Does Rest Matter: 1/29 Advanced Stats Check-In

Seven-time Big Ten coach of the year Gene Keady had a saying: “It’s not who you play, it’s when you play ‘em.”

He could have been talking about a number of things. You don’t want to play a team on an emotional high from the surprise return of one of their best players. (See: Penn State at Ohio State). You do want to play a team that just lost the nation’s best shot-blocker. (See: Indiana at Minnesota).

A basketball season is unpredictable in a lot of ways. Even the first part of that saying, “Who you play”, isn’t a sure thing. When this year’s Big Ten schedule was announced, playing Northwestern twice was seen as contributing to making your schedule harder. Playing Ohio State twice made the schedule easier. Turns out that was totally backwards.

But when we’re talking about the difficulty of a schedule, there’s another wrinkle that we have to consider, particularly this year. The Big Ten plays on Mondays and Fridays now. The Big Ten also plays its tournament a week earlier than usual, so there’s less calendar available for conference games. Combine those two things and you get a Big Ten schedule where your team might have to play Monday/Wednesday, or Wednesday/Friday, or Saturday/Monday. Short rest, in other words.

In Big Ten conference games played through Saturday night, there have been fifteen times a team has had to play on only one day’s rest. And there have been at least three fanbases who have been vocally upset about it, probably more.

So, let’s investigate. Does playing on short rest matter?

Let’s invoke another Big Ten coaching legend. Thad Matta was famous for playing a short rotation. When asked if his boys’ legs didn’t get tired, he said (I’m paraphrasing): “These guys are 18-22 years old. They can play hard for 40 minutes, especially with all the TV timeouts.”

By that same logic, these guys are 18-22 years old. They can play two games in three days. If you want to win a National Championship, you’ll need to play on only one day’s rest three times in the NCAA tournament. Hell, Michigan won four games in four days in last year’s Big Ten Tournament, and that was after dealing with the aftermath of a plane crash. So I tend to buy Thad’s argument.

Of course, Thad Matta got fired, and this year Chris Holtmann is using his bench more extensively. Holtmann won his first nine conference games as a Big Ten coach, something that hasn’t been done since the 1920s.

But that’s anecdotal. Let’s look at some data.

Through Saturday night, there have been 67 Big Ten conference games. 67 home teams, 67 road teams. (OK, one game was in Madison Square Garden, but I considered that a home game for Ohio State for purposes of this analysis.) For each of those games, I determined how many days of rest each side had. If you played Thursday/Saturday, that’s one day’s rest (Friday) for Saturday’s game.

We can use the KenPom rankings to calculate expected winning percentages for each game played so far, so that’s what I did, adjusting for home court advantage. Then we can compare expected wins vs. actual wins for each number of days of rest and see if there’s a meaningful difference.

Home Games

Here are the results for home teams:

• One day’s rest: 13 games, 8.11 expected wins, 9 actual wins
• Two days’ rest: 36 games 23.10 expected wins, 24 actual wins
• Three days’ rest: 12 games, 7.82 expected wins, 8 actual wins
• Four days’ rest: 5 games, 3.08 expected wins, 3 actual wins
• Five days’ rest: 1 game, 0.94 expected wins, 1 actual win

Actual results are dead-on from what we expected.

But that’s home teams. Where you sleep in your own bed the night before and only have to travel five minutes to the gym against five or more hours. Surely fatigue would be a bigger factor on the road?

Road Games

Here are the results for road teams:

• One day’s rest: 12 games, 4.13 expected wins, 5 actual wins
• Two days’ rest: 22 games, 9.58 expected wins, 9 actual wins
• Three days’ rest: 24 games, 8.02 expected wins, 6 actual wins
• Four days’ rest: 6 games, 1.74 expected wins, 2 actual wins
• Five days’ rest: 2 games, 0.44 expected wins, 0 actual wins
• Six days’ rest: 1 game, 0.04 expected wins, 0 actual wins

(Before you ask, that last one was Wisconsin in between games against Nebraska and Purdue.)

Once again, those numbers are more or less dead on.

Conclusion

“It’s not who you play, it’s when you play ‘em” does not refer to how many days off a team has had before facing you. Thad was right. These are kids in the prime of their lives; in high school and AAU they played multiple games on the same day; in November they go through daily three-hour practices that are way more grueling than actual games. Lack of rest is not an excuse.

In other words, no, Big Ten fans, the schedule-makers aren’t out to get you. Rest doesn’t matter. Find another excuse.