Minnesota 77, Indiana 74: back to Earth.

You are correct, no doubt, but I think there is still some shellshock from the last three seasons, particularly the way the last three seasons finished. I don’t think we are going to lose out or even miss the tournament, but as you note, Minny isn’t a bad team. Neither is Penn State, Northwestern, or Nebraska. This is a sobering reminder that for all of this team’s accomplishments, there are no gimmes in this league this year, home or away.

His blog is still up, I think. Search the archives and you’ll probably find several useful articles on Pomeroy, Sagarin, and tempo-free stats, albeit from five years ago.

A Purdue grad asking an IU liberal arts grad to educate another Purdue grad on stats? Heh. I’ll do my best. I certainly can’t claim to understand the fundamentals beneath either what Pomeroy or Sagarin does, but I think the important distinction between what they do and what the RPI does is that they are trying to predict, not reward, whereas the RPI is basically evaluating wins and losses and strength of schedule. A team that is 0-4 with four one-point losses is treated the same by the RPI as if it had lost every game by 20. Pomeroy and Sagarin would be much more kind, and would consider such a team unlucky and probably not too bad.

I agree with Lonestar’s recommendation for Travis or anyone else of the big Ten Wonk’s archives, particularly this post, as an introduction. It’s interesting to me to see just how much influence guys like Pomeroy and Gasaway have influenced the quality of basketball writing. Back in the mid 2000s, it was common to read stuff like, for example, that Northwestern was a good defensive team, and that would be based only on the fact that they were in low scoring games (disregarding their slow pace and low number of possessions). Or that Wisconsin won despite an anemic offensive, when on a per-possession basis their offense was ruthlessly efficient. It’s also really affected the way I think about rebounding. A stat like overall rebound margin is often useless. This one is still pretty persistent. If, say, Michigan State shoots 15-20 from the field and 5-5 from the line in a half, and gets 3 offensive boards, and Illinois shoots 10-20 and 4-8 from the line (assuming all were live-ball misses) and gets 5 offensive boards, then some sportscaster will breathlessly report that “Illinois is losing , but they outrebounded MSU on offense.” Well, as you well know, Matt, first, they aren’t “offensive rebounding” against each other. One team’s OR opportunity is another team’s DR opportunity. And in the above exaggerated example, MSU would have capitalized on 60 percent of its 5 OR opportunities, while Illinois would have capitalized on only 36 percent of its 15 opportunities. It’s one of those things that seems so obvious once you think about the game that way—of course the number of rebounds can’t be judged in a vacuum, it has to be judged based on the number of rebound opportunities—but it’s still an elusive concept. Turnovers are the same way, of course. Sure, 20 turnovers is almost always a bad number, and 6 is almost always a good number, but the pace of the game says much about whether, say, 13 is an acceptable number.

The final thing I love about Pomeroy is how user-friendly his site is. It’s clean, no ads, no fancy graphics or pulldown menus, and practically every stat is clickable and/or sortable. You won’t like this example, Matt, but last year after Purdue’s loss to VCU I mentioned that it was, by far, the worst defensive performance (on a points per possession basis) in the Painter era. Someone (not you or Travis, for that matter) asked how long I spent looking that up. The answer, thanks to Pom’s awesome website, is probably less than 30 seconds. Given how much I use it for the site, the $20 membership to Pom is well worth it for time savings alone. StatSheet has all of these numbers as well, but it is at the opposite extreme of usability. Considering that I write my posts when I can squeeze them into my workday, I probably made back the $20 I spent on Pom within a week, considering how long it would take me to use StatSheet.