Plenty of words and pixels have been devoted to Major League Baseball’s new instant replay system. Whether you think of it as the game’s savior or as merely another stepping stone to a dull, anti-septic, technocratic future, it is here to stay, and debating its worthiness is pretty played out already one week into the season.
Instead, let’s talk about an overlooked and perhaps even unexpected aspect of the replay system: the strategic knowledge of managers will be tested in a big way.
The first major whiff came on Opening Day in the game between Arizona and San Francisco. In the fourth inning, the Diamondbacks had one out, center fielder A.J. Pollock on first base, and pitcher Wade Miley batting. Matt Cain threw a pickoff attempt and had Pollock leaning towards second, but the center fielder appeared to just beat the tag and return to the base safely.
After 30 seconds of watching the replay — the same 30 seconds the Giants dugout gets to pass a signal to Bochy — I couldn’t tell if Pollock was safe or out. Apparently, the Giants thought they had something, and Bochy challenged the call. This was a poor decision for a number of reasons.
At the time, San Francisco led 4-2. Their win probability sat at a neat 70.3 percent according to FanGraphs. More importantly, this was not exactly a high-leverage situation. FanGraphs spits out a leverage index of 1.52 for the play — more significant than the average situation, but not significant for a close game. Worse, a successful challenge would have only raised San Francisco’s win probability to 74.1 percent, according to Baseball Prospectus’s most recent win probability tables. The “challenge leverage index,” if we wanted to go to such lengths as to calculate it, would have been much lower than 1.52 in this case.
There is a clear and obvious point to only giving managers one failed challenge for the game. There are going to be close plays like this pickoff call in every game, and it would be ridiculous if every single one went to replay. By forcing the managers to hold the card for a play when it either matters or is absurdly clear, this should limit the number of frivolous challenges. Bochy would have won the right to another challenge had he succeeded, but the combination of the uncertainty of the play and the minimal impact of the call makes it clear that this was precisely the kind of frivolous challenge the system should weed out.
After the failed challenge, Miley popped up a bunt (another reason why not to waste the challenge there), but with two outs, Gerardo Parra doubled, putting runners at second and third. With the Diamondbacks trailing by two, the next plate appearance had a leverage index of 2.61, almost double the leverage at the time of the challenge and the second-highest leverage index of all plays in the first six innings of the game.
With a 1-1 count to Aaron Hill, Buster Posey allowed a passed ball. There was a play at the plate, and it appeared Cain had successfully tagged out Pollock attempting to score from third. Pollock tried to step over the tag, but replays showed he clearly grazed Cain’s glove before he touched the plate. A challenge would have easily succeeded, but since it was within the first six innings, the umpires were not allowed to make their own review call. The run counted, and the Giants’ win probability was down to just 59.9%, compared to 77.1% with a correct call and the inning over.
The Diamondbacks stranded Parra at third base, but they managed to take the lead with two runs in the sixth and went on to win the game 5-4. Obviously, Bochy’s wasted challenge is not the only reason the Giants lost. Defensive miscues, quiet bats and bad pitches were just as responsible for the loss as the poor challenge. But Bochy’s job, as with all managers, is nothing if not to put his players in the best position to win, and Bochy clearly failed in this regard.
Leverage index has one clear thing to say about challenges early in the game: if you’re not sure, hold onto it. Bochy challenged a play that mattered more than the average play in a game, but the “average play” includes all those first inning at-bats with nobody on, and all those late-inning at-bats in games decided well before. As the leverage index chart at Inside the Book shows, most plays in a close game in the fifth or sixth inning would have been at a significantly higher leverage index. And if the game had gotten out of hand by the fifth or sixth inning, in this case, it certainly would not have been because Bochy held on to his challenge.
According to Bloomberg Sports, 16 of 44 replay reviews have resulted in an overturned call, a 36 percent rate. Although reversing 16 calls over the course of a week seems worthwhile to me, a replay system that only overturns calls 1 in 3 times will grow tedious quickly.
It’s unclear, as Drew covered last week, if the manager challenge is the ideal way to do things. But one would expect managers will figure out the proper strategy quickly, especially in a world in which so many teams employ people in analytics. Putting together a cheat sheet of when to challenge a borderline call would probably take one of these office-dwellers about half an hour. Managers should be able to adapt to the new system and start holding their challenges for the moments they’re really needed. If they can do so, we should expect to see a higher success rate, a lower overall challenge rate, and hopefully, a replay review system that intrudes less on the game and makes a difference when it does.