Stealing, Bunting, and Mike Trout

This will be an article about a (probably) bad bunt. Have you read an article like that before? Yes? Well, too bad, because you're going to read another one. This time it will be the story of the Angels, the story of a bunt, the story of a bunt that came to be while Mike Trout, a god among men, was on first base. Yes, the same Mike Trout that successfully stole 49 of 54 bases last year. But we're getting ahead of ourselves.

Here's how the inning started:

Top of the ninth, no outs, Angels down by one. Arguably the best closer in the game, Aroldis Chapman, on the mound. And arguably the best closer in the game was facing arguably the best position player in the game, the aforementioned king of all that is good and powerful and just in the world, Michael "I love you please marry me" Trout. In case you forgot. this guy:

What a guy.

Oh, and because we're a saber-inclined website where we use numbers and nonsense like that, I suppose I should give you some numbers regarding this situation.

Angels' probability of winning: 15.3%
Angels' run expectancy: 0.46 runs
Angels' chances of scoring at least one run: 29.3%

The first part of the story isn't too important. Mr. Trout, doing what he does best, beat the Cuban Missile, lining a single to the right side. He ends up on first, and looks handsome doing it.

Here's where things get interesting, and by interesting I mean frustrating as hell. But first, more numbers:

Angels' probability of winning: 26.8%
Angels' run expectancy: 0.83 runs
Angels' chances of scoring at least one run: 43.7%

Hey look, the Angels almost expect to score a run now! Of course, these numbers fail to take into account an important factor: one of the fastest players in the game, and one of the best base stealers, is on first base. And when you're down by one run in the 9th inning and one of the best base stealers in the game is on first base, you generally want to steal.

In fact, let's look at the hypothetical numbers if Mike Trout had in fact stolen second base. Remember, these are hypothetical, not what actually happened (because that's what hypothetical means):

Angels' probability of winning: 34.9%
Angels' run expectancy: 1.07 runs
Angels' chances of scoring at least one run: 63.2%

Look at those numbers, then look at the numbers right above them. That's a pretty significant shift, isn't it? Just by stealing a base, Trout increases his team's probability of winning by 8%, and the chances of scoring at least one run (this at least tying the game) by 20%!

On the other hand, Trout doesn't have a 100% chance of stealing successfully every time, obviously. Last season, he stole successfully 90.7% of the time. That's a pretty insane rate, and probably not one he can continue. So let's significantly regress his success rate to, say, 80%. So, if he attempts to steal, he has an 80% chance of being successful, and a 20% chance of being unsuccessful. What would happen if he was unsuccessful?

Angels' probability of winning: 8.1%
Angels' run expectancy: 0.263 runs
Angels' chances of scoring at least one run: 17.3%

Yikes. So if Trout got caught stealing, his team's chances of winning would decrease by 18.7%. Based on these numbers, we can figure out the expected "value" of a stolen base attempt: 0.8*8.1%+0.2*-18.7% = 2.74%.

So, assuming Trout has an 80% chance of stealing successfully, a stolen base attempt has an expected win probability added of 2.74%. However, there's one very important thing to note: the batter at the plate. Batting after Trout was Erick Aybar. Aybar is not a power hitter. He is a singles hitter, but he does not get on base at a great rate for the number of singles that he hits (read: he doesn't walk a lot); last year his line was .290/.324/.416. When singles hitters who don't get on base at a great rate are at the plate, it is a much better idea to steal than when good power hitters are at the plate.

In fact, I talked about this a while back, and even calculated the breakeven stolen base rate while each individual hitter is at the plate, sorted by number of outs. Aybar, with no outs, had a breakeven rate of 68.5%, which, though not at the bottom of the list, was definitely in lower end of qualified hitters.

Long story short, this means that it was an even smarter idea to steal second when Aybar, who doesn't get on base much and when he does hits singles, was at the plate than the average batter.

Instead, Mike Scioscia decided to do this:

The bunt was successful, and probably seemed like a fine idea at the time because Albert Pujols was coming up next. However, consider the win probability numbers after the bunt:

Angels' probability of winning: 21.9%
Angels' run expectancy: 0.64 runs
Angels' chances of scoring at least one run: 40.6%

Compare that to the numbers above for while Trout was at first, and you see that all three dropped, even the chances of scoring just one run. Now these numbers aren't perfect because of the fact that Pujols was coming up after Aybar. However, even if the bunt didn't hurt the Angels, it was certainly a worse idea than having Trout steal second. And, once that happened, it may have actually been a good idea for Aybar to bunt Trout to third, because that would bump their chances of scoring at least one run to 66.2%, higher than if he was on second with no outs.

But, alas, as often happens when teams decide to give up outs, the Angels' plan did not work. Pujols lined out to right, and Hamilton did this, ending the inning and the game"

Here are the final numbers for the end of the game, because I'm just a mean person:

Angels' probability of winning: 0%
Angels' run expectancy: 0 runs
Angels' chances of scoring at least one run: 0%

Let that be a lesson to the managers and armchair managers of the world. If you have one of the best base stealers in the game on first base, and the batter that is up tends to hit singles and not much else, steal second before you think about bunting. Sure, the risk is higher, but the reward - winning the game - is well worth it.

Note: To be clear, I don't think that this was the sole reason, not the primary reason, that the Angels lost the game. I mostly wanted to point out what I believe to be a mistake in strategy. This absolutely was not the main reason that the Angels lost this game.

Actual WPA and RE values taken from FanGraphs. Theoretical ones taken from the Hardball Times and Baseball Prospectus. Chances of winning values taken from Tangotiger.