April 17, 2012 at 12:32 am by Franklin Rabon under Atlanta Braves
Much has been made of the decision to walk David Wright with a 2-0 count to face Ike Davis, where Ike Davis subsequently hit a three run homer that drastically changed the game.
First, we must understand the game situation.
At the time the score was tied 1-1, in the top of the 6th inning. David Wright was ahead in the count 2-0 when the decision was made to intentionally walk him in order to face Ike Davis. There were two outs and Ruben Tejada was on third.
First, let’s just compare game situations with a man on third with two outs and a man on first with two outs. We will use the run expectancy charts in The Book: Playing the Percentages in Baseball .
The overall expected number of runs with 2 outs and a man on third in any inning is 0.387. The expected number of runs with 2 outs and men on first and third is .538. That comes out to 1.39 times more runs by having men on first and third. This indicates that in general, you better have a really darn good reason to intentionally walk a hitter in this situation, because the vast majority of the time it doesn’t make sense. You aren’t setting up a double play because there are already two outs. You are simply taking the bat out of one player’s hands, and putting the bat in another player’s hands in a more advantageous situation. You better believe that the player you walked was at least 1.39 times better than the player you will now be pitching to.
So, next, let’s see if David Wright actually was 1.39 times better than Ike Davis, given the situation. First, we’re going to ignore hitter-batter numbers, because those are unreliable due to sample size issues. However, we will look at Davis and Wright’s splits against righties, which should be useful, since Hanson typically is a fairly normal RHP with regards to his lefty right splits, using wOBA, probably the single best pure measure of hitter ability.
Over Wright’s career, he has a .366 wOBA against RHP. Ike Davis is a .358 career wOBA guy against RHP. While these differences are real, they aren’t substantial, and certainly come nowhere near the 1.39 we needed, given the change in base-out state we talked about above, this difference is very close to being statistically insignificant. ie for all intents and purposes, trading the righty Wright for the Lefty Davis is a wash, given a RHP.
However, that isn’t the end of the analysis, because Wright wasn’t 0-0 when the decision was made. If the decision had been made to walk Wright right off the bat, it would have been a very poor decision. Trading virtually equivalent hitters, but putting him in a place where, on average, more runs will be scored. We also must consider that Wright was in a hitter’s count at the time. To see that, we will now see what increase in productivity Wright see’s after he works himself into a 2-0 count.
Over David Wright’s career, he has been an astonishing .485 wOBA hitter any time he manages to work a 2-0 count (this means all counts including and after a 2-0 count is reached). Given that .340 is a roughly league average type number for any count, after 2-0 counts wright is an extremely dangerous hitter. However, what is important isn’t how Wright fared in 2-0 counts, but how Wright fared in 2-0 counts to righties. As it turns out, the best way to figure this out is simply take his 2-0 count and after delta (ie how much better he did after 2-0 counts than all counts) and add that to his overal RHP wOBA. We find that Wright’s 2-0 and after wOBA is .101 higher than his overall wOBA. Because this delta for most players is the same regardless of what handed pitcher they are facing, we will then add it to his RHP overall wOBA. As stated above, Wright’s wOBA against RHP is .366, meaning that trading Wright in a 2-0 count for Ike Davis is effectively like trading a .467 wOBA hitter (wright after 2-0 counts against righties) for a .358 wOBA guy (Davis against righties). That is quite obviously no longer an insignificant difference.
Now things have gotten a lot murkier. We are trading a much better hitter (David Wright in a 2-0 count against a RHP), for a worse hitter (Ike Davis against a RHP), but the worse hitter is in a more advantageous situation.
Is the 109 wOBA point difference between the two worth the 1.39 times more runs that are scored on average in the resulting situation? To answer this question, we must move to wRC+, a number based on wOBA that gives an idea of what percentage more runs than average given players will score.
Ike Davis is a 126 wRC+ v. RHP. This means he is worth 26% more runs than a MLB average player. Using similar methods as to what we used with wOBA above, we get that Wright is likely a 191 wRC+ player in 2-0 counts against RHP, ie 91% better than a league average hitter in any count in any situation. So, in all base/out situations we figure that Wright is worth 65% more runs than Davis, given the count Wright faced at the time. Let’s apply these numbers to the given base/out states we had when the players were up:
Davis: 126 wRC+ applied to .538 expected runs from an average player given men on 1st and 3rd with 2 outs = .678 expected runs.
Wright: 191 wRC+ applied to .387 expected runs from an average player given a man on third with 2 outs = .739 expected runs.
So, it does look like Fredi made the right call, though on average the difference is .06 runs (a relatively insignificant number). Ultimately, a large exercise and a lot of words and thoughts to describe a decision that really was a wash either way. We made a few assumptions in getting to these numbers, but the point remains that given the statistical tools we had at the time, it’s hard to really question the move, even though it didn’t work out. I hesitate to say it was the right move, given how insignificant the differences were, but by no means was it clearly the wrong move. It was just a move.