Hey Gabe, it would be interesting to compare those numbers with the numbers we have in a 30 team league only. Are you able to look at just the period from 2000-2001 to present? I suspect these streaks will be less common over that time.

I’m sure it’s just a coincidence that the teams with fanbase problems: ATL, Tampa, Phoenix, Columbus, LA kings and the Panthers, are part of the losing-est franchise list.

There’s a 1985 paper by Gilovich, Valone, and Tversky (here) on the “hot hand” in basketball, which showed that, contrary to popular belief, a basketball player who’s made shots recently is no more likely to hit their next one than their long run average would indicate. So why do so many people insist that there really is such a thing as the hot hand when there isn’t, with even pros trying to get the ball to the guy who made his shots recently?

Well they do some extra experiments, and it turns out that humans seem wired to think that streaks will be more rare than they actually are. They show people the result of digital coin flips, which they manipulate to make streakier or less streaky, and ask people to pick out the one that shows true randomness… and the one people pick is pretty reliably one that they forced to alternate heads and tails more than they really do. So when we see REAL streams of random events, we’re surprised at how common long streaks are, and start inventing reasons for them.

This means that hockey teams will regularly come up with longer strings of wins and losses than most of us would expect, that we will call shooters “hot” or “cold” when they are being perfectly consistent in their odds of making each shot, and so on.

Corollary: “Momentum” is mostly also bunk, but it doesn’t stop our perceptual system reading it in.

"Momentum" is mostly also bunk, but it doesn’t stop our perceptual system reading it in.

Thank you.