PennStateHoops.com Discussion Forum

Résumé Tracker 19-20

Consistent with the guidelines laid out by the NCAA Committee, will track our own predictive/results-based metrics and those of our opponents. We’ll work with KenPom until we get our first NET rankings and then adjust to that.

Current PSU Résumé (as of 2/3)

Results-based metrics (NET: 24, KPI: 19, SOR: 16) Average: 20
Predictive Metrics (BPI: 16, KenPom: 16, Sagarin: 20) Average: *17

Record by Quadrant
Quad 1: 5-3
Quad 2: 4-1
Quad 3: 2-1
Quad 4: 5-0

Current NET of our NC opponents:
UMES: 352
Wagner: 328
Georgetown: 48
Bucknell: 250
Yale: 50
Ole Miss: 111
Syracuse: 66
Wake Forest: 107
Alabama: 41
CCSU: 349
Cornell: 320

Current Conference NET
Sparty: 10
Maryland: 11
Ohio St: 20
Iowa: 21
Rutgers: 28
Illinois: 30
Michigan: 31
Wisconsin: 32
Purdue: 39
Minnesota: 44
Indiana: 52
Northwestern: 149
Nebraska: 170

Wins by NCAA Team Sheet Quadrant
Quad 1: @Georgetown, Maryland, VS. Iowa, Ohio St, @Michigan
Quad 2: VS. Syracuse, Yale, Alabama, Indiana
Quad 3: Wake, @ Nebraska
Quad 4: UMES, Wagner, Bucknell, CCSU, Cornell

Losses by Quadrant:
Quad 1: @OSU, @Rutgers, @ Minny
Quad 2: Wisconsin
Quad 3: VS Ole Miss
Quad 4:

Quadrant criteria;
Quadrant 1: Home vs a 1-30 team, Neutral site vs 1-50, Away vs 1-75.

Quadrant 2: Home vs a 31-75, Neutral site vs 51-100, Away vs 76-135.

Quadrant 3: Home vs a 76-160, Neutral site vs 101-200, Away vs 136-240.

Quadrant 4: Home vs a 161-353, Neutral vs 201-353, Away vs 241-353

7 Likes

Georgetown is currently a Q1 win? I’m kinda shocked.

1 Like

As long as their NET doesn’t drop below 75, it’s a Q1 win. They’ve had 2 very nice wins since they lost those guards.

3 Likes

Georgetown meets Syracuse this weekend. It’d be nice for Syracuse to edge out a close win so we can flip them into Q1 and GTown won’t really fall.

We’ve also got a shot of sneaking Yale into the top 75. That would be a really nice bonus Q2 win if they keep this up

3 Likes

Feels like we’re starting from a really good base of metrics this year (obviously hard to tell the whole picture pre-NET)

1 Like

That would be huuggee bonus… makes that comeback even better

That’s because it’s a true road game for us.
AND if you think about it, that’s the “W” we’re still hanging our hat on, so to speak. That’s still our best win (until tonight, that is!!! :slight_smile: )

1 Like

Don’t know if it matters either way once NET reshuffles. I suspect Cuse will be one of those squads with higher KP vs. NET given how many losses they already have.

1 Like

#16 in Sagarin, #18 in SOR, #19 in BPI and #23 in Kenpom

2 Likes

15 in T-Rank

Interesting to me that Kenpom didn’t budge much.

1 Like

So Kenpom wasn’t overly concerned by the OSU loss but wasn’t overly impressed with the MD win

1 Like

Well we won a game we were favored to win, so maybe that makes sense. And maybe that means we’re good.

1 Like

It’s hard to make big jumps in KenPom right now unless a team does something very unexpected. We were 1 point underdogs and won by 7, that’s not going to cause a massive jump in his system.

1 Like

Just heard on the Eye on College Hoops Podcast that the first NET rankings are coming out on Monday.

3 Likes
1 Like

Evaluating a team’s resume is a two-step process:

  1. How good was your schedule?
  2. Given the answer to #1, how well did you perform against that schedule?

NET is being used by the selection committee to answer question one. It’s an organizational metric to determine schedule strength.

Because NET is better at predictively ranking teams than RPI, the NCAA is now better equipped to answer question number one. The problem is question number two — which relies on the quadrant system.

Instead of relying on NET to tell us exactly how strong a win or loss is, the quadrant system bins results into arbitrary categories. It’s a very human way of doing things.

On an episode of Solving Basketball last year, Ken Pomeroy said it best:

The fact that humans are trying to look at a 30-game schedule and evaluate what all those wins mean compared to another team’s 30-game schedule and evaluate what all those wins and losses mean — it’s an impossible task for humans to do.

The most immediate problem with the quadrant system is the arbitrary binning.

  • A win over the #1 team in the country and a win over the #30 team in the country are essentially the same
  • A win over the #30 team in the country and a win over the #31 team in the country are hugely different

The next problem with the quadrant system is that — like Ken said — humans are just not good at interpreting these resumes. How does the committee compare a team that went 4-8 versus Quadrant 1 opponents to a team that went 0-2? We’ll go back to Ken again for the answer:

You’re basically judged on…What are your five best wins and what are your five best losses? You’re still not going to get credit for beating teams in the 100 to 200 range like I feel you should if you can do that on a consistent basis over a 20-game league schedule.

The underlying metric being used to form the quadrants is now better after the change from RPI to NET, but the problem is the quadrant system itself. And in all of the mayhem regarding the roll-out of NET last year, that’s been lost in the national discussion.

2 Likes

It doesn’t matter to KenPon whether you win or lose, it’s all about how you play the game. Play more efficiently and your rating will get better. Play worse and it will go down.

2 Likes

I just love guys who make fun of Kenpom without the foggiest idea of how it works and what it is representing.

2 Likes