It’s a good weekend to bring this back. I have some essays about information theory and sports contests and maybe you missed them earlier. Here goes.

**How Interesting Is A Basketball Tournament?**I say: 63.**What We Talk About When We Talk About How Interesting What We’re Talking About Is**and so I say 63*what*.**But How Interesting Is A Real Basketball Tournament?**in which I say: Nah, 63 is too interesting. There’s some games you really don’t have to watch to know how they turned out. They’re the ones where the number 1 seed beats up the number 16. And nearly all the ones where the number 2 seed plays the number 15.**But How Interesting Is A Basketball Score?**If you want to know more than just who won and who lost ? I say: oh, something like 5.4.**Doesn’t The Other Team Count? How Much?**Since you can be pretty sure the two teams in a basketball game won’t tie, what’s that tell you about the two teams’ score? Nothing much, turns out.**A Little More Talk About What We Talk About When We Talk About How Interesting What We Talk About Is**and I finally say what these 63 or 5.4 or 10.8 things*are*. You know the name of them.

And then for a follow-up I started looking into actual scoring results from major sports. This let me estimate the information-theory content of the scores of soccer, (US) football, and baseball scores, to match my estimate of basketball scores’ information content.

**How Interesting Is A Football Score?**US football has an annoyingly complicated set of scoring rules. But it’s also got enough historical data that I can make an estimate: about 8.7.**How Interesting Is A Baseball Score? Some Partial Results**Baseball is even better about keeping statistics than football is, but I couldn’t get at convenient summaries. So I made do with what I could find.**How Interesting Is A Baseball Score? Some Further Results**With the help of someone else’s aggregation of data I made a slightly more precise estimate.**How Interesting Is A Low-Scoring Game?**This uses distributions, estimates of what kinds of results are likely for low-scoring games such as soccer, hockey, or baseball to draw some conclusions, including questioning whether distributions that mathematicians like are actually good fits for sports.

Don’t try to use this to pass your computer science quals. But I hope it gives you something interesting to talk about while sulking over your brackets, and maybe to read about after that.

Advertisements