Should Kershaw start tonight’s game on 3 days’ rest?

Posted: October 7, 2013 in Managers, Post-season, Uncategorized

The big unknown is how it will affect his performance tonight and in the future. Since starters don’t generally pitch on 3 days’ rest, even the great ones, and even the great ones with bodies that can presumably handle a big workload, we have to assume that it will hurt their performance and/or their chance of future injury.

The Dodgers are up 2 games to 1, so Kershaw is guaranteed to pitch tomorrow unless they win and don’t need him. Conventional wisdom says that you only pitch your ace on 3 days’ rest when his team is facing elimination.

Is that right in this situation? Once we estimate Kershaw’s change in talent, we can pretty much figure out which alternative is correct.

According to The Book, modern pitchers that have pitched on 4 days rest pitched 17 points in wOBA against worse than usual. That is the equivalent of .56 runs in RA per 9. That is a lot. In deference to Mattingly and Kershaw (keep in mind that the pitchers who were studied were also deemed well-suited to pitch on 4 days’ rest, or at least not ill-suited), we’ll make that an even .5 runs. For 6 innings (we will also assume that he will be on somewhat of a short leash after throwing 124 pitches in his last outing), that is a difference of .33 runs plus another .11 runs for that extra inning that Kershaw does not pitch, for a total of .44 runs, or a win percentage of .0044, or 4.4%. So the

Dodgers are 4.4% worse off in game 4 with him pitching then in game 5.

Now let’s do the math. In today’s game, Vegas has the Dodgers as a 67.7% favorite. Presumably if he pitched on 4 days rest, they would be be a 72.1% fave, or a 64.1 % favorite in Atlanta). With Nolasco on the mound, the Dodgers were a 60% favorite in LA, which makes them a 52% favorite in Atlanta.

Chances of Dodgers winning:

1) With Kershaw tonight: .677 + .323*.52 = 84.50%
2) With Kershaw in game 5: .6 + .4 * .644 = 85.76%

So, by pitching Kershaw tonight on 3 days’ rest, assuming that he is .5 runs per 9 inning worse than he would be on the normal 4 days’ rest, costs the Dodgers only 1.26% in win expectancy for the series.

What is the break-even point, in terms of how much worse Kershaw has to be for it to be a tossup? About .22 runs per 9 innings.
For what it is worth, on the broadcast commentary, Pedro was asked about Kerhsaw pitching on short rest. He said that as long as he had time to prepare – to adjust his training routine – that is shouldn’t be a problem. He implied that if it was a last minute decision, then it was problematic. We don’t know how long Kershaw has known or suspected this. Hayhurst (as in our friend Dirk) didn’t think it was a good idea. He called it a “panic” by Mattingly. That is probably a bad choice of words as that would clearly apply to the team that was losing the series.

An astute reader on The Book blog pointed out something which I failed to consider. With Kershaw pitching tonight, Greinke can go tomorrow rather than Nolasco, and Greinke is the better pitcher. (Also, the entire rotation for the NLCS changes, right?)

With Greinke, rather than Nolasco, I estimate the the Dodgers gain 3.7% in win expectancy for that one game, which is a lot.

So now, we have:

1) With Kershaw tonight and Greinke in game 5: .677 + .323*.557 = 85.69%
2) With Kershaw in game 5 and Nolasco tonight: .6 + .4 * .644 = 85.76%

So now it is virtually a tossup! Of course if you use Greinke tomorrow, you have to start the NLCS with Nolasco (then Kershaw and Geinke), etc. If you use Nolasco tonight and Kershaw tomorrow, not only would you get to start the NLCS with Greinke if there is a game 5 in NLDS, but if Nolasco were to win tonight, you can start the NLCS with Kershaw. I am too lazy to see how the different pitching scenarios for the NLCS would pan out and, more importantly, affect the Dodgers’ chances of winning that.

  1. Steve Staude. says:

    Just wondering whether the 17-point difference in wOBA for 3 vs. 4 days of rest accounts for differences in the offenses faced. I imagine when a starter is pitching on 3 days’ rest, it tends to be against above-average offenses.

    • Lorraine says:

      You’re great Donna! I envy your work, i should have chosen Hotpal-ndscaring instead of Hort-Postharvest Physiology, haha. That time while we were in college, landscaping is not in yet, as very few projects are available, mostly corporate and rich clients only do landscaping jobs. Now, i feel like i should have chosen that field too, i am envious. Formal gardens are not for us though, maybe because we are an archipelago and limited lands for those type of gardens.

  2. MGL says:

    I don’t think so, but I am not sure. Why would you think that? We are talking about regular season only.

    • Steve Staude. says:

      Well, OK, I guess an injury to another starter is probably the main reason a starter would pitch on 3 days’ rest during the regular season. But I’d guess the second reason is that it’s an important game against a strong team (especially late in the season). Maybe not. I’ll try to look into it.

  3. MGL says:

    I think that maybe the reason why there was such a penalty in the post-season (besides small sample size noise) is that it may often be a last minute decision. If you believe Pedro Martinez, then preparation and advance notice is the key.

    In the regular season, I’m not sure what the circumstances usually are. It doesn’t happen very often. When the regular pitcher can’t go at the last minute, someone from the pen usually takes his place. The next day’s starter rarely does that. Maybe it usually occurs at the end of the season in big games. I wouldn’t think that would have much to do with the opposing offense though.

    • Steve Staude. says:

      Yup, you were right, MGL. According to 2013 data anyway. There were no differences in the wOBAs of the offenses that were pitched against on 3,4,5, or 6 days of rest. And it definitely wasn’t the case that these were ace pitchers being moved up.

      However, those who pitched on 3 days of rest actually put up slightly better numbers in those games (about 0.004 wOBA) than their overall average.

      The sample size for 3 days was tiny, though — only 8 games in 2013 met the inclusion criteria set in The Book. I think I’ll give more years a shot. It was a real pain to do, though.

  4. TomC says:

    Last night almost had to be a late decision- nobody would be crazy enough to leave Kershaw out for 124 pitches in a 5-run game intending to bring him back on short rest, or even with the thought that it might be necessary. I wouldn’t be surprised if Nolasco winds up off the NLCS roster and we find out about some kind of injury, discomfort on his throw day, etc. That seems far, far more plausible than Mattingly just flipping out and panicking having just won a blowout to go up 2-1 facing Freddy Garcia in G4 and throwing Kershaw in G5.

  5. MGL says:

    I don’t really know. They clearly didn’t plan it when Kershaw started his first game, but we don’t know when they made the decision. My guess is that it was discussed from the getgo, that Kershaw may have prepared for the contingency, but that the final decision was not made until fairly late. I don’t think there is anything wrong with Nolasco. I don”t think that he had anything to do with it other than they don’t think he is all that good.

Leave a Reply to Steve Staude. Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s