In a playoffs defined by managers' creative use of pitchers, Game 7 offers one of the boldest examples: Corey Kluber, Cleveland's ace, will take the mound for his second consecutive start after three games' rest. The team has a shallow rotation in the absence of Carlos Carrasco and Danny Salazar, so they're hoping to squeeze one last good start from Kluber.
It's not unprecedented. In 2001, the Arizona Diamondbacks started Curt Schilling in Games 1, 4, and 7, getting 21.1 innings from him, during which he allowed four earned runs while being asked to throw more than Kluber (Schilling twice threw more than 100 pitches; Kluber's thrown just 88 and 81 pitches in his first two starts). If Kluber pitches a good game and wins, it'll go down as one of the great postseason performances of all time. If he loses, he'll be a cautionary tale, though one who pitched an outstanding postseason overall.
Pitchers have thrown at least one start in the postseason on three days' rest with some frequency. At FiveThirtyEight, Rob Arthur found 12 such starts in the World Series since 2000, and the results were not good: "pitching on short rest added about 2.60 to their ERAs. Without complete time off, these starters got much worse in other key indicators: walks per nine innings rose by 0.6 while strikeouts per nine fell by 0.9."
I was curious if that remained the case for all postseason play. The advantage is a bigger sample size, though the disadvantage is that you're getting starts from earlier in the playoffs, when pitchers have a bit less mileage—and there's good evidence that the cumulative effect of pitching a lot is much more important than pitching often in a short span.
So, keep that in mind. But the results of expanding the sample are pretty interesting.
Using Baseball Reference's Play Index, I found 62 instances of pitchers starting on three days' rest in the playoffs since 2000. Their average ERA is, yeah, not good: 6.97. But ERAs can get very high very quick—Kevin Brown gave up five runs in 1.1 innings against Boston in the 2004 ALCS, for an ERA of 33.75. The median ERA is a more respectable 3.64, meaning about half the starts were, at least, fine.
Another way of looking at it is the Game Score statistic. As Scott Lindblom puts it, "the beauty is in its simplicity." And it's meant to provide a simple answer: did a pitcher have a good start or not? The baseline game score is 50. Above that is average to good, below that is average to bad. It's not a perfect stat, but it's handy.
And looking at the Game Scores of those 62 starts yields an interesting result: the average is 49, the median 50.5. So pretty average.
Win Probability Added? The average is -0.02, the median is 0.002. Again, close to average.
If you expand the sample back to 1990, you get 128 instances, but the results are pretty similar. The average ERA is 7.99 (removing one infinite-ERA sample), but the median is 3.60. The average Game Score is 50.8, the median 52. The average Win Probability Added is 0.006; the median is 0.024.
In conclusion? It's both uninteresting and interesting. The average start on three days' rest in the playoffs is… pretty average. Sometimes it works great, sometimes it's a disaster, just like baseball usually is. Here's one more way of looking at it.
Granted, the pitchers starting on three days' rest in the playoffs tend to be pretty good: Mike Mussina has two of the top 30 by Game Score, as do Curt Schilling, Andy Pettitte, John Smoltz, and Dave Stewart. So average, for them, is a bit below average, given that really good pitchers tend to average in the low 60s and high 50s.
A Game Score of 50 from Kluber isn't really what Cleveland wants—in one of Kluber's two starts in the playoffs on three days' rest, he pitched five innings, struck out seven, gave up four hits (including a home run) and walked two, for a Game Score of 56. They lost, 5-1. In his other such start, against the Cubs, he gave up five hits in six innings, with six strikeouts, one walk, and one earned run, for a Game Score of 63. It's about the difference between a pretty good game and a good game.
Is Francona taking a risk? Yes. But it might not be any bigger than starting a lesser pitcher than Kluber, and not a lot bigger than starting any good pitcher in any baseball game.