Rutgers has been involved in 9 games decided by 3 points or less, most in the country. 4 wins, 5 losses. These things tend to even out. The real good teams win the majority of them, the bad teams lose them all.
I think you were right the first time - these things tend to even out - but not the second. They even out no matter how good or bad you are.
Many years ago (twenty, in fact), I noticed that a coach who the schools’ online fans were calling on to be fired had won a boatload of one-possession games. In the days before I was introduced to analytics, my intuition was there could be no better indicator of coaching prowess. But before I posted, I thought I ought to sanity check the numbers.
Big mistake. I looked at the records of coaches the consensus deemed as great, all of whom had great teams. (The one example I remember was Coach K). Just about all of those great coaches hovered around .500 in those kinds of games (even when the games were against much-lesser opponents). It changed my intuitive sense of how these outcomes occur.
One-possession games are a dice roll and there is probably a normal distribution curve of the outcomes (0-10 on the left tail, 10-0 on the right, a big hump around 5-5) that is not obviously correlated to “good teams win them” and “bad teams lose them” except with something like hindsight or outcome bias.
I believe the accurate stock answer about this is not that ‘better teams win the close ones’ but that ‘better teams don’t get in the close ones’ - they get ahead and they put teams away. Mostly through talent and, for lack of a better term, a killer instinct, occasionally with a hand from in-game coaching decisions.
Fluoxetine is the one who has it right here.
Interestingly, KenPom has Rutgers at 0.00 in luck rating; its record matches its underlying efficiency. Until recently (when things began to even out), KenPom had Rutgers conference SOS at 14th. (It’s now at 11, even though the most recent opponents were on the bottom end, and that upward trend is about to be ‘with a bullet’, as Kasey Kasem might say.)
Torvik has a stat that is similar to the baseball WAR stat, PRPG!. (Basically points over average D1 player at their usage rate.)
If you average out what Rutgers has on the floor (%min played * PRPG!), Rutgers averages a team of guys who score 0.52 points more than a team of perfectly average D1 guys. Among peers, for example, Penn State‘s average is 0.93*, Northwestern 0.8. Purdue’s at 1.08.
Exceptional performances from RHJ (and streaky Geo) can lift RU in any given game, Cliff and Mulcahy are solid contributors, but there isn’t a lot of firepower among the rest of the minutes. And about 40% of the team’s minutes are going to guys other than those four.
That can change as guys get more experience and a team figures itself out. (Eyeballs-wise, Reiber seems to be developing rapidly.) Unfortunately, the remaining RU schedule is going to make that tough.
If you machine-language analyzed the posts here, none of this is surprising other than maybe whether RU has under-performed (it probably hasn’t), and the one-possession game stat is interesting but not meaningful.
I watch Rutgers a lot because I enjoy how Pikiell’s teams compete when they are playing the way he wants them to play. It certainly seems hard to sustain, though. Everyone in this league seems to play hard.
_____
* Penn State has played the toughest league schedule so far. Michigan the easiest.
