a higher overall expected value of team output more than the opponent can. I still think we could do a better job though in the architecture and design of those pieces to generate more efficient leverage, and in turn a higher winning percentage.

    ReplyDelete

Twitter

Blog Archive

Subscribe

My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books

Stat Counter