FanPost

STATSWATCH: Its A Blizzard of Georgetown Basketball Statistical Analysis!

Net Player Efficiency Ratings

Thanks to the snow day, I had some time to work with the line-up data I've been gathering. Some quick notes on the data before we get to the analysis:

  • I calculate the statistics from play-by-play feeds and I'm dependent on the other schools to provide information. In the Big East, Georgetown, DePaul and Providence do not publish play-by-play feeds and when I can't find one, I use the ESPN feed to calculate the tempo-free box scores. ESPN does not provide sub data though.
  • Some of the sub information is low quality. There will be data entries where 1 player subs in and 4 players sub out. I try my best to correct for the errors, but its not going to be perfect. However, most of the mistakes occur at the end of games when coaches go with "offense/defense" subs and so the impact is less than it seems.
  • This data set includes 8 of the post-Josh Smith games. DePaul and both Providence games are excluded because I don't have the data. I'm considering adding some of the other games and excluding line-ups that include Josh Smith, so let me know what you think.

From the data set I calculated a variety of metrics comparing the performance of the team when certain players are on and off the court. Although there is always going to be noise, this is useful to identify possible ways in which players contribute to the team, but are not necessarily visible in individual statistics. It takes a while to walk through the different fields for each player, so I'll save those for later write-ups.

For every player on the team, I calculated the offensive and defensive rating of Georgetown when the player is on and off the court for the combined 8 games in the data set. This is different than the offensive and defensive ratings that appear on Sports Reference, which are calculated based on the relationship between a player's individual stats and the team's stats. The difference is much more pronounced on defense, where the alternative method is biased towards front court players that have more rebounds and blocks. The net method is much better for identifying players that can impact the team outside of traditional box score statistics.

For offense, a positive number indicates that Georgetown scores X more pts per 100 possessions when the player is on the floor than on the bench and for defense, a negative number indicates that Georgetown surrenders X less points when the player is on the floor than on the bench.


Off Rating Def Rating Net Rating
CAMERON,REGGIE 22.3 -3.1 25.4
HOPKINS,MIKAEL -0.3 2.7 -3.0
LUBICK,NATE -2.7 14.1 -16.8
SMITH-RIVERA,D'VAUNT 12.5 -4.7 17.2
STARKS,MARKEL 20.4 6.1 14.4
BOWEN,AARON -10.7 -2.1 -8.6
CAPRIO,JOHN 7.4 25.0 -17.6
AYEGBA,MOSES 0.8 -5.7 6.6
TRAWICK,JABRIL -19.2 -8.3 -10.9

Some Comments

  • Jabril has only 64 possessions and I think his low net offensive rating is a result of small sample size and should change with more games.
  • I was surprised that Moses didn't seem to impact the offense much given that his skills are limited on that side of the court.
  • Reggie's ability to space the floor is clearly valuable on offense. I know that there was some talk before he came in that he wasn't a great defender, but it appears that Coach Thompson has found ways to use him in the scheme.
  • Markel's efficiency is down this year, but this is clearly the result of him having to create shots for a team that can struggle to score at times. His +20.4 net offensive rating though shows how valuable he really is.
  • Nate's contributions to the offense don't seem to outweigh his deficiencies on defense, but he still gets a lot of minutes. I don't want to jump-the-gun too much on numbers based on 8 games, but as an internet blog community, I really think we should all continue to monitor this.

BPI Distribution Data

Most indices present information as an average score across all performances with various weightings and adjustments, and then projections are made using the adjusted scores. A good example of this are the Log5 Win Probabilities that can be calculated with Sagarin or KenPom ratings. Once the tournaments get near, you'll start to see sites posting win and advancement probabilities using this method.

The Log5 Method is considered to be one of the better measures for estimating win probabilities and I use it in my own model for weighted bracket games. But its still a single input model that solely relies on the first moment of the team's performance (the average adjusted score). Think of the infamous "tournament resumes" where some teams play consistent throughout the season and others have great wins, but a lot of bad losses. To model the outcome of a single game, to some extent, I think volatility matters.

So to get a sense of what the distribution of Georgetown's performances look like, I ran some descriptive statistics for four series: the past two seasons, the current season and the current season excluding games without Jabril. I want to do some further research to put these numbers into perspective, but I thought the results were interesting enough to share:

2012 2013 2014 2014 w/ Jabril
N 32 32 24 19
Average 80.58 81.84 70.72 79.88
St. Dev 19.83 18.68 24.87 20.59
Skew -1.41 -1.25 -0.79 -1.82

Here's a histogram of the data series. Let's get this to the selection committee.

Butler Tempo-Free Box Score

1st Half 2nd Half Total 1st Half 2nd Half Total
Pace 30 29
Points 29 42 71 25 38 63
Efficiency 99.7 145.8 123.2 91.1 132.9 112.5
Dunk % 100.0% 100.0% 100.0% 100.0% N/A 100.0%
Layup% 66.7% 87.5% 76.5% 50.0% 90.0% 72.2%
(Dunk + Layup)% 70.0% 88.9% 78.9% 55.6% 90.0% 73.7%
2-Pt Jumper% 38.5% 54.5% 45.8% 14.3% 16.7% 15.0%
2-Pt% 52.2% 70.0% 60.5% 30.4% 62.5% 43.6%
3-Pt Jumper% 16.7% 0.0% 8.3% 50.0% 25.0% 40.0%
Jumper% 31.6% 35.3% 33.3% 25.0% 20.0% 23.3%
FT% 100.0% 82.4% 84.2% 50.0% 88.2% 81.0%
Dunk Rate 2.9% 2.8% 2.8% 2.7% 0.0% 1.4%
Layup Rate 25.8% 22.5% 24.2% 21.8% 30.8% 26.0%
(Dunk + Layup) Rate 33.5% 30.5% 32.0% 30.2% 33.9% 32.1%
2-Pt Jumper Rate 37.3% 31.0% 34.1% 38.1% 18.5% 28.9%
2-Pt Rate 70.7% 61.5% 66.1% 68.3% 52.4% 61.0%
3-Pt Jumper Rate 17.2% 16.9% 17.1% 16.3% 12.3% 14.4%
Jumper Rate 54.5% 47.9% 51.2% 54.4% 30.8% 43.3%
FT Rate 2.5% 21.1% 11.9% 4.8% 23.0% 13.3%
Effictive Field Goal % 46.6% 53.8% 50.0% 39.7% 57.5% 46.9%
True Shooting % 48.5% 62.7% 56.0% 40.6% 69.1% 54.1%
Assist Rate 53.8% 50.0% 51.9% 30.0% 45.5% 38.1%
Unforced TOV Rate 11.5% 2.8% 7.1% 10.9% 6.2% 8.7%
Forced TOV Rate 2.9% 2.8% 2.8% 5.4% 9.2% 7.2%
TOV Rate 14.3% 5.6% 9.9% 16.3% 15.4% 15.9%
Block Rate 6.9% 3.8% 5.5% 0.0% 10.0% 4.1%
OR% 33.3% 46.7% 40.0% 45.0% 33.3% 41.4%
DR% 55.0% 66.7% 58.6% 66.7% 53.3% 60.0%

Providence Tempo-Free Box Score

Home - Georgetown Away - Providence
1st Half 2nd Half Total 1st Half 2nd Half Total
Pace 32 33
Points 32 51 83 36 35 71
Efficiency 102.8 153.4 129.2 109.2 109.0 108.2
Dunk % N/A 100.0% 100.0% 0.0% N/A 0.0%
Layup% 25.0% 100.0% 50.0% 37.5% 40.0% 38.9%
(Dunk + Layup)% 25.0% 100.0% 53.8% 30.0% 40.0% 35.0%
2-Pt Jumper% 60.0% 54.5% 57.1% 50.0% 40.0% 43.8%
2-Pt% 44.4% 68.8% 55.9% 37.5% 40.0% 38.9%
3-Pt Jumper% 30.0% 33.3% 31.8% 45.5% 20.0% 33.3%
Jumper% 45.0% 43.5% 44.2% 47.1% 30.0% 37.8%
FT% 63.6% 89.5% 80.0% 90.0% 76.5% 81.5%
Dunk Rate 0.0% 2.6% 1.3% 5.5% 0.0% 2.5%
Layup Rate 22.3% 10.4% 16.2% 22.0% 23.0% 22.5%
(Dunk + Layup) Rate 25.1% 15.0% 19.9% 29.1% 28.2% 28.6%
2-Pt Jumper Rate 27.9% 28.7% 28.3% 16.5% 23.0% 20.0%
2-Pt Rate 53.0% 43.7% 48.2% 45.6% 51.2% 48.7%
3-Pt Jumper Rate 27.9% 31.3% 29.6% 30.2% 23.0% 26.3%
Jumper Rate 55.8% 60.0% 58.0% 46.7% 46.0% 46.3%
FT Rate 13.5% 21.8% 17.8% 12.1% 17.2% 14.9%
Effictive Field Goal % 44.6% 60.7% 52.7% 50.0% 36.7% 43.0%
True Shooting % 48.7% 70.1% 60.0% 57.3% 46.7% 51.5%
Assist Rate 45.5% 53.3% 50.0% 63.6% 40.0% 52.4%
Unforced TOV Rate 5.6% 0.0% 2.7% 11.0% 6.9% 8.8%
Forced TOV Rate 2.8% 5.2% 4.0% 2.7% 6.9% 5.0%
TOV Rate 8.4% 5.2% 6.7% 13.7% 13.8% 13.8%
Block Rate 10.7% 10.7% 10.7% 3.7% 0.0% 1.8%
OR% 22.2% 35.7% 28.1% 18.8% 45.0% 33.3%
DR% 81.3% 55.0% 66.7% 77.8% 64.3% 71.9%

Adjusted 3-pt Makes

This doesn't account for shot selection, but can help to calm down everyone's emotional swings by trying to adjust for volatility. A positive number indicates that the expected points were greater than the realized points. For players w/ 25+ attempts, shooting percentages are from this season and if not shooting percentages for each player are from the 2012-2013 season when available. Otherwise, the team's 3-pt percentage for the 2013-2014 season is used in the calculation for each player.

BUTLER -1.6 / GTOWN +7.1 / GTOWN +8.7

PROV -0.1 / GTOWN -2.6 / PROV + 2.7

Stay Casual, my friends.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Casual Hoya

You must be a member of Casual Hoya to participate.

We have our own Community Guidelines at Casual Hoya. You should read them.

Join Casual Hoya

You must be a member of Casual Hoya to participate.

We have our own Community Guidelines at Casual Hoya. You should read them.

Spinner

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9347_tracker