Big Ten Basketball Efficiency Margin

thePhilipJFry's picture
January 16, 2014 at 12:21p
17 Comments

I really miss John Gasaway.  ESPN hired him and put all of his stuff behind a paywall and now I never get my Tuesday Truths.  Tuesday Truths was when Gasaway would look at every team’s points per possession offensively and defensively during conference play and rank them by margin.  Since ESPN has taken this away from me, I decided to do it on my own.  I may add to this in the future by including turnovers and offensive rebounding percentage as well as a scheduling balance component.

  pace offense defense margin
Wisconsin 64.5 1.23 1.02 0.210
Michigan 58.2 1.24 1.04 0.199
Michigan State 65.6 1.11 0.93 0.185
Iowa 69.1 1.14 0.96 0.178
Ohio State 70.6 1.08 1.01 0.070
Purdue 64.9 1.09 1.08 0.013
Illinois 67.6 1.00 1.04 -0.036
Minnesota 65.2 1.09 1.14 -0.048
Indiana 67.3 1.08 1.13 -0.049
Penn State 65.2 1.00 1.13 -0.129
Nebraska 64.8 0.94 1.13 -0.185
Northwestern 59.6 0.86 1.14 -0.284

You can see the top 4 have clearly separated themselves from the pack.  I don’t know if Michigan is really there.  Their schedule so far includes only 4 of the bottom 5 teams and their efficiency margin was earned almost entirely in two home court wins over Penn State and Northwestern.  Their upcoming stretch of at Wisconsin, versus Iowa and at Michigan State should tell us where they belong. 

Wisconsin thus far in Big Ten play is attempting 34 free throws a game at home, and 7 free throws per game on the road.  I guess that’s just the Kohl Center difference.  Almost certainly that difference won’t remain all season, but it gets you close to understanding why Fran flipped out.

As for Ohio State who finds themselves all alone yet somehow in the middle, Ken Pomeroy’s numbers have the Ohio State defensive unit as the best in the country.  Through four games in Big Ten play they are above average, but trail MSU and Iowa.  Ohio State has only themselves to blame for this as they allowed 1.2 points per possession to Iowa while handing over the ball 17 times one game after committing 21 turnovers against MSU.  Those turnovers not only make the Ohio State offense bad, they make the opponents defense look good.

After I wrote this, Gasaway tweeted that his Tuesday Truths will be posted on his website starting next week.  He posted a preview where his numbers for the Big Ten are different than mine.  Since points scored are more fact than a subject of debate, this must come down to number of possessions.  I used Dean Oliver’s formula of pos = fga – or + to + 0.4 * fta.  I suspect Gasaway is just counting possessions off the play by play data.

http://johngasaway.com/2014/01/16/coming-soon-tuesday-truths-for-a-realigned-world/

Comments

Jack Fu's picture

+1 to you, sir. Wonk on.

d5k's picture

You really need some kenpom.com.  Opponent adjusted efficiency > plain-old efficiency.  He doesn't isolate conference play though (perhaps if you subscribe).

thePhilipJFry's picture

The subscription portion of kenpom.com gets you to individual players offensive ratings (across a multitude of offensive statistics), predicted results for future games and a more thorough breakdown of team statistics.  It does not give you conference only stats.
There are reasons to look at conference stats only.

  1. The college basketball season can be broken down into three parts.  Non conference (teams play schedules where opponents strength varies wildly), conference (teams within a conference play fairly similar competition) and tournament.  Some teams have shown an ability to cream terrible opponents in the non conference schedule while looking more pedestrian when facing worthy opponents.  Even when doing an opponents adjustment things can slip through.  Wisconsin in 2012 was 7th in kenpom's final rankings.  In big ten play they had a 0.06 efficiency margin.  A good team, but not the 7th best in all of college basketball.
  2. The conference slate starts roughly halfway through the season.  Many coaches and players make adjustments to their play based on the results of the nonconference schedule.  Coaches may shorten their bench, defensive rotations and communication can be improved, etc. 

I guess what I'm trying to say is, look at this as another piece of information that is available.  Kenpom's rankings are great, and I look at them.  But I think there is value here too.

d5k's picture

This has to do with the impact of the scoring margin on games against much lesser competition.  In summary, whether you beat the 250th best team in the country by 40 or 60 isn't really that good of a predictor of future success.  Pomeroy actually admitted this and adjusted his formula this year to dampen margin of victory and he claims it fits past data better than his old model.
http://kenpom.com/blog/index.php/weblog/entry/pomeroy_ratings_version_2.0
He mentions Wisconsin but while it would've hurt them a bit that year it wouldn't have dropped them a ton.  The problem with only looking at conference games is it limits the sample size a bit.  Limiting to games vs. the top 150 of some ranking system might be a good model to rank the tournament caliber teams.

Zimmy07's picture

I'm curious about the formulas.
Is pace equal to possessions per game?
Is offense then points per possessions & defense points surrendered per possession?

thePhilipJFry's picture

You are correct on all three.  If I post these again updated I'll add that, so thank you.

Zimmy07's picture

These are cool stats.  I was very surprised to see OSU leading in possessions.

thePhilipJFry's picture

Gasaway has Iowa at 72.2 and Ohio State at 69.2.  I suspect that he is correct and that the 0.4* free throw attempts in Dean Oliver's formula is underestimating the number of Iowa possessions that end with a free throw and overestimating the number of Ohio State possessions that end with a free throw.  This would mean that Iowa is getting a lot of made baskets plus a foul shot (or missing the front end of a 1 and 1), while Ohio State is going to the line more for two shot fouls (or three).  Either way they are the two fastest paced big ten teams.

Go1Bucks's picture

More stats = always good info.
Thanks for doing the legwork.
 
Go Bucks!

"I will pound you and pound you, until you quit." -Woody

thePhilipJFry's picture

Ohio State just put up 0.84 points per possession against what is (or used to be anyway) the worst defense in the Big Ten.  Looking through the numbers more closely Ohio State's offensive efficiency in Big Ten play is being skewed heavily by the 17 minutes that Amedeo Della Valle played against Nebraska.  Removing those minutes puts the OSU offense at 0.98 points per possession in Big Ten play and looks to be a lot closer to what you should expect out of the Ohio State offense the rest of the year.

Jack Fu's picture

Yep. Last year OSU put up 1.03 PPP in Big Ten play, which is very close to completely average. And that was with Deshaun Thomas. Without him, I wouldn't be surprised if they end up under a point a possession this season. Everything's riding on the defense.

thePhilipJFry's picture

This year Ohio State is forcing turnovers on 20% of opponents possessions, which is 3rd in the B1G.  They are however allowing opponents to grab 34% of available offensive rebounds, which is 8th in the B1G.
Offensively if Ohio State doesn't solve the packed in the lane version of the 2/3 zone that they have seen the past two games, they will finish the season in the 0.92-0.95 points per possession range (with the B1G average at 1.07).  With their 1.01 points per possession allowed that would put them in the 9-15 conference loss range.

Indy_Buck87's picture

Come on you think the buckeyes will lose 9-15 conference games? If statistics were an exact science I would be a billionaire from the stock market.    I will go on record and say that OSU likely finishes with 90% probability the regular season at 25-5. They will lose 2 more games this regular season and that does not include a loss to MSU at home. 

I know of only two things that are infinite, space and human stupidity.....and I'm not sure about space". Albert Einstein.

thePhilipJFry's picture

Last year 19 teams finished the conference schedule with efficiency margins between -0.04 and -0.10.  If Ohio State continues to play against the packed in 2/3 zone in the same manner that they have the past two games, their efficiency margin will fall in that range.  Those 19 teams had a conference winning percentage of 0.36.  OSU is already 2-3, so over the next 13 games we should expect them to win 4.7 of those games (6.7-11.3 for a final conference record).

RedsBuckeyeBoy's picture

On defense, I love the Craft/Scott combo on the floor at the same time.
On offense, I think that same combination is contributing to a lot of the offensive woes since teams can pretty much disregard them as a threat to hit a jump shot.

thePhilipJFry's picture

Yeah in the old Iowa high school girls 6 on 6 games a starting defense of Craft, Scott and Williams would be tough to score against.  Unfortunately in 5 on 5 basketball players have to go both ways.

Jack Fu's picture

Gasaway's updated the Truths taking into account games through last night. OSU is 5th in the Big Ten at +0.02, scoring exactly 1 point per possession and allowing 0.98 PPP.

    w-l pace ppp opp ppp em
1. Michigan State 6-0 64.1 1.11 0.91 +0.20
2. Iowa 4-1 72.0 1.14 0.96 +0.18
3. Michigan 5-0 61.6 1.18 1.03 +0.15
4. Wisconsin 3-2 66.3 1.17 1.03 +0.14
5. Ohio State 2-4 68.2 1.00 0.98 +0.02
6. Purdue 3-2 67.4 1.04 1.03 +0.01
7. Minnesota 3-3 65.6 1.05 1.10 -0.05
8. Indiana 2-3 66.2 0.99 1.06 -0.07
9. Illinois 2-4 66.1 0.97 1.04 -0.07
10. Nebraska 1-4 66.9 0.93 1.06 -0.13
11. Penn State 0-6 67.6 0.96 1.10 -0.14
12. Northwestern 2-4 61.3 0.84 1.05 -0.21
AVG     66.1 1.03    

As you can see, there is a gigantic drop-off from the top 4 to OSU and Purdue, then a medium-sized drop-off to Minnesota, Indiana, and Illinois, then another big drop-off to the dregs of the conference. Which includes Nebraska, who the Bucks lost to last night. As of right now OSU is a significant step behind the class of the conference. But you should go read the whole article.
Of particular note is that Michigan lost Trey Burke, the national player of the year, and, so far at least, both their offense and their efficiency margin have actually improved (they scored 1.12 PP in Big Ten play last year). Beilein is one hell of a coach.