Good morning and happy Pac-12 is back week! There’s something very satisfying about saying that something is “back”. I get why Texas and Notre Dame fans say it every year. It’s got a “Trust the Process” level of comfort to it, except TTP let’s you lose and be happy, while “Texas is back” guarantees that you will be sad by the end of the season (or by the 3rd game).
Anyway, we’re going to do something a bit different today and talk about a common college football misconception: the big-win hangover.
You've heard it before. A team get's a "big" win over a highly ranked opponent, or maybe an upset win as an underdog. The next week, what do the analysts all say?
"Coming off a big win last week, can they keep that same energy up and stay focused? Expect a bit of a hangover and they lose this one."
Commenters say it so often that it almost sounds like college football folklore. But is there any weight to this "hangover effect"? Do teams actually struggle after a big win, still riding the highs of the previous week and unable to focus on the task at hand.
We looked into the numbers and found a very interesting story. There is no identifiable hangover effect. In fact, there might be a "big win boost" effect, where underdogs are shot full of confidence in themselves and start playing up to their opponents' levels, knowing that they've already hung with the best of them.
Wins Breed Wins
The first thing I want to look at is the difference between winning a game and losing a game. Is there even an inkling of evidence that a team might lose more often after a win?
It turns out winning football games is a good indication of your ability to win more football games.
As you can see, the average win-rate in game #2 is almost always higher, sometimes by a wide margin, after a win in game #1 compared to a loss in game #1. The only decline is a negligible decline in games where you lose game #1 to an easy opponent, and then come back in the second game to perform better against a second easy opponent, but the sample size was so small that this likely amounts to rounding error.
The result that I think most “analysts” are talking about is either a) playing an easy game and being “sleepy” after a big win, or b) playing two tough opponents back to back and not being able to live up to the hype after your first win. In both cases, teams play better after winning the first game than losing it. This is intuitive but not the common narrative. Don’t be fooled by it!
Sidenote: This also debunks the idea of a "bounceback" win after a loss, indicating that somehow by sucking one week, you will be more inspired to suck less the next week. While this may be the attitude of many teams, it doesn't usually work out in practice. As it turns out, your next opponent could care less about your need to bounce back. You play better if you won the previous game.
Let’s look at just games that were won for a moment. Controlling for the quality of your second opponent, does the “toughness” of win #1 affect your outcomes in game #2?
We would say no after looking at this. When your second opponent is a cupcake (defined as ~ 85% or more win probability), your win rate is ~ 95% whether you’re coming off a “big” win against a tough opponent (you had < 15% win probability) or another easy game.
When your second opponent is a tough matchup, you win about 12% of the time whether you’re coming off another tough win or a cupcake win. So strength of previous opponents do not affect next-game results.
Ok, before we fully debunk this myth, let’s look at one more scenario referenced above. You won in an upset win. You’re riding a high. Now you have an easy game you should win. How do you perform in this game after an upset compared to just a normal win between an evenly matched opponent?
Remember, the game #1 result was a win, so you can only go down from here in win rate. So what is the difference in win rate in that 2nd easy matchup vs. your game #1 win?
When teams pull off an upset win, they perform slightly better against their cupcake matchup the next week than if they had just won an average game. However, these were small samples. There is no evidence of any hangover here.
Underdogs Play Better in Back-to-Back Tough Games
Take a look at this for a second. This is just the difference in win rates from one game to the next, controlled for opponent. Teams that play two games in a row where they are favored actually tend to struggle more in the 2nd game.
Meanwhile, underdogs actually step up and play better in game #2 than they do in game #1! This is exciting because it means those underdogs are getting better and maybe learning from game #1 struggles against similarly tough opponents.
I’d look out for any teams that have just played a tough game–win or lose–and see if they’re playing another big opponent next. They might just shock them. Similarly, I’d look for any team that is favored big in back-to-back games and maybe short them in that second game. Remember, this part has nothing to do with whether they win or lose the first game. It’s more indicative of the experience they gain in the previous game. You can see the same thing put similarly below, but you’ll notice that most of this discrepancy is within the margin of error.
That’s it for this week. If you’re looking for Week 10 win probabilities, you can view them here. The list is getting absurdly long now that everyone is back. There are a good number of mismatches between Elo and Vegas this week, so study up!
Enjoy your weekend and thanks for reading!
– Kyle