The Probability of an MVP

In recent weeks, as Ryan "The Clean in the Clear" Howard's performance on the field has reached new heights and as the Phillies have continued to flirt with the National League Wild Card race the same way a twelve year old boy flirts with a girl -- able to pull the pigtails but not sure what to do once he gets the desired attention -- there are growing discussions of Howard's MVP-worthiness.

Those who support Howard will point to the fact that without Howard leading all of baseball in home runs and runs batted in by large margins and his seventh best batting average in the National League, .316, the Phillies would be trying to decide between Myrtle Beach or Vegas right now for their October getaways.

Howard's detractors will bring up the fact that while he is certainly impressive, Albert Pujols is still a better player on a better team, and they might even say something similar about Carlos Beltran.  He's the best player on the best team in baseball.

But these are all subjective arguments.  He means more to his team.  No, he means more to his team.  Yada, yada.  Even the home run and run batted in totals are subjective to a certain extent as they are influenced by the number and types of opportunities a player has almost as much as talent influences the totals.

One statistical measure, however, is more of the objective type of thing we are looking for:  win probability added.  This is a statistic that attempts to quantify how much a player has tilted the balance towards his team in a game.

Using win expectancy tables based on Retrosheet.org's 1979-2004 data, one can get the probability of a team winning a game given a certain situation.  For instance, if a team has the bases loaded in a tie game in the bottom of the ninth with two outs, that team will win the game, on average, 66.4% of the time, or .664.  [Regular readers of this site have seen graphs of these probabilities for various games posted here.]

Staying with that example, if a player were to get a hit in that situation and win the game, he would obviously increase his teams chances to 1.000.  When calculating win probability added, that hitter would then be credited with .336 win probability points.

Of course, as with any baseball statistic, size matters.  Sample size, that is.  After nearly 150 games of play, with probability points added up for the entire season, some reliable trends have certainly emerged.  It would be safe to say that by now, the more probability points a player has, the more he has helped his team.

To this point in the season, Ryan Howard leads all of baseball with 7.66 win probability points.  Albert Pujols is close on his heels with 7.55 points while Carlos Beltran is a distant third with 4.82 [incidentally, they are not just the top three in the NL, but in all of baseball too].  Using WPA as a measuring stick, we are now looking at a two-man race for the National League MVP.

When an MVP race is perceived to be as close as this one is by voters, every bit of information helps.  If voters are open-minded enough, win probability added could tip things in Ryan's favor.

-----------------

More on win probability added:

The One About Win Probability [The Hardball Times]
WPA Leaderboard [Fangraphs.com]