Welcome to your weekly Big Ten advanced statistics column.
In this space every week, we’ll check in on on the Big Ten’s KenPom and Sagarin ratings—and maybe a few others—and look at the projected standings, how the Big Ten stacks up nationally, and the differences between the computer numbers and BT Powerhouse’s own Power Rankings.
But before we look at the numbers, there are a few things we need to point out.
1. It’s still early.
There obviously isn’t a whole lot of data yet for the 2017-’18 season, so the bulk of the ratings are based on preseason projections. Preseason projections are obviously going to get a lot of things wrong, and while the computers are getting more sophisticated in being able to project a team’s performance based on the prior year’s results and how that team’s roster has changed, it doesn’t really make much sense yet to say “KenPom says so-and-so is a top 25 team.” That’s the best guess and the degree of confidence associated with that guess is very low until mid-to-late December.
The AP poll is probably as good (or better) than the computer numbers at this point. In fact, there’s actually evidence that the preseason AP poll actually does a better job of projecting NCAA Tournament performance than the poll published immediately before the tournament starts. This is due to cognitive errors on the part of the AP’s human voters, most notably recency bias and the tendency to only look at wins and losses rather than margin of victory and location of the game.
2. But that doesn’t mean the numbers are meaningless.
If the computers say your team is bad, your team is probably bad. Don’t give me any arguments that “the computer results don’t mean anything until January.” They absolutely do mean more and more as we play more and more games. It’s not a switch that gets flipped on a given day.
That said, your team is undervalued because they’re exceeding expectations. If that’s the case, then we should see a steady rise in their computer rankings over time. That’s the type of thing we are going to try to point out in this column.
In other words, early in the season, pay attention to which direction your team’s numbers are moving. If they aren’t moving much, then you should have a pretty good idea about where they stand as the season continues.
3. The RPI is not an advanced metric.
The RPI is not intended to be predictive. It’s an outdated metric that needs to be euthanized (and there’s evidence that the NCAA is going to do just that after the season). Yes, the Selection Committee’s official resume reports include RPI on them. No, it doesn’t matter as much as you think. Either way, this is not a bracketology column, so you won’t see much mention of the RPI around these parts. Especially not in November, because the RPI is ridiculously noisy during the early part of the season.
4. The computer numbers aren’t biased.
The author of this column may be biased as hell. For instance, I’m not nearly as high as most people on Penn State as others because Penn State didn’t schedule like they thought they might be a bubble team. That type of thinking would show up in my rankings if I were an AP voter.
However, the computers just run data through algorithms and produce rankings. You can argue that the algorithms aren’t as well-calibrated as they ought to be, but you can’t argue that Sagarin or KenPom hates your team. Looking at you, Maryland fans.
As for my commentary, feel free to call me out as a moron if I deserve it.
5. We’re going to use KenPom as a source more often than other systems.
That’s not because I think KenPom is necessarily the best system, but because it includes a lot more detail than other systems. In future columns, we’ll dive into offensive and defensive efficiencies and rebounding rates and maybe look at ratings for individual players.
As a way of thanking, Mr. Pomeroy for all the information he provides which this column will be taking ample advantage of, I encourage all of you to sign up for a KenPom subscription. For a college basketball fan, it’s the best $20 you can spend.
The site is incredibly intuitive, and the FanMatch page which displays upcoming games and then the results of those games has a much better layout than the CBS or ESPN scoreboards. I can’t imagine trying to scout an upcoming opponent without pulling up their KenPom page.
6. There may be a lag between the numbers used when I write this column and what the numbers are when you read this column.
If you want the most up-to-date numbers, get a KenPom subscription. I’m hoping you come here for the insightful analysis and commentary around the numbers. That analysis should hold up despite real-time updates bumping a team.
With all those caveats out of the way, let’s take a look at the Big Ten KenPom standings as of 11/12. National ranking in parenthesis.
11/12 Big Ten KenPom Ratings:
- Michigan State (9)
- Purdue (17)
- Northwestern (28)
- Wisconsin (30)
- Michigan (35)
- Minnesota (36)
- Maryland (39)
- Penn State (45)
- Iowa (54)
- Ohio State (74)
- Indiana (79)
- Illinois (92)
- Nebraska (102)
- Rutgers (124)
11/12 Big Ten Sagarin Ratings:
- Michigan State (6)
- Purdue (11)
- Wisconsin (25)
- Michigan (27)
- Maryland (32)
- Minnesota (37)
- Iowa (45)
- Northwestern (46)
- Ohio State (56)
- Illinois (59)
- Penn State (74)
- Indiana (76)
- Nebraska (90)
- Rutgers (160)
Not much has changed since our preseason comments, so I won’t break these down much this time around. Come back next week as we get some real changes on the numbers with more (and better) matchups.