![]() |
(Photo courtesy of Michael Allen Blair/News-Herald) |
I hate bunting. It doesn’t take long to find that out about
me. There are very few situations in which a bunt is the optimal course of
action. The third inning is never one of those times. Roberto Perez is not
going to be exempt from this rant, even though his bunt became a ROE (reached
on error) that subsequently resulted in a completed sacrifice bunt by Jose
Ramirez.
The game in question occurred on Wednesday afternoon. The
Indians, mired in a four-game slump, were facing left-hander John Danks. For
those unfamiliar with John Danks, he’s not very good and he is, laughably, the
highest paid player on the White Sox roster this season. Danks gave up a
leadoff single to left-handed batter Lonnie Chisenhall. At this point, the
Indians win expectancy in a 0-0 game with Chisenhall at first and nobody out is
59.3 percent. Thanks to John Danks’s throwing gift, Roberto Perez was safe on
his ill-advised bunt attempt.
The Indians had runners on first and second with nobody out.
Their win expectancy was 65.5 percent. That means that they had a 65.5 percent
chance of winning the game. At this point, in the third inning, facing
switch-hitting Jose Ramirez, somebody thought it was a good idea to bunt.
Whether that somebody was Ramirez himself or the call came from Terry Francona,
the sacrifice was successfully completed. Yay! Right? So, how much more likely
were the Indians to win because Ramirez advanced the runners? They were not
more likely to win. The sacrifice bunt did not help nor hurt the Tribe’s
chances of winning. Their win expectancy remained 65.5 percent.
Michael Bourn mailed in his at bat to hit a ground ball to
second on the first pitch he could handle for an RBI groundout. That bumped the
Indians win expectancy all the way up to…wait for it…67.1 percent! In that
sequence, the Indians improved their chances of winning by 1.6 percent.
At this point, you’re probably asking yourself how that’s a
bad thing. For starters, the White Sox loaded the bases in the top half of the
fourth inning and the game was essentially a 50/50 proposition. When Alexei
Ramirez scored to load the bases, the Indians were at 50.1 percent to win the
game. Fortunately, JB “Aw” Shuck(s) popped out foul and Trevor Bauer threw a
dart on the outside corner to end the inning without harm. When the Indians
came to bat again, their win expectancy was 69.6 percent.
If the bunt helped us win, why was it bad a decision? Is
that the question you’re asking? It should be. The answer lies in the concept
of run expectancy. Based off of the tremendous work by Tom Tango, run
expectancy is a chart that illustrates how many runs a team will score, on
average, given a certain situation. The situations are broken down by number of
runs, the bases the runners are occupying, and the number of outs in the
inning. Unfortunately, Tango’s work only goes up to 2010, but Baseball
Prospectus has a neat little run expectancy chart updated year by year.
Let’s look back at the situation in the third inning and
start from the top. With nobody on base and no outs, the average runs scored
per inning in that situation is 0.4552. Lonnie Chisenhall singled. That raised
the run expectancy to 0.8182 runs scored for that inning. Roberto Perez put
down a bunt that was intended to be a sacrifice, but John Danks could not
complete the out. With runners on first and second and nobody out, the average
number of runs scored in that inning is 1.4023. (As a side note, if Perez is
thrown out at first, the run expectancy drops to 0.6235 runs; so the sacrifice
bunt was detrimental to scoring runs.)
Now, the play in question. Jose Ramirez is at the plate for
his first at bat of the day. The Indians are projected to score 1.4023 runs in
the inning when Ramirez steps into the box. Ramirez bunts. The expected runs
scored on average in that inning drops
to 1.2714. Remember the old adage, “If you play for one run, you get one run”?
The Indians scored exactly one run and it nearly didn’t matter when the White
Sox came to bat again.
Let’s put on our hypothetical thinking caps. Had Ramirez
made an unproductive out, where the runners stay put, the expected runs scored
in the inning falls to 0.8623. This is bad. Had Ramirez stood there while Danks
unloaded a wild pitch that advanced both runners, the run expectancy jumps to
1.8707. Had Ramirez walked or singled without a run scoring, the run expectancy
climbs to 2.2337. If Ramirez singles, the Indians get a run and are set up to
score an average of 1.4023 more runs in that inning. If Ramirez doubles, the
Indians score one run and score another 1.8707 runs on average.
Context matters. The math is there, but not all situations
are created equal. One can make the argument that Jose Ramirez is the
ninth-place hitter, so putting runners in scoring position makes sense, since
he is, in theory, the worst hitter on the team. This would make sense, if he
were not bunting in front of Michael Bourn and Mike Aviles. These are two below
average hitters. That’s an argument for another day, one that one of my
colleagues is sure to write about soon. Also, there’s a case to be made for
Danks making a mistake pitch, either from being pissed off about the error or
from losing mental focus from having to field and then get back on the mound
and face a hitter in a mess that he created.
The overall point to make here is that the Indians settled
for one run when they were projected to score more than that. There were a lot
of positive outcomes and only one negative outcome. Even a fielder’s choice
that places runners at first and third with one out carries an average of
1.1261 runs per inning, which is not that bad of an outcome, all things
considered.
The Indians took a 1-0 lead and won the game 4-2. There’s
more to it than that. There’s a chance that the Indians could have taken a
commanding lead of the game. Unfortunately, when The Hardball Times re-designed
their website, they took down one of my favorite tools on the internet, their
Win Probability Calculator. The best I can find is this Win
Probability Calculator that uses Retrosheet box scores from 1957-2013,
minus the 1999 season. This percentage differs a bit from Fangraphs, but the
concept will be the same regardless. It differs because it is based off of
actual results, not mathematical expectations.
A chart is probably the best way to show this:
Outcome
|
Win Prob (%)
|
68.6
|
|
Sac bunt
|
67.1
|
Wild pitch/passed ball
|
77.9
|
Walk/Single (No RBI)
|
78.1
|
Double
|
84.6
|
Triple
|
89.2
|
Home run
|
86.4
|
Unproductive out
|
64.3
|
Fielder’s choice 1st & 3rd
|
68.8
|
Fielder’s choice 1st & 2nd
|
64.4
|
Remember, this is based on actual results over a 55-season
sample size. It’s not hypothetical math. The home run lowered the expectancy
because there is a large sample size of games in which the home team held a 3-0
lead in the third inning. There’s a little bit of small sample size bias in
some outcomes, but not a lot.
It’s also important to keep in mind that it is the third
inning. The opposition still has at least 18 outs to use. The effectiveness of
a bunt and the value it carries would be different in the late innings. There
are a handful of situations where bunting is a good decision. In the third
inning, with six innings of baseball left to play, it’s hard to make any kind
of case for it to be the best course of action. Even if the Indians don’t score
in the third, a 0-0 game into the fourth gives them a 58.8 percent win
expectancy based on the sample size in the win probability calculator.
More often than not, if you look at the math, a sacrifice
bunt is not worth it. Giving up outs, when they are so precious, is a bad idea.
If you need to play for one run in the late innings, it might be worthwhile.
When you play for one run in the early innings, it’s often a misguided
strategy.
There’s another element to this that is hard to quantify,
but the Indians are still terrible defensively. How often is one run going to
stand up? Traditionalists will argue that there’s a mental component to scoring
first or taking the lead. And, yes, teams that score first win more often than
teams that don’t. But, in this particular situation, you’re making a lot of
assumptions about the outcome of Ramirez’s plate appearance. The Indians may
have scored first regardless of the outcome and Michael Bourn may have had a
better at bat rather than just putting a ball in play to get the run home.
Technically, the Indians gave up two outs to score one run. Over the long haul,
that’s not a winning strategy.
The Indians won the game and that’s what matters. Still,
this is an exercise worth looking at. If the Indians would have lost, these are
the situations you point to and wonder what could have been. What if Ramirez
had swung away instead of giving up an out? Would the Indians have scored two
runs? Three? More than that?
Let’s briefly look back over the last few seasons and see
how teams that led the league in sacrifice bunts did in terms of run scoring.
The National League is silly and still lets pitchers hit, so of course they’re
going to skew this a little bit. Because of that, I’ll look only at American
League teams, since bunting strategies are a little different in the NL.
The Indians (surprise surprise!) led the American League in
sac bunts last season. They finished seventh in runs scored. The Tampa Bay Rays
were next. They finished dead last, 15th, in runs scored. The Texas Rangers
were next. They finished 10th. In 2013, the top three AL teams in sac bunts were the Astros,
Rangers, and Angels. They finished 14th, seventh, and sixth, respectively, in
runs scored. Sensing a theme here?
Bunting sucks.
0 comments :
Post a Comment