Oscar Robertson's Triple-Double Season
Wikipedia reports: "Oscar Robertson is the only player in NBA history to achieve this feat [of averaging a triple-double]. During the 1961–62 season, Robertson averaged 30.8 points, 12.5 rebounds, and 11.4 assists per game."
My son said, "But Dad, he only had 181 triple-doubles in his career. Wouldn't this mean 82 of them came that one year?"
I said, "No, those were his averages. He had games below those numbers in each category."
"OK, but shouldn't that still be about 70?"
"Hmm, let's see: if his distribution on each of these is a Bell curve, I'd expect that maybe in 25 or 30 games he'd miss a double in assists, maybe in 20 or so in rebounds, and in a handful in points. So perhaps..."
"Dad, right here..."
"Wait a second! Hmm, perhaps in 40 games he actually had a triple double."
"Dad, right here on the page it gives the actual number: 41!"
Ah, the power of statistical reasoning!
My son said, "But Dad, he only had 181 triple-doubles in his career. Wouldn't this mean 82 of them came that one year?"
I said, "No, those were his averages. He had games below those numbers in each category."
"OK, but shouldn't that still be about 70?"
"Hmm, let's see: if his distribution on each of these is a Bell curve, I'd expect that maybe in 25 or 30 games he'd miss a double in assists, maybe in 20 or so in rebounds, and in a handful in points. So perhaps..."
"Dad, right here..."
"Wait a second! Hmm, perhaps in 40 games he actually had a triple double."
"Dad, right here on the page it gives the actual number: 41!"
Ah, the power of statistical reasoning!
When I have this kind of conversation with Clark, if we have time I'll say, "OK let's not look yet, let's see if we can figure out the answer."
ReplyDeleteI don't have children, but I often do that just to test myself. It's weird because I don't have much of a background/training in statistics or math, and I often cannot explain or describe such concepts (though, if applying some effort, I can memorize and recite the concepts), but for some reason my "guestimations" are eerily close to the true answer.
DeleteI will admit, that when I do this around other people and I'm on the money, that I revel in it. It's funny, because they'll think that I'm really smart even though I'm not. Eh, I'll take it.
The takeaway from my prior comment is that I am basically a good guesser (since I cannot explain the mechanism, I *must* be guessing).
DeleteFor instance, if you give me a written test where there is an exact answer, I will do horribly (the exceptions being essay and constructed-reponse tests). However, if you give me a multiple choice test, I will almost always score in the 95th percentile or greater. I've never been able to explain this.