Papi vs. A-Rod
Given all of the talk about the AL MVP, I was hoping that I could make my own quantitative, objective comparison of the raw value that both David Ortiz and Alex Rodriguez have contributed to their respective teams this season. Unfortunately, I have failed (Gah!). Actually, my original intent was to calculate the total offensive WPA (Win Probability Added) for both Ortiz and ARod throughout the course of the entire season. I mentioned this statistic in an earlier post about closers and I would also suggest this site as a good starting point if you’re interested in exploring the stat yourself. In short, WPA is a metric designed to quantitatively measure a player’s “value” or significance of his overall contribution to the team’s performance.
Think of it this way: at any one particular point during a game, you can describe the game situation in terms of inning, outs, runners on base, and the score. For each one of these game situations, we can assign a probability of victory for either team. There are many strategies for coming up with these probabilities, but perhaps the easiest and most straight-forward method is to use historical data and directly use the percentage of times a team has won the game given it has been in a particular game situation. This site is an example that uses such historical data from games played between 1979 and 1990. In any case, when a hitter comes to the plate, the game is in a certain game state as described, and his team has a certain probability of winning. After his plate appearance, the game is in a new game state and the probability of winning is a new probability. The change in probability from before his plate appearance to after can be directly attributed to the result of his at-bat. Thus WPA accumulates these changes over all at-bats throughout the course of the season.
To make it even simpler, WPA gives more weight to what many would refer to as “clutch” hits. For example, hitting a walk off RBI single would be a significantly greater boost to a hitter’s WPA than a grand slam in the top of the 9th of a blowout. Conversely, popping out with the bases loaded in a tie game would be more of a negative impact to WPA than grounding into a double play with a man on first when you’re down by 10 runs. Thus WPA does not necessarily accurately rank players according to their talent or their raw quality of performance during a season. It rather measures the overall timeliness of a player’s hits as they are important to the outcomes of each particular game. (Whether or not this “clutch” ability is due to anything other than pure randomness is an entirely different debate however)
At any rate, given Ortiz’s flare for the dramatic and unusually high quantity of late inning heroics, I suspect that his WPA would be significantly higher than A-Rod’s. Even though, looking at their raw numbers, A-Rod had a slightly better pure offensive season (not to mention he also has some value as an above average third basemen), I still think that Big Papi’s “big hits” meant more to the Red Sox and makes him the Most Valuable Player in the truest sense of the word. Alas, though, I was unable to efficiently calculate the WPA’s for both players. In order to accomplish this, I would have had to track the WPA over every single at-bat and the only play-by-play data that I could find for every game of the season were in the game recaps provided on ESPN.com. I started to try and do it by hand, but it just took too long and I could not find a usable data set that could be automatically parsed using a reasonable amount of programming effort. Regardless, though, I still think that my suspicion holds true (about Ortiz), and I would like to see WPA become more a part of the mainstream SABR community.
December 1st, 2005 at 1:13 am
The MVP race came down to A-Rod & Ortiz. The defensive case was made lound and clear as A-Rod edged out Ortiz in a very close race. I was hoping big papi would have won.