On inconsistency (or the lack thereof)
Posted by kj on Thursday, February 26th, 2009
Following last night’s narrower-than-expected win, I’ve seen several observers (both here and elsewhere) say that MSU is “inconsistent.” This is a particularly sensitive topic for us Spartan fans, since last year’s team (which scored 36 points in a loss against Iowa and 103 in a win over Indiana) had the very definition of an inconsistent offense.
Last year, TAFKATBTW introduced a statistic he calls the “Winehouse Factor.” This figure is calculated by finding the standard deviation of a team’s game-by-game points-per-possession figures in conference games. The 2008 MSU team had an offensive Winehouse Factor of 0.23–the second highest figure (on either end of the court) among all the major conference teams in the land. (Technical: I think the article was written prior to MSU’s final regular season game.)
So what do the numbers say this season? Here are the current Big Ten offensive Winehouse Factors (through last night’s games):
| Purdue | 0.12 |
| Iowa | 0.12 |
| MSU | 0.12 |
| Minnesota | 0.13 |
| Wisconsin | 0.13 |
| Indiana | 0.14 |
| Michigan | 0.14 |
| Penn St | 0.15 |
| Ohio St | 0.16 |
| Northwestern | 0.16 |
| Illinois | 0.19 |
We currently stand at #3 in terms of offensive consistency, in a statistical tie with Purdue and Iowa. A glance at the Kenpom game plan tells the story: We’ve had just three games below a point per possession on offense–and just one below 0.95 points per possession.
This team is nearly 50% less inconsistent than last year’s team was. That’s a dramatic improvement–particularly when you consider we’ve played much of the conference season without one of our two top scoring options. With fewer scoring options, you’d think the odds of some offensive duds in a few games would go up, not down.
Is there room for improvement for MSU on offense? Absolutely. We could stand to turn the ball over 2-3 fewer times and make 1-2 more 3-pointers per game. But are we “inconsistent”? Not in any sense outside the normal parameters of major conference college basketball.
In fact, here are the offensive Winehosue Factors for the consensus top four teams in all the land (3 of whom also rank in the top 4 nationally in adjusted offensive efficiency):
| UNC | 0.12 |
| Pitt | 0.18 |
| Oklahoma | 0.11 |
| UConn | 0.13 |
As you can see, our number is right in the mix. At the end of the day, college basketball isn’t the NBA, where a team can mechanically impose its superior talent on its opponent night in and night out. There are going to be a few nights when things don’t click in an optimal fashion. Last night was one of those nights–but we still put up 1.08 points per possession.
Finally, while we’re at it, here are the current Big Ten defensive Winehouse Factors:
| Iowa | 0.10 |
| Indiana | 0.11 |
| Northwestern | 0.11 |
| Michigan | 0.12 |
| Purdue | 0.12 |
| Ohio St | 0.14 |
| MSU | 0.14 |
| Minnesota | 0.15 |
| Illinois | 0.16 |
| Wisconsin | 0.16 |
| Penn St | 0.18 |
We’ve been a little less consistent defensively, but it’s been more a case of improvement than of volatility. We allowed 5 of our first 9 conference opponents to score 1.05 or more points per possession. In the last 6 conference games, meanwhile, only one opponent has scored more than 0.95 points per possession.
I understand why Tom Izzo is never going to concede his team is playing with consistency: It’s his job to keep the players motivated. But I think we fans should recognize that this team is doing exactly what a good Tom Izzo team is supposed to do: forcing tough shots, crashing the glass, and getting to the free throw line. And they’re doing it with a relatively high level of regularity. If the team can improve just a tad in the turnover and 3-point shooting departments, the sky’s the limit.
« « The Word from Germany | Thursday Night Links » »
Filed in stats analysis5 responses so far
5 Responses to “On inconsistency (or the lack thereof)”
SpartanDanon 26 Feb 2009 at 6:42 pm 1I wonder if this data, combined with Pomeroy’s consistency rankings, indicates that our offense and defense are highly correlated. Pomeroy’s consistency factor is simply standard deviation of margin of victory, and shows us at 282nd (although ahead of three Big Ten teams: OSU, NW, Iowa); a high inconsistency rating from Pomeroy but a low one using Wonk’s method suggests that our offense and defense are individually consistent but tend to either both be good or both be bad on the same night.
(Then again, conference-only calculations wouldn’t factor in the 35-point beatdown by UNC or the annihilation of Alcorn State, which are obviously going to skew things in Pomeroy’s math.)
kjon 26 Feb 2009 at 10:20 pm 2Hard to say, Dan. I tend to take the view you have to dismiss the nonconference data–both because the competition is so uneven and because we’re a much different/better team now.
DP99on 27 Feb 2009 at 12:26 pm 3These ratings were really surprising to me. I guess we tend to focus on individual performances when watching a game, whereas on the whole it looks like the team is managing ok. I think we expect to see at about Lucas at 16-6 every night, Roe at 7-7, Suton at 12-8, and so on, but that just not the way it happens. Lucas goes for 20 one game and 8 the next, but at the same time Roe goes for 4 in the first game and 16 in the next. In my head I think “just think how awesome it will be when all these cylinders are firing at once!!” but that’s just not how it’s going to happen.
Bubblin | UM Hoops.comon 27 Feb 2009 at 2:25 pm 4[...] On inconsistency (or the lack thereof) KJ looks an consistency issues in the conference [...]
Indiana Game Recap (3/3/09) | Spartans Weblogon 03 Mar 2009 at 10:48 pm 5[...] my inclination is to write this game off as an outlier in what has otherwise been a stellar (and consistent) run through the conference season. The glitches on offense don’t concern me too much; the [...]