About those crazy graphs you’ll start seeing hereBy
Who’s ready for some Opening Day baseball? We at River Ave. Blues sure as hell are. There is nothing more pleasing than knowing that we’ve just begun seven months of baseball.
Hopefully, this nerdy and math-related concept won’t ruin the mood. Anyone familiar with my old blog, The Sporting Brews, is also familiar with Win Expectancy (or Win Probability Added [WPA], which is how I’m referring to it this season, because WPA is a better acronym than WE). For those of you unfamiliar, I’ll give the short short version.
In any given situation during the game, a team has a certain probability of winning. This is determined by the inning, score, runners on base, and outs. The probability of winning in each situation is determined by using historical data. That is, it looks at all games in MLB history in which any situation occurred and finds out what percentage of the time the home team won. For example, if the Yankees are up 1-0 in the top of the fourth, with Tampa Bay having runners on first and third with no outs, the WPA calculator goes back and searches for all games where the home team was up by one run in the fourth with runners on first and third with no outs. It then spits out the percentage of the time the home team wins. That would be the Yankees WPA at that given moment.
Individual players are debited and credited WPA, too. If Carl Pavano surrenders a home run in that situation, his WPA score changes based on the difference between the aforementioned WPA (up one run, fourth inning, first and third, no outs — 51.1%, incidentally) and the WPA that resulted from the play (27.8 percent): -23.3. If, say, Pavano induces a DP, but the runner scores, the WPA would change from 51.1% to 58.5%, meaning Carl would be credited 7.4%.
That may seem strange to some of you. After all, Tampa Bay just tied the score. How is it that the home team has a better chance of winning the game when they no longer have the lead? I could go into minute details, like the fact that they just took away two of Tampa Bay’s precious 27 outs, as well as two out of their super-precious three at a time. And I could mention how since 1977 that the home team has won 54 percent of the time, and that increases to 59 percent if they set the visiting team down without scoring in the first. But let’s cut to the chase.
There have been 8,109 games since 1977 in which the visiting team was batting in the fourth with two outs, the bases empty, and the score tied. Of those, the home team won 4,765, hence the 58.8 percent chance that the home team wins.
Yes, I understand that WPA isn’t a flawless system, as it doesn’t take into consideration who is at bat, who is on the mound, and who are the specific men on base. However, it does offer us an excellent historical perspective as to what the chances are during each game situation that your team wins. It’s also accompanied by a graph, sometimes filled with pictures or a bit of snark:
The graph, of course, tracks the Yankees chances of winning throughout the game. At this point, unfortunately, I don’t have the Excel savvy to mark each inning. I’ll work on it, though. Plus, I normally point out the bigger plays within the graph.
Questions? I know there must be questions out there. If so, leave ‘em in the comments section, or e-mail me: RABJosephP (at) gmail (dot) com.