GW Hoops

You are not logged in. Would you like to login or register?



3/02/2022 3:13 pm  #1


Maybe The Advanced Metrics People Can Help Me Out Here

My beloved Badgers clinched at least a tie for the Big 10 regular season title last night by completing a two-game sweep of Purdue.  Not to count chickens but the sole title is theirs with a home win against Nebraska on Sunday.  And yet, they've lagged throughout the season in the NET rankings and other metrics.  I gather this is due to two things:  first, they were unranked to begin the season (picked to finish 10th in the Big 10 by the coaches and media...can you imagine?) and because they tend to play a lot of close games which must mean their efficiency ratings aren't too hot.  A quick comparison to Duke, who played in  a considerably easier ACC than did Wisconsin in the Big 10:

Duke: 26-4, 5-1 in Q1, 11-3 in Q1 & 2, 7-1 in Q3, 8-0 in Q4
Wisc: 24-5, 8-3 in Q1, 16-4 in Q1 & 2, 4-1 in Q3, 4-0 in Q4

Of course Duke is ranked higher but their name has a lot to do with that.  As for the NET, Duke is #6 and Wisconsin is #20.  If this were flipped, many would argue that Duke had the better season due to having almost as good a record against a much more difficult overall schedule.

What's the big difference?  It would have to be that Duke plays more dominantly on the court.   Greater margin of victories.  Better efficiency numbers.

So, here's my thought.  Wisconsin added to what I believe is an amazing streak last night.  In solely games that have been decided by 6 points or fewer, Wisconsin has now won its last 15 times.  You have to go back to the 1980's to find a streak this long.   Doesn't this have to send a signal that when a game is on the line, Wisconsin has an excellent chance of winning?  And correct me if I'm wrong but this sounds like it means absolutely nothing in the advanced metrics world.  How can this be?  Isn't there something to be said for consistently pulling out close victories?  Wouldn't you rather have this know-how than be a team who comfortably wins during the regular season only to be confronted by tougher opponents in the tourney who may have a leg up at winning a tight game?

Am open to learning what I may be missing here.

 

Last edited by Gwmayhem (3/02/2022 3:45 pm)

 

3/06/2022 1:21 am  #2


Re: Maybe The Advanced Metrics People Can Help Me Out Here

So the answer to your question is essentially that the primary computer metrics that get referenced (KenPom and the NET) tend to me more predictive based on data than they are contextual.

For example, with KenPom, based on the data points that are going into his formula, he is generating a prediction for how each game will play out (basically a point spread). Based on KenPom's metrics right now, if Wisconsin played Penn St in Madison tomorrow, Wisconsin would be roughly a 11.5 point favorite. Assuming it was the same type of spread when they met earlier in the season, Wisconsin's two point win actually hurt their ranking in KenPom rather significantly because they missed the prediction by 10 points. Wisconsin and Providence's knack for winning close games is almost derided by the "luck" category in KenPom, where the badgers rank 4th and the Friars 3rd.

My issue with winning more close games than the law averages suggests being classified as "lucky" is that the computer is not watching games and evaluating "lucky" plays throughout the course of a game. For example a banked in three from 32 feet with 12 minutes left in the first half goes into the dataset the exact same way that a swished three off beautiful ball movement does. Now Wisconsin's two baked in shots to beat Purdue did involve quite a bit of luck, but the reason Wisconsin and Providence have fared so well in tight games this year has more to do with the fact they don't turn the ball over late and always seem to generate a shot in closing moments and they defend without fouling on the other end. These are skills and not just things that can be defined by expected values and the corresponding results in small sample sizes.

My hope for both programs is that the committee will value their results more so than their metrics, as Wisconsin will have a legit case to be a 1 seed if it wins both the Big 10 regular season and tournament titles and Providence should get consideration for a 2 seed if it can do the same in the Big East.

 

Board footera

 

Powered by Boardhost. Create a Free Forum