Not everybody is familiar with the win probability graph I included as part of yesterday’s game recap. As Samir summed up quite succinctly in the comments, the graph shows each team’s likelihood to win the game after each play. Basically, people with much, much better database and analytic skills than me compiled data on thousands of games, and based on the score and state of the inning (how many runners, how many outs) and the way games have historically played out, we can say what chance each team has to win the game assuming that each team is a league average team. That assumption is key – at the start of the game, we say that the Yankees have a 50% chance to beat the Orioles, even though the probability is probably closer to 60% in favor of the Yankees. Because of that assumption, though, we can derive a statistic called WPA, or “Win Probability Added,” to show how much better players and teams are than the average player.
For more on the subject, I recommend this article by The Hardball Times’ Dave Studeman.
For an excellent example of Win Expectancy and WPA in progress, check out this article by one of the leaders in Sabermetrics, Tom Tango. It takes you through the Bartman inning play by play, so even if you already know everything you want to know about WPA, it’s a fun read just because it allows us to relive the awesome Bartman game.
I’m a sabermetric writer, so I will be using some advanced statistics. First, here is the Glossary from FanGraphs, which is a good starting point. If you would like to learn some more of these statistics and concepts relatively quickly and easily, here are two excellent resources that have appeared in the blogosphere this offseason.
By Steve Slowinski, another excellent course on Sabermetrics.