Thursday, January 28, 2010
The shorter the game, the more the upsets
This is so obvious as to not even merit discussion. But, there are two points to make:
1. It may be obvious, but even smart people don’t know it.
2. The interesting part is “how much”, which is the whole point, really, to why I like being a saberist
In baseball, the chance of the home team winning to start (an obviously tied) game is 54.2%. By the time it gets to the 9th inning, it’s down to 51.8%. So, we can say that turning a 9-inning game into a 1-inning game means that you are turning a “true talent” .542 team into a “true talent” .518 team. This is a good time to say that when we call a team a “true talent” .542, we mean it in the context of… the context. That they play a certain distribution of teams, and more importantly, the LENGTH of the game.
Gabe produced a chart along these lines, but he did two things better: 1) he did it as a chart rather than a table, and more importantly 2) he did not take an average home team, but a really really good team. And we get this:
We see that when the game starts, the great team has a 68% chance of winning. With about one-ninth of the game to go (meaning 400 seconds to go), these two lopsided teams, still tied, gives the better team a 60% chance of winning. Basically, the +18% is cut down to +10%. That’s a better return than baseball (+4.2% cut down to +1.8%). It would be better if I (or someone) looked at the good v bad teams like Gabe did to make a better comparison. In any case, it’s fairly close. You probably get a few more upsets in baseball than hockey, but not noticeably different. If you do the same in basketball, I have no doubt the chart would be much flatter than Gabe’s or my chart, meaning way less chance of getting upsets. In football? I don’t know.
But, Gabe then shows something superduper cool, and it’s because of the NHL’s system of rewarding ties.
Basically, the really good team, as long as it is tied with the bad team, will expect to get 1.5 points of the 2 points available to it. But, the bad team, the longer it can keep that very same game tied, will go from 0.75 points at the start to 1.5 points by the conclusion of the game! It’s a ridiculous system the NHL has set up.
Gabe: if you want to try something cool, show the same chart, but assume we have the better scoring system: 3/2/1/0, where you get 3 points for a regulation or OT win, and 2 points for a shootout win. (Obviously the change in the scoring system means that his data won’t necessarily apply. But, at least it will show something that is more reasonable.)
I highly encourage reading this thread from a few years ago. It’s one that I quite enjoyed at the time, and still do now.


Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date