Nooo, don't do it. - Greg Fiume
The honeymoon is long over. The new administration has been in place long enough for some to begin crying for their dismissal. Even the sabremetrically-inclined fans that preach patience and large sample sizes are wondering if their new leaders are as numbers-based as they first appeared to be.
One nexus for these complaints is the use of the bunt. Is Terry Collins any less eager to use the oft-degraded tool than Jerry Manuel was? It seems that Collins loves the bunt like any conventional manager. Maybe a comparison of the two managers with respect to the bunt will uncover some of the differences in managerial styles.
The ideal bunt, if there is one, is completed late in the game when the team is down by a run. It may decrease the expectancy of scoring many runs - that's what happens when you give an out away - but it would increase the expectancy of scoring one run. The ideal bunt would be put down by a pitcher because any other hitter has a chance of being more productive with the bat. So let's look through the bunting (situations), the bunts (and their success) and the bunters that bunt them.
When have the Mets bunted this year compared to last year? Early, actually. 68% of their bunts have come in the fifth inning or earlier. Last year the team laid down 60.8% of their bunts that early in the game. If it's folly to give outs away for the most part, it's folly to give outs away earlier in the game.
Then again, that's not entirely fair - sometimes a game is close early. Leverage index can help us out here. In a close game with runners on, the leverage index goes up and tells us that the moment was important. This year, 73.7% of the sacrifice bunts have come with a leverage index of 1.5 or higher. Last year, that number was only 54%. So at least Collins is bunting in more important situations, even if those situations are coming earlier in games on average.
If you already have a batter in scoring position, it really doesn't make much sense to move him to third if it costs an out to do so. This year, 72.7% of the Mets bunts have come with a runner on first. Last year, that percentage was 64.8% - call this a win for Collins so far. Especially since, when you remove the squeeze bunts, those numbers change to 76.2% and 65.7%, respectively. Collins uses the bunt to move a runner into scoring position 3/4 of the time, and that seems decidedly okay.
One thing that seems crazy to do is to bunt when you're ahead. Why play for one run when you have that run already? Last year, 35% of the team's bunts came while ahead. Amazingly, if Baseball-Reference is to believed, 18.9% of their bunts came while up by two runs or more. Well, in one way, Collins fares worse here. 40.9% of the team's bunts have come while ahead this year. On the other hand, at least only 9% have come while up two runs or more (24.2% nationally in 2010). Both times, R.A. Dickey performed the feat - perhaps Collins doesn't trust a lead with the knuckler on the mound? At least Collins isn't as bad at bunting while up multiple runs.
Collins is calling for more bunts - .51 per game to .46 per game for Manuel - but it's a little early in the season to call that a trend and the difference doesn't seem huge.
A successful bunt should make it so that your team is more likely to win. But let's not be too harsh on the Mets here - there were 1543 sacrifice bunts last year, and only 147 of them added win probability to their team (9.5%). Of those 147, 60 actually only added value because the other team muffed the play. That means that only 5.6% of all sacrifice bunts last year would have added win probability to their team regardless of the play of the opposing defense.
Three of the 22 bunts (13.6%) Terry Collins has demanded have added value to the team, but two of those included errors by the opposing team. That means Collins' 4.5% rate is below the national average. Well, Jerry Manuel was worse. He only had two 'successful' bunts, but out of 74 completed bunts. That's a 2.7% rate, and both of them came on errors. There's hope still.
The very worst is when your manager asks your position player to bunt. It's just not a good idea. Players are supposed to be able to hit, and even the worst hitters have a better chance to be a positive influence on the game swinging away than with the sac bunt.
This year, 68.2% of the bunts have been perpetrated by pitchers. That's aight. It's also better than the 50% number from last year. Way better. The national average last year was 39.7%, so there are plenty of worse managers out there asking their position players to drop down the bunt. If this is the only takeaway, it does suggest that the Mets have at least this part of the game figured out.
Of course, we can quibble with some of Collins' decisions - why Jason Bay and Daniel Murphy have been asked to lay down bunts this year is a mystery. Angel Pagan and Jose Reyes, too. Maybe Jason Pridie and Chin-Lung Hu make more sense, but still.
Sometimes a single epic fail can stick and convince a fan of the stupidity of their manager. On April 20th, Terry Collins asked Josh Thole to bunt with Jose Reyes on first and no outs. By almost all the ways we've outlined here, that was a dumb idea before it was even attempted. That Thole flied out and Reyes was doubled off first just added insult to injury.
The problem is, it doesn't seem like Terry Collins has actually been that bad once you zoom out. He has bunted earlier in games than Jerry Manuel, but he's limited his bunts to higher leverage situations and has been more 'successful' on those bunts when measured by win probability added. He's asked his position players to bunt less often, and has bunted less often with a batter already in scoring position. It's those last two features of his approach that make him a more numbers-savvy manager with respect to bunts.
Even if he did once order a game-killing bunt by his catcher.