We Were Always Coming Back
A brain shaped by win expectancy.
The professional poker player Sam Greenwood writes a newsletter called Punt Of The Day about his own mistakes, which is something I enjoy and admire. Sam is also a big ball fan. You can tell he’s my kind of ball fan by this question he sent me:
Do you know if anyone has written about what the peak win expectancy is for the average losing team in a given MLB game? … What I am aiming to find is that most sports fans are miserable because most of the time when their team loses they were a reasonably big favourite at some point during the game
Nobody has written about it that I know of, or that Ben Lindbergh knew of, and Ben is the living database of everything ever written. So I asked Dan Hirsch, a developer at Baseball Reference, if he could run the query and he did. Here are the results, in two parts:
About 85 percent of losing teams were “winning” at some point, according to win expectancy.
This makes sense if you think about it, or even if you don’t. In standard win expectancy methodology every game starts at 50/501 and updates its state after each batter2. If the first batter of the game makes an out, the visiting team becomes a win expectancy underdog, which means that the only way for a visiting team to win without ever being “behind” is to
a) get the leadoff man on base, or else immediately drop to 48 percent win expectancy;
b) score in the top of the first, or else drop to 45 percent win expectancy; and
c) move the runner around the bases fast enough not to lose the win expectancy edge. Runners on first and third with two outs before scoring? You’ve technically become the underdog at that point, with 49 percent win expectancy.
For a home team to go wire to wire they must get the first two batters of the game out, since a runner on first with one out gives the visiting team a 50.16 percent win expectancy. And so on.
Is this what Sam was looking for? Probably not. Nobody blows a 50.16-49.84 lead and goes off to write The Myth Of Sisyphus. So where is the bummer line? I’m going to give you three situations, and you tell me which ones make you the most miserable.
Situation 1. Your team is on the road. It’s a cold April day and the opposing starter has a hard time getting loose. The first two batters single; the third batter works a walk. The bases are loaded in the first inning with nobody out, and you’ve got your cleanup hitter coming up!
You go on to lose.
Situation 2. Your team is on the road. Both teams score a couple runs early, but then a stubborn tie takes hold. It stays 2-2 until the seventh inning, when a two-out triple and a wild pitch give you a run! You hand a 3-2 lead to your high-leverage relievers with nine outs to go!
You go on to lose.
Situation 3. Your team is at home. The game is tied going to the bottom of the ninth. With two outs, you mount a rally: Single, single, walk. The bases are loaded and the pitcher looks gassed!
You go on to lose.
These are very different situations. I expect one’s emotional response to them would be:
Situation 1: You’d be really optimistic at the high point but not excessively sulky if you eventually lose.
Situation 2: You’d be pretty optimistic at the high point and very sulky if you eventually lose.
Situation 3: You’d be only cautiously optimistic at the high point, but devastated by the loss.
The twist is that these three situations all carry the same win expectancy: 66 percent for the batting team. The other twist (kind of) is that
the average losing team in a major league game blows a 64.3 percent peak win expectancy
which makes these three situations almost average. This is more or less the typical amount of heartbreak for each day’s 15 losing teams.
