ALCS Rainout a Mixed Bag for Pitchers

The Yankees and Astros both won over 100 games in the regular season but nobody beats Mother Nature. When a rainstorm causes cool terms like “bomb cyclone” or “explosive cyclogenesis” to be bandied about, you know you’re not expecting a light drizzle. Yankee Stadium is currently dry, but with a system the size of the mid-Atlantic barreling up the coast, it didn’t make sense for MLB to pretend that tonight’s ALCS Game 4 was going to take place. Just look at the radar, courtesy of the National Weather Service:

Yikes. The rain provides the Yankees and Astros with an extra off-day now at the expense of losing an off-day between a possible Game 5 and Game 6. This isn’t a big deal for the hitters, but it will result in some revised pitching plans. In a five-game divisional series, teams can generally muddle through with a three-man rotation. Due to the 2-2-1 format, no team plays on three consecutive days, and while the Game 1 starter would have to pitch Game 4 on short rest, the Game 2 starter can pitch in a possible rubber match on normal rest. This extra rest gives teams the flexibility to either stretch their best three starters or, as the Nationals demonstrated, use starting pitchers in relief more aggressively.

But short of baseball going to some kind of impractical 2-2-1-1-1 format, that doesn’t quite work in a seven-game series. So unless you’re going to have your entire rotation do it 1930-style, you’ll need to use a fourth starter. That isn’t an ideal situation for either the Yankees or the Astros. From a pure projection standpoint, it’s actually doesn’t move the probabilities. The Astros get an immediate benefit in that they avoid a Bullpen vs. Bullpen Game 4; ZiPS takes bullpen depth into consideration and Yankees enjoy a significant projected edge in any such bullpen game. Before the rainout, ZiPS projected the Yankees to have a 56%-44% edge in a home bullpen duel, so it’s a nice game to delay if you’re Houston.

The problem you run into with this model is that the rainout doesn’t really add an extra day of rest, it simply moves it. Since there are only two days of rest for a Game 4 starter to pitch in Game 7 now, both teams end up repeating the dilemma of either using a fourth starter — particularly problematic for the Astros — or going with a bullpen game.

To get a sense of how the probabilities shake out rainout vs. no rainout, I’ve broken down the ZiPS series win probability for the 10 remaining win “patterns” still possible in the final four potential games of the ALCS:

ZiPS ALCS Projections – No Rainout
Game 4 Game 5 Game 6 Game 7 Probability
Astros Astros Unnecessary Unnecessary 22.1%
Astros Yankees Astros Unnecessary 11.7%
Yankees Astros Astros Unnecessary 15.3%
Yankees Yankees Astros Astros 8.3%
Astros Yankees Yankees Astros 5.5%
Yankees Astros Yankees Astros 7.3%
Total Astros Odds 70.1%
Game 4 Game 5 Game 6 Game 7 Probability
Yankees Yankees Yankees Unnecessary 12.9%
Yankees Yankees Astros Yankees 6.7%
Astros Yankees Yankees Yankees 4.5%
Yankees Astros Yankees Yankees 5.9%
Total Yankees Odds 29.9%

ZiPS ALCS Projections – Rainout
Game 4 Game 5 Game 6 Game 7 Probability
Astros Astros Unnecessary Unnecessary 24.7%
Astros Yankees Astros Unnecessary 14.3%
Yankees Astros Astros Unnecessary 13.4%
Yankees Yankees Astros Astros 6.9%
Astros Yankees Yankees Astros 5.6%
Yankees Astros Yankees Astros 5.3%
Total Astros Odds 70.1%
Game 4 Game 5 Game 6 Game 7 Probability
Yankees Yankees Yankees Unnecessary 11.3%
Yankees Yankees Astros Yankees 7.2%
Astros Yankees Yankees Yankees 5.9%
Yankees Astros Yankees Yankees 5.5%
Total Yankees Odds 29.9%

The overall odds don’t move, but the likelihood of each pattern does. There’s an old saying in statistics, usually attributed to English statistician George Box: All models are wrong, but some are useful. I like to believe I’m fairly clever, but ZiPS is just a model, not an all-encompassing theory of everything in the baseball universe. Box encourages statisticians to focus on the big things, what’s importantly wrong, because “it is inappropriate to be concerned about mice when there are tigers abroad.”

ZiPS is, I believe, a good generalized model, but there are small bits that are hard to capture. For example, what if the Astros bullpen, with the fourth-highest hard-hit percentage against in baseball in 2019, is particularly poorly suited to Yankee Stadium and gets a larger-than-average boost from not pitching a full game in New York? On the flip side, the Yankees bullpen has thrown 15.1 ALCS innings compared to Houston’s 8.1; what benefit do the Bombers get with an extra day of rest right now? And which effect is stronger? How much does the immediate day of rest increase the probability that Giancarlo Stanton is in the lineup and playing close to normal strength? These are all very real questions and while not all of the answers will be significant factors, some of them will be, and you can probably think up hundreds more of these little what-ifs with enough pondering time. I’m not afraid to say “I don’t know” to some of these; knowing what your tools can’t do is as important as knowing what they can.

There’s one last small advantage I think the Astros do get, not so much in the ALCS, but in a possible Nationals-Astros World Series matchup. While the Nationals heading to the Fall Classic well-rested and able to set up their rotation without any rest-day considerations, the Astros aren’t necessarily assured of that luxury. The probabilities of the Astros winning the ALCS are unchanged, but the probability that they win in fewer games increases with the rainout, with a win in five games the scenario with the biggest boost. The Nats aren’t a creampuff opponent, beating the 106-win Dodgers in a series that ZiPS projected as nearly a coin flip despite the regular season difference in win totals. Washington’s built to win in the postseason and the ALCS winner ought to want to be as close to 100% as possible when the World Series starts on Tuesday.

In the end, who wins the American League championship will come down to what the Astros and Yankees do, not the rain. Which, really, is how it should be.

Dan Szymborski is a senior writer for FanGraphs and the developer of the ZiPS projection system. He was a writer for from 2010-2018, a regular guest on a number of radio shows and podcasts, and a voting BBWAA member. He also maintains a terrible Twitter account at @DSzymborski.

Newest Most Voted
Inline Feedbacks
View all comments
4 years ago

I think what it does-
No rainout- there was only a 22% chance that Houston wins in 5. So not having to use Verlander and able to set up their WS rotation Verlander/Cole 1-2
with rainout- there’s now a 52.4% chance that Houston wins in 5 or 6 games. Verlander goes on Friday now so he’d be set to go game 2, with Cole in game 1.

In terms of the series though going on, there’s not that much difference. Only 2.6% more of a chance that the series ends in 5 games.