statistical replicability

[Marc-Gregor Campredon]

2/16/2020 – Michigan 89, Indiana 65 – 16-9, 7-7 Big Ten

Sports fans are prone to wild swings in mood, often with little justification. Everything that just happened will keep on happening, and this goes double when things are bad. Back when I ran the Blogpoll voters tended to overrate their own teams a hair when things were going well, but it was a dead certainty that they'd under-rank them significantly after a loss. Every voter, every time.

Lose painfully for a month and your perspective gets jaundiced. When Michigan played at Nebraska a few weeks ago they were down Zavier Simpson and Isaiah Livers. Then Franz Wagner got in foul trouble. Michigan spent a big chunk of the first half with Adrian Nunez and CJ Baird on the court. Our photoshopper-in-residence Abraham wondered on twitter why he was watching a random MAC game, and I laughed sardonically.

At some point in January I said I wanted to sim to the end of the season and get the Howard croots on campus. This season felt like a snakebit write-off: Livers couldn't stay on the court, the trident was haunted, Michigan would get a million good looks they miss while opponents poured in every variety of garbage known to man.

At the same time I tried to argue that Michigan's January was a massive statistical outlier that could not last, because Michigan was not the second-worst team in the country at all things from behind the three-point line. And lo:

Opponent M 3P% Opp 3P%
Rutgers 47 25
OSU 32 39
MSU 39 26
Northwestern 35 23
Indiana 53 25

Those five games are the five they've played in February.

The situation is now flipped. Michigan's probably ahead of its skis a little. But you've seen the shots. You've seen Michigan tee up open corner threes over and over again as the opponent issues a contested jack from NBA range. Reality is somewhere between 59 and 1. It's a lot closer to 1.

[After THE JUMP: The Mona Lisa of floor slaps]