clock menu more-arrow no yes

Filed under:

Brutal bunting: Taking away more runs from the league's worst offense

New, 65 comments

Bunting is usually harmful to scoring runs, and that's not something the Braves can really afford given the state of their offense.

When Freddie bunts for a hit, it's okay. This post is not about that.
When Freddie bunts for a hit, it's okay. This post is not about that.
Dale Zanine-USA TODAY Sports

This post was initially going to go up prior to the Braves relieving Fredi Gonzalez of managerial duties, and may have been more fitting prior to Brian Snitker's ascension as interim manager. Among other things, Fredi Gonzalez often took flak for relying on bunts, though in reality, he did not do so much: up until 2016, Braves' position players did not bunt* notably more than their counterparts on other teams. Only in 2016, with an embarrassingly bad cast of offensive characters, did Fredi Gonzalez begin to utilize the bunt with more frequency.

* When I say "bunt" in this article, I'm referring to a sacrifice bunt, rather than bunting for a hit. Bunting for a hit is basically an alternate way to attempt to reach base on a single (or in rare cases, a double), and doesn't figure into this calculus.

Through yesterday's game, the Braves have recorded seven position player sacrifice bunts, tying them for fourth place in MLB. (The teams ahead of them are the Angels, Phillies, and Rockies, which prompts all sorts of head-scratching.) Mallex Smith and Erick Aybar have three sacrifice bunts each, while Ender Inciarte, who's missed much of the season so far, has one.

I got the idea to look closely into bunting patterns when Ender Inciarte bunted in the series opener with the Pirates. It was the top of the first, and after a leadoff walk to Nick Markakis, the Braves opted to freely yield their first of 27 game-given outs to move Markakis over to second. Markakis did not end up scoring, as Freddie Freeman and Jeff Francoeur made outs to strand him at the keystone. As most of you can probably figure, even if you don't agree, I think that bunt was a really dumb move. In fact, in most cases, successfully bunting lowers a team's chances of scoring runs. For some teams, the lowered chance is not an impediment: they have good offenses that can overcome bad decisions. For the Braves, who are hurtling towards a historically pathetic offensive performance, there is zero excuse to further minimize the team's chances of plating a run in any situation.

Let's take a look at the Inciarte situation in more detail. Across all 2010-2015 teams and situations, the run expectancy (the average total number of runs scored in and after the situation in a given inning) for man on first, no outs (referred to in shorthand as 1xx_0) is .859. That means, on average, about 0.86 runs score in every inning after that situation occurs, not counting runs that have already scored in that inning. Moving that runner over, the run expectancy drops to 0.664. The expected runs tally drops by over 20%, just by giving up the out.

This is even true if you're thinking about just playing for one run (which the Braves had no reason to do in the top of the first). In all innings with none out and a runner on first, there is a 42% chance of one or more runs scoring, but after a successful bunt, these odds drop to about 40%. It's not as big of a drop as the total average number of runs scoring, but experience shows that bunting doesn't even help to get a single run on the board in that situation. (Some of this might be confounded by the fact that teams who feel they need to bunt in that situation are just worse offensively, but that should reinforce the intention to not be like those teams.)

All these run expectancy data are from Tom Tango's repository of information, available here.  The graphic below illustrates possible successful bunt scenarios, and how they generally harm run expectancy.

Some bunts are worse than others, but all of these bunts reduce run expectancy, based on actual historical outcomes. This pattern even holds if we talk about the chances of scoring at least one run, which is a useful lens to look through if it's a close-and-late situation, such as a potential walk-off run or go-ahead run in extras.

Essentially, you can see that in game-winning run situations, it makes sense to bunt that run over to third, so long as you do so with zero outs. Otherwise, you're still better off doing something other than giving up a precious out to move the runner along a single base.

Part of the complication here, of course, is that all of these run expectancy tables from Tom Tango aggregate a whole bunch of data from different offensive teams and spread across all sorts of offensive players. In reality, though, each manager has to make decisions about the batters in his lineup. Luckily for us, Fangraphs just debuted a really fun tool that shows run expectancy, sliced-and-diced by wOBA (an aggregate offensive stat similar to wRC+, but scaled to OBP rather than to 100) and by run environment. This lets us examine bunting decisions in an even more tailored context.

Let's go back to the Markakis-Inciarte example.For his career, Inciarte has a .311 wOBA, but that falls to just .261 against lefties. The batter behind him, Freddie Freeman, is at .362 and .326, respectively, indicating he's a much better hitter than Inciarte, especially when he lacks a platoon advantage.

Per the run expectancy tool, assuming an average offensive environment, the run expectancy before a bunt, with Inciarte at the plate, is 0.79 runs. (It would be 0.66 if we used Inciarte's against-lefties wOBA, rather than his overall wOBA). Post-bunt, the run expectancies are 0.72 (Freeman vs. everyone) and 0.63 (Freeman vs. lefties). So it's not quite so egregious of a difference as the general data, but still not a great idea.

But! The above assumes an average run environment. The 2016 Braves, however, play in anything but an average run environment, due to their offensive ineptitude. They are currently last in the league, averaging 3.2 runs per game, and it was closer to 3.0 before the veritable offensive explosions of the first two games of the Pirates series. Unfortunately, the run expectancy tool isn't tailored to this level of punchlessness, so the lowest available run environment is 3.7. Still, if we use that baseline rather than an average one:
  • Against the world, the run expectancy with Inciarte bunting and Freeman hitting behind him falls from 0.71 to 0.67.
  • Against lefties, the run expectancy with Inciarte bunting and Freeman hitting behind him is about unchanged at 0.58. (It still falls, just from 0.583 to 0.579.)
There are two ways to read this. One is that the worse a team's offensive performance, or the expectations for any individual hitter or series of hitters on a team, the less the decision to bunt or not matters, since the odds are stacked against any runs scoring anyway. But, on the flip side, the harder it is to score runs given a team's talent level, the more critical it becomes to squeeze out any drops of tactical oomph in order to maximize your team's offensive potential.

What's potentially interesting here is that the Braves rarely bunted (correctly or not), when they had powerful offenses and all-around strong teams in 2012 and 2013. The bunting increased when the offensive performance was worse in 2011, 2014, 2015, and especially during the first six weeks of 2016. This suggests an actual vicious spiral: the worse the offense, the more desperation there appears to be to score any runs, which leads to suboptimal tactical decisions. Rather than bunting, even weaker offensive teams would be better served by trusting their (poor offensive) players and letting it ride.

A few more minor notes about this analysis:
  • There are various ameliorating or worth-considering factors when it comes to bunt tactics, such as the possibility of reaching on a bunt intended to be a sacrifice, the potential for an error, and avoiding a double play. All of these are valid concerns, but the run expectancy matrices actually take all these occurrences into account, and are kind of the "final result" after all these different possibilities have been tallied. In any specific situation, it's important to consider the defense in play, the speed of the hitter, the leverage of the situation, and so forth, but in general, bunting seems like a bad call much of the time.
  • Another reason why bunting can be suboptimal: at best, as shown above, a bunt reduces run expectancy. But bunting is not a 100% guaranteed success, and the pre-bunt run expectancy already incorporates all outcomes that are essentially botched bunts (as well as incorporating all successful bunts, just like the post-bunt run expectancy figure incorporates everything that happens after a successful bunt). In other words, while a good bunt lowers run expectancy, a bad bunt lowers it even more, and there's no guarantee you'll get a good bunt. When letting a hitter swing away, they'll reach base in accordance with their true talent level OBP, when asking for a sacrifice bunt, you're running the risk of decreasing run expectancy even more than what a matrix suggests.
  • Bunting with pitchers can make sense because their wOBAs/OBPs are so low that you at least get something out of their PAs if you bunt, as opposed to a mode outcome of a strikeout. In the same vein, bunting ahead of your power hitters is a bad idea, as power hitters are more likely to hit balls in such a way that makes the sacrifice bunt redundant. This is the same reason you generally don't want to steal ahead of power hitters. Bunting ahead of singles hitters (like perhaps one's leadoff hitter, or number nine hitter) might make sense in some contexts, but bunting ahead of Freddie Freeman will generally be a bad move.
As noted earlier, the Braves have actually bunted with position players to move a runner over six other times than the one analyzed above. In addition, there have been a few botched/failed bunts not recorded there.
  • Runner on first, none out, Mallex Smith bunts against a righty with Markakis following him. It's hard to predict what Smith will or won't do in this situation given that we have no idea of his true talent level, but given the way he's hit righties so far (130 wRC+, no doubt propped up by his two-homer game against the Pirates), this was a huge run expectancy sink. It's less of one, but still notable, even if he was worse against righties. The game was tied, but it was only the fifth inning, meaning that one run was not particular meaningful there, even if the bunt call did somehow increase run expectancy there.
  • Runners on first and second, none out, Mallex Smith bunts against a righty with Erick Aybar following him. Given how bad Aybar has played, this was absolutely brutal, and Aybar hasn't really hit well in his career in general (.304 wOBA against righties). It was a 2-0 game at that point, and, unsurprisingly, Aybar hit a sacrifice fly and the Braves were unable to get a two-out hit to tie the game in the seventh.
  • Two games later, another Smith bunt, again with a man on first and none out, this time ahead of Daniel Castro. The Braves were down by one in the fifth, and given that Castro was the beneficiary, another weird call to make.
  • Aybar's bunts were also very silly. The first was in a tie game against the Cardinals in the 7th. He bunted a leadoff walk over to second, but this just led to an intentional walk of Freddie Freeman. (Again, stop bunting ahead of power hitters.) The Braves didn't score. The second came with two on and none out, and again got Freddie Freeman intentionally walked. The bunt kind of set up two add-on runs against the Cubs, but the Braves were already leading, and those runs only occurred due to infield errors and miscues by the Cubs. The third ended up being pointless, as in a tie game, Aybar moved Smith from second to third by making the first out of the inning. The next batter walked, and then Inciarte hit a double, making the base advancement entirely moot.
So yes, if you're keeping score at home, bunting with position players has yet to help the Braves so far in 2016. They've been among the league leaders at it so far, but let's hope we see a lot less of it now that a managerial change has been made.