Wednesday, February 10, 2016

Building a Mystery

Please note the following chart:



Kenpom's O-Rating is essentially points per possession times one hundred.  For reference, the scale loosely translates to:

124 - Damion Lee
115-124 - All CAA 1st Team Player
105-115 - Very good CAA Player
100-105 - CAA Starter
95  -100 - Potential CAA Starter
90 - 95   - CAA 6th Man
Below 90 - CAA Liability

So that Kenpom rating is a bit of who we are on offense.  I stress on offense because this next chart is where things get interesting.  It shows a players +/- rating over a 40 minute span.  That's how much Drexel outscores or gets outscored by their opponents when this player has been on the floor:



This illustrates a much larger point.  Anyone good, when they want to, can make numbers tell virtually any story that they want.  If you want to argue Mohammed Bah is terrible, you can from the above.  If you want to argue that the team would be lost without Mohammed Bah, you can from the above.  It's with investigation, context, and study, that you can find the underlying story behind the digits.  Knowing that, let's take a deeper look into plus/minus.

First, note the two plus/minus columns before.  In the first column, I have removed any part of games following the final sub before there were two minutes remaining.  The reason is simple, and is best seen when looking at Terrell Allen and Rashann London.  When subbing in a close game during that time of the game, coaches will go offense/defense.  Bruiser generally has Terrell in on offense only, and London in for defense.  That will skew a stat like +/- because Rashann London can only see a negative impact if he only plays on defense, and visa versa for Terrell Allen.

Beyond that, when looking at the +/- there is, among other issues with the statistic, a sample size issue.  None of these numbers have stabilized yet, and with Mohammed Bah's limited playing time, his is even more unstable than the other significant contributors on the team.  With those disclaimers in place, lets allow ourselves to be blown away:

Mohammed Bah has a positive +/- on a 3-20 basketball team.  With an O-Rating in the 70's!

Additional perspective:  This "really bad", "you don't need to guard us" (thanks to Bruiser Flint for the quotes) team has outscored their opponents when Mohammed Bah has been in the game.  And that's not just any opponents - the Dragons have played a top 50 strength of schedule nationally.  They've outscored some very good opponents, but only when Bah was on the floor.

And that's where numbers end.  They have provided the question, and from there the smart fan looks to the tape, examines him closer on the floor, and tries to find out what he is doing that is so beneficial.  Could it be that past DU teams only had one forward scorer on the floor at a time, and he is a better fir in that system?  Perhaps Bah's high screen's/ high post offensive sets clear out the bodies from under the basket making it easier to get to the rim?  Is it a defensive contribution, helped by him being larger than most of the other forwards?  Is it the plays that the coaches are calling when he is in the game?  Whatever it is, it's worth looking at, because Drexel seems to be a much, much better team when Mohammed is on the court.

We find the questions to then look for in the tape.  That's how numbers can be best used in the modern game.  And that is what has given us this mystery.

So fans what is your hypothesis?

No comments:

Post a Comment