The History of Starters Relieving on Short Rest

A few hours from now, Yu Darvish will throw the first pitch of Game 7 of this year’s World Series. Shortly thereafter — or, less shortly thereafter, if the Astros are lucky — Lance McCullers will take the mound. I don’t need to tell you what Game 7 means. Nobody does. It’s plainly evident: This game is everything. It’s everything that anyone plays for.

Because of the stakes, and because there’s no tomorrow, patterns you might be used to no longer apply. Both teams will employ an all-hands-on-deck approach, hoping for sufficient adrenaline to counteract fatigue. Darvish, of course, will want to go as long as he can. The same goes for McCullers. They’ve probably both dreamed of going the distance. But that’s almost certainly not going to happen. The Dodgers and Astros are likely to dip into their bullpens. And that’s where it gets extra fun.

Both teams have their full complement of arms. The Dodgers might have more faith in their relievers than the Astros do, but the Astros’ relievers also ought to be more rested. Yet there’s an additional twist. It’s hard to find a writer who doesn’t expect to see Clayton Kershaw. It’s just as hard to find a writer who doesn’t expect to see Dallas Keuchel. There’s also been chatter the Astros might make brief use of Justin Verlander. Kershaw and Keuchel would be on two days’ rest. Verlander, zero. With one game remaining, one game that means more than all others, we should examine the playoff history of this.

No such appearance would be unprecedented. Keuchel has even attempted this before. When many people think of starters making short-rest relief appearances in the playoffs, their minds go right to Madison Bumgarner. It was Bumgarner whose incredible five shutout innings slammed the door on the Royals in 2014’s Game 7. Bumgarner was pitching on two days of rest. A.J. Hinch and Dave Roberts would love nothing more than to have that kind of option. Bumgarner kind of made everything easy.

But there are individual appearances, and then there’s the whole group of them. I made liberal use of the Baseball Reference Play Index. I examined the wild-card era, stretching from 1995 – 2017, and I looked for appearances made on zero, one, or two days of rest. I then manually selected only those appearances made by pitchers who had recently started. I decided to eliminate Tim Wakefield from the sample, because he was just weird. He played under different rules. In any case, I found 47 appearances, spanning a combined 66 innings. It’s not a huge sample, but it’s something, and honestly more than I expected. Over those 66 innings, the short-rest starters in relief allowed 29 runs.

That doesn’t mean much of anything by itself. To set some context, I calculated the weighted averages for how all those pitchers combined to do in the playoffs. I also calculated the weighted averages for how all those pitchers combined to do in the regular season. Here is a small table!

The pitchers, as a group, were a little worse in the playoffs than they were in the regular season. I think this is to be expected, because the average playoff hitter is better than the average guy you face in the first six months. The quality of competition rises steeply, so you see an increase in RA/9 and an increase in walk rate, with a small decrease in strikeout rate.

The most important row is the top one. That reflects the combined performance in the short-rest relief appearances. RA/9 goes back down. There’s another small bump in the walks, but there’s a significant leap in the strikeouts. That’s good! That’s encouraging.

It might be obvious, or it might not be: I think we’d expect that. These are starting pitchers. In the short-rest appearances, they’ve all served as relievers, save for one exceptionally short-rest start by Derek Lowe in 2004. One of the fundamental truths that’s come out of baseball analysis is that relieving is easier than starting is. You get to throw at 100% the whole time, instead of worrying about saving your bullets. You also get to avoid your worst pitch if you want, since you’re typically not seeing batters more than once. You’d think that there would be a performance boost. The short-rest appearances have led to deeper counts, and more whiffs. Runs have still scored — plenty of runs have still scored — but the history would suggest that Keuchel and Kershaw could be perfectly useful tonight. Asking for several innings would probably be too much, but I imagine they could handle one or two, with an outside shot at three.

As far as Verlander goes, I don’t know how realistic that might be. He’s had no rest at all. He just threw 93 pitches. Only one of the appearances in the data sample occurred on zero days’ rest — Randy Johnson pitching in Game 7 of the 2001 World Series, a day after he started. Johnson retired all four batters he faced. People have tried to compare Verlander’s durability to Johnson’s. Verlander might indeed be one of the freaks. It’s something to watch for, should the Astros need a couple crucial outs.

The easiest way to conclude would be by saying it’s a mixed bag. It’s always a mixed bag, of good results, of bad results, and of results somewhere in between. Kershaw has made one of these appearances, in last year’s NLDS, and he retired Daniel Murphy and Wilmer Difo. Keuchel has made one of these appearance, in the 2015 ALDS, and he wasn’t good. He allowed three runs in his inning, with a double and a homer. I have no idea what that’s supposed to mean for today. As long as any pitchers warming up in the bullpen have their velocity, I don’t know how you could tell anything else. You’d have to let the hitters show you whether a pitcher should be pitching in the most important game of the year.

Long story short: Starters have relieved on short rest in the playoffs before. Everyone remembers Madison Bumgarner. Most everyone remembers Randy Johnson. Fewer people remember Dallas Keuchel. Fewer people remember Jack McDowell. There are definitely perks to being a major-league manager. This is one of those days that I’m ever so glad I don’t have their job.