Presenting an Extra Innings Tactics Checkup

The first few days of baseball have brought us our first taste of this year’s new extra innings rules. Sure, the rules were around in the minor leagues before now. Sure, teams theoretically care about their prospects winning. But for the most part, this is new — high stakes games with untested rules to try out. There have now been five extra inning games. Let’s walk through the decisions in each of them to see whether teams are playing the odds or acting rashly.

Angels at A’s

The game between the Angels and A’s was the first extra innings contest of the year. In the top of the 10th, the Angels played it by the book. With first-ever ghost runner Shohei Ohtani on second, Jared Walsh swung away. Whoops:

What can we say tactically, other than that you shouldn’t do that as a runner? Not much. Matt Olson made an excellent read, Matt Chapman made an excellent scoop, and it’s probably a bad break for the Angels that their first automatic runner was the player who had the most on his plate in summer camp, between rehabbing from Tommy John and the usual rigors of two-way work.

After that goof-up, the A’s escaped the top of the inning unscathed. Per the numbers, it was time for a bunt. Ramón Laureano stepped to the plate…and swung away. He got to a 2-2 count before being hit by a pitch. Time to re-up, first and second with no one out…and Matt Chapman, the next batter, also swung away. Chapman struck out. Two batters, neither of whom bunted. It didn’t end up mattering — Matt Olson hit a grand slam to end the game — but let’s look at why the A’s didn’t bunt.

Recall, if you will, that bunting in extras when tied adds 1.3% to your win probability. Let’s look for reasons that 1.3% might be worth less to the A’s. First, what if Laureano were a bad bunter? Our bunt rates are based on the average success rate of a bunt with a man on second, while the A’s know which of their specific players can bunt. Start turning successful bunts into unsuccessful ones, and the margin dries up.

It doesn’t dry up that quickly, though. There’s a lot of margin in a 1.3% win probability edge. If Laureano gets a successful bunt down only 61% of the time rather than the 71% league average, bunting and swinging away come out roughly even. That’s a pretty big cushion. Laureano has all of three successful sacrifice bunts in his professional career, however, and only one since 2015. Is it possible that in an abbreviated run-up to the season, he didn’t have a chance to get comfortable enough bunting to use it here? Surely.

Let’s say, for the sake of argument, that Laureano didn’t bunt because he didn’t feel comfortable. What about Chapman? He, too, isn’t exactly a great sacrifice bunter, with a grand total of zero in his professional career. In his situation, moreover, there was an additional headwind; the identity of the next batter. The worst thing you can do with a runner on third and less than two outs is strike out, and Khris Davis, the next batter up, is no stranger to strikeouts.

To get more specific, here’s what Davis’s projected results look like:

After running the numbers, that means that Davis is a poor batter to bring to the plate with a runner on third and less than two out. If his projected numbers are right, the A’s score 60% of the time with Davis up, as compared to 65% for the average batter. That drags the run value of bunting down by quite a bit; an average bunter with Khris Davis batting after them should be indifferent between bunting and swinging away. The strikeouts and low batting average are simply too much of a penalty.

In fact, maybe this weighed on Laureano’s decision as well. Walk Chapman, and the Angels could always face Davis with one out. Maybe the A’s personnel simply didn’t allow them to bunt in that situation. It’s entirely possible that the A’s simply won’t bunt in this situation, but given their specific circumstances, swinging away seems fine even if they plan on bunting somewhat often.

Braves at Mets

The Braves knew the drill when they hit extra innings; swing away and go for a big number. They started the inning off with two singles and capped it with a double; three runs in an inning that would have scored two runs even without the free runner. Nothing to see here; optimal strategy means playing like it’s a tie game, which is exactly what happened.

On the Mets’ side, there’s not much to say. Down three runs, any kind of tomfoolery is out of the question. They played it straight, and for a moment it looked like it might work — Jake Marisnick and Pete Alonso started the inning with singles. Then gravity took hold, and the Braves held on for a 5-3 victory. But the fancy second-base runner played no part in this game; it was simply an old-fashioned rally.

Royals at Indians

Ah, finally some small ball. The Royals bunted pinch runner Brett Phillips from second to third before Maikel Franco drove him in with a sacrifice fly. Wait, they were the visiting team?? Oh boy. Does this one make any sense?

Not really. You can always bunt here — the rules of baseball don’t prevent you from doing so — but it’s not even a great decision in the absence of the new automatic runner. Scoring from third isn’t automatic, and the other team can score in the bottom half of the inning anyway. Giving away outs as the visiting team rarely makes sense.

In the end, it worked out. The Royals scored one run, and the Indians played the percentages by playing it straight in the bottom half of the inning. Greg Holland came in to wake up the echoes of 2015, and did so in spectacular fashion; a hit batsman and then three straight strikeouts.

Why bunt in the top of the inning? I spent a while trying to come up with a numbers-based explanation for it, but at the end of the day, I needn’t have bothered. You can’t make it work; trading an out for a base just isn’t worth it when the opposition gets a baserunner of its own.

Was it the fact that Erick Mejia was the batter? That certainly doesn’t hurt — Mejia is a weak hitter (64 projected wRC+) with plenty of experience bunting. There’s only problem with that theory. Mike Matheny brought him in as a pinch hitter! Yes, Ryan McBroom, a league average bat, was due up, and Matheny went with the bunt specialist instead. I don’t — I can’t — look, let’s just say that this probably shouldn’t be standard practice, and leave it at that.

Blue Jays at Rays

The Jays backed into a workmanlike inning. After Brandon Drury popped out, pinch runner Santiago Espinal stole third (after a replay review) and came home on a sacrifice fly. One neat side effect of this temporary rule is more baserunning decisions, and I don’t know any baseball fans who don’t enjoy a nice steal of third base.

As for the Rays, the rule is simple. Down by a run? Play it out. Play it out they did — after José Martínez drew a walk on the ninth pitch of his at-bat, Kevin Kiermaier sent everyone home with a bases-clearing triple. Stolen bases, triples — this game had it all. What it didn’t have was any controversial decisions, so it merits only a short footnote here.

Brewers at Pirates

Pity the Brewers, who ended up with Omar Narváez on second base and no spare catchers on the bench to start the 10th. He managed to reach third base with one out after a wild pitch, but was stranded there. The Pirates were sitting pretty — Jarrod Dyson started the inning at the plate. While you’d probably prefer to have him standing on second, he’s also an excellent bunter, and he attempted to bunt on three straight pitches.

Yeah, uh, he knocked two foul and struck out looking on an awkward attempt to pull the bat back on the third pitch. Turns out that bunt isn’t automatic, even if it’s quite likely. Strikeouts on pulled back bunts are more or less impossible to track, but adjust the breakeven down a little further. Particularly after an abbreviated summer ramp up that has given batters less time to get used to pitching, teams will want to consider their batters’ comfort level with bunting. That doesn’t mean the Pirates shouldn’t have bunted with Dyson — it was a layup, a good decision that didn’t work out — but it’s worth keeping in mind for future extra inning games.

This game was wild, by the way. The Brewers scored four runs in the ninth to tie the Pirates. After the aforementioned dry 10th inning, they scraped a run across in the 11th before David Phelps managed two clutch strikeouts with a runner on third and one out to close out the game. The new format certainly creates drama.

What did we learn from all of this? Sadly, it’s too early to draw any strong conclusions. There was only one circumstance where a bunt made sense, and even then, the exact context of the game argued against a bunt. High-strikeout batters are an interesting wrinkle to consider; a sufficiently high combination of strikeout pitcher and strikeout-prone batter can make a runner on third less interesting.

More than that, however, we learned that baseball is going to feel essentially the same even with an automatic runner on second base. Walk-off grand slams? Bases-clearing triples? Those aren’t some weird side effect of an extra runner. They’re just good, clean fun.

At the same time, teams don’t always make optimal decisions. There’s no way to rationalize the Royals bringing in a bunting specialist in the top of the 10th. But c’mon; did you think a new rule was going to make Mike Matheny suddenly stick exactly to the numbers and manage like an automaton? In a way, it’s satisfying. The glaring imperfections are a reminder that expecting all of baseball to follow some grand plan with the precision of a watchmaker never made any sense. Baseball is messy, and a lot of the fun of it is in the deviations from the ideal. That was absolutely on display this weekend.