(I stupidly pushed my final NCAA piece too early. You get to see me edit in real time -Arturo 3/20 6:30 PM EST)
“All the time you’re saying to yourself, ‘I could do that, but I won’t,’–which is just another way of saying that you can’t.”- Richard P. Feynman, Surely You’re Joking, Mr. Feynman!
I’ve always loved NCAA tournament analysis and simulation pieces and I’ve always thought it would be cool to build one from the ground up. Logistics and time had always gotten in the way. Last year we took a first stab at it but i wasn’t quite what we wanted. The crux of the matter is that it had always been too damn hard to build a full on wins produced model for the NCAA they same way we did for the NBA.
I had to build a whole new way of doing it.
That is a word cloud using Final Four odds as worked out using the brand new and yet to be fully revealed NCAA wins produced model that I have been helping Professor Berri put together over the last few months. A fuller reveal is coming and you can get a hint as to what shape it’ll take if you listen to what Dave has been saying recently.
A huge shout out goes to College Basketball Reference for compiling the data in one place and to Ken Pomeroy for providing a great,great measuring stick and a place to validate my assumptions. I feel like throwing up the results of the full sim in right now:
That is the result of me taking the full wins produced numbers as well as the KenPom numbers for 2013 and converting them to a projected margin of victory in the tournament versus an average NCAA team. For the Wins Produced numbers I adjusted for strength of schedule (using the KenPom schedule adjustement) as well as adjusting for a shorter bench (as per the Half Baked Notion). For the KenPom numbers, I adjusted for pace (which I did not have to do for Wins Produced) and repeated the schedule and depth chart adjustment I did for Wins Produced. A nice handy-dandy formula to keep in mind is:
Point Margin per Game = – 18.2 + 38.0 Win Percentage
That was the key Formula in doing all my conversions. The end result was the previously shared Power Rank:
That is your Tournament Power rank with everyone’s expected point Margin against average competition. In simple terms, you want to project margin for a game? Look up the teams Point Margin and work out the difference using simple subtraction.
Now, I have a particular and unique method I use for working out the win probabilities which I will publish at some point and I find more accurate than what’s currently out there. In essence I worked out the actual math behind margin for a basketball game as opposed to approximating it. I am keeping that for myself for now but I am willing to put it in a table for you.
I’m also willing to share the results in an excel file.
How to close? I know, how about a table of the Top 200 Wins Produced per 40 Minutes Players sorted by team and an observation.
This leads to the following fascinating table:
|Team||Tourney Point Margin||Stars on Team (Top 200)||SuperStars on Team (>.250 WP40)||Star Points|
|North Carolina State||10.57||1||1||3|
|South Dakota State||6.24||2||0||2|
|San Diego State||14.00||2||0||2|
|Florida Gulf Coast||5.37||1||0||1|
|New Mexico State||7.57||1||0||1|
Kansas, Florida and Louisville are supremely top heavy teams with Creighton, Duke and Gonzaga not far behind. That depth would lead me to think that these teams could outperform expectations. A team like Indian is also interesting because they don’t play Oladipo enough.
-Arturo with an assist from DJ
PS As a bonus, here’s a cinderella table (5 seeds and up):