RPI is no longer to be used for NCAAT selection

[CuseJacket]

In all seriousness... I would love to hear how margin of victory capped at 10 is a good idea.

Describe to me how winning by 2 when the clock hits 0:00 is materially different than being up by 2 with 25 seconds and the other team fouls 4 times to lose by 8. It simply is not.

This is a problem if true. And since we don't know how much it matters, it is entirely fair to speculate that it is possibly going to be given too much weight.

 

[CuseJacket]

Also, @ESPNjacket, if you truly believe the formula measures the right things, let's pretend it's perfect. How could anyone game it? How does one game Kenpom and offensive/defensive efficiency without just simply being a better team? I think you didn't respond to that because of my bad example at the end.

Also, preemptively I will acknowledge that the original article posted said the NCAA was unsure if they'll share the formula, so for all I know they'll share it tomorrow and my issue with this is a moot point.

 

[ESPNjacket]

You are the one taking shots at something you don't know much about. Did you read the Sporting News link? Start there. Make a useful point and I'll be happy to respond.

 

[CuseJacket]

You are the one taking shots at something you don't know much about. Did you read the Sporting News link? Start there. Make a useful point and I'll be happy to respond.

Yes, I did, prior to my last series of posts. Would you mind responding to them? I'm not sure what the disconnect is. Maybe you know more than me or what's posted in the articles?

I'll try stating my first point one last time and drop it, unless you give me something to chew on. I fully understand and acknowledge that KenPom is better than RPI. I simply have no clue how it's factored in to this new formula. Please educate me. Sincerely asking, because the rub for me is everything I've read suggests no one knows. And the NCAA may intentionally keep the formula secret.

My 2nd point, I'll try to make a different way, at risk of you taking this single example and ignoring the rest of this and my prior posts again. My boss has been talking about instituting a new performance management process for our department for years. Lot of new ideas throughout the years, and nothing ever executed or executed well. If he comes to me and tells me he consulted Jack Welch, Steve Jobs, etc. and he has a new plan. it'll mean very little to me until I understand how it works. I wouldn't have a ton of faith that it'll get done well. Much like the NCAA's bungling of a number of matters of importance historically.

 

[mstranahan]

My fear is your first point. If they don't tell people how the new system works, I fully expect the following will happen in the smoke filled room of secrecy that is the selection committee.

Committee Member #1: Hey, how did small conference Team X get in and my conference's Team Y didn't?
Committee Member #2: They had a better NET score
Committee Member #1: But we all know Team Y is better and my conference will get a bigger tourney payout if Team Y gets in
Committee Member #2: And who cares if small conference gets a bigger payout
Committee Member #1: I'll make sure to support you when you want your bubble team to get in
Committee Member #2: Sold! Team Y gets my vote. Screw Team X

 

[mstranahan]

Also, the selection committee only has to figure out the last 8 or so spots in the field.

68 total bids
32 automatic bids
Power conferences get about 20 of those spots, before we even talk about bubble teams (ACC, Big East, SEC, Big Ten, Big Twelve, PAC 12)

That leaves about 12 spots for the bubble teams. Some of them are always easy because they're strong teams in a power conference or they won regular season in a non-power conference but lost in tourney finals. Call that 4 teams each year

That leaves roughly 8 bubble teams that the committee will actually look at NET or RPI or whatever.

 

[orientalnc]

Also, the selection committee only has to figure out the last 8 or so spots in the field.

68 total bids
32 automatic bids
Power conferences get about 20 of those spots, before we even talk about bubble teams (ACC, Big East, SEC, Big Ten, Big Twelve, PAC 12)

That leaves about 12 spots for the bubble teams. Some of them are always easy because they're strong teams in a power conference or they won regular season in a non-power conference but lost in tourney finals. Call that 4 teams each year

That leaves roughly 8 bubble teams that the committee will actually look at NET or RPI or whatever.

None of this addresses the problem Cuse has with NET.

I think your problem is simply a fact of any metric based system whose results are somewhat accurate, but not precise enough for total buy-in from every team/coach. Our town had an election for the town commission 2 years ago where the final vote count was exactly tied between two candidates for one seat. The state election board ordered a recount that found one more vote for the candidate they declared the winner. But, close elections can be appealed, which is what happened. This second recount found another vote. Tied again. This time they tossed a coin and the loser was satisfied all that could be done was, in fact, done. It was a close election and someone had to be the winner.

 

[ESPNjacket]

I fully understand and acknowledge that KenPom is better than RPI. I simply have no clue how it's factored in to this new formula. Please educate me. Sincerely asking, because the rub for me is everything I've read suggests no one knows. And the NCAA may intentionally keep the formula secret.

Kenpom's formula is also a secret yet you fully understand that it is better than RPI. I'm pretty sure any model that uses “game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency, and the quality of wins and losses" and was trained with actual results from past seasons is inherently better than RPI.

If you have a contrary example feel free to share it. If you have another bad analogy, please feel free not to.

 

[Peacone36]

The obvious answer here is......Syracuse was often seen in a favorable light by the rpi and that may go away with the new metric

 

[CuseJacket]

Kenpom's formula is also a secret yet you fully understand that it is better than RPI. I'm pretty sure any model that uses “game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency, and the quality of wins and losses" and was trained with actual results from past seasons is inherently better than RPI.

If you have a contrary example feel free to share it. If you have another bad analogy, please feel free not to.

Here's my last bad analogy This is not far from how I interpret this thus far.

NCAA: "Hey KenPom, thanks for the insight. We'll include you in our new formula but we think we can do better. Dave Gavitt believes the NCAA can get this right. Just take us at our word."

Anyway, it really doesn't matter a whole lot either way. Not going to lose sleep over it. Just seems crazy that we will reverse course in transparency. I think it'll be interesting, if kept secret, what will happen on selection Sunday when a team with a NET ranking of 53 is selected over a team with a NET ranking of 41. How angry are we supposed to be if we don't understand the magnitude of 12 NET spots?

 

[CuseJacket]

The obvious answer here is......Syracuse was often seen in a favorable light by the rpi and that may go away with the new metric

Now you gon' did it. They were certainly the benefactors of the awful quadrants system last year.

I'm choosing to believe Syracuse is headed back to performance of the 10 years prior to probation, not the last 4. Maybe the RPI did help them get these seeds: #4, #5, #3, #1, #3, #1, #4, #4.

I don't know why you and @kg01 incessantly bring up Syracuse on a GT forum.

 

[Peacone36]

Now you gon' did it. They were certainly the benefactors of the awful quadrants system last year.

I'm choosing to believe Syracuse is headed back to performance of the 10 years prior to probation, not the last 4. Maybe the RPI did help them get these seeds: #4, #5, #3, #1, #3, #1, #4, #4.

I don't know why you and @kg01 incessantly bring up Syracuse on a GT forum.

My guess would be because the guy running the show has a pro-cuse handle

 

[AE 87]

Here's my last bad analogy This is not far from how I interpret this thus far.

NCAA: "Hey KenPom, thanks for the insight. We'll include you in our new formula but we think we can do better. Dave Gavitt believes the NCAA can get this right. Just take us at our word."

Anyway, it really doesn't matter a whole lot either way. Not going to lose sleep over it. Just seems crazy that we will reverse course in transparency. I think it'll be interesting, if kept secret, what will happen on selection Sunday when a team with a NET ranking of 53 is selected over a team with a NET ranking of 41. How angry are we supposed to be if we don't understand the magnitude of 12 NET spots?

Now you gon' did it. They were certainly the benefactors of the awful quadrants system last year.

I'm choosing to believe Syracuse is headed back to performance of the 10 years prior to probation, not the last 4. Maybe the RPI did help them get these seeds: #4, #5, #3, #1, #3, #1, #4, #4.

I don't know why you and @kg01 incessantly bring up Syracuse on a GT forum.

 

[CuseJacket]

My guess would be because the guy running the show has a pro-cuse handle

He's probably an upstanding, classy, always on topic guy, if I had to guess.

 

[Peacone36]

He's probably an upstanding, classy, always on topic guy, if I had to guess.

Yes @kg01 is typically that guy.

 

[AE 87]

Yes @kg01 is typically that guy.

The dictionary just called. They want to use this post as an example of irony.




Eeeeeyyyyyooooo

 

[kg01]

You guys are all jerks.

Once @CuseJacket dies in a hail of gunfire hollering, "You'll never take me alive, coppers!", and I take over as mod ... you'll all regret mocking me.

 

[ESPNjacket]

Here is a link to the early NET rankings. GT is at 81 and Northwestern 83.

http://extra.ncaa.org/solutions/rpi/Stats Library/NET Nitty Gritty - November 26, 2018.pdf

 

[YlJacket]

Just a couple more

Sagarin has NW at 58 and GT at 60
A very flawed RPI has GT at 82 and NW at 138.

I will be curious to see if Vegas has a big spread to NW here.

 

[MWBATL]

Here is a link to the early NET rankings. GT is at 81 and Northwestern 83.

http://extra.ncaa.org/solutions/rpi/Stats Library/NET Nitty Gritty - November 26, 2018.pdf

So you're saying we've got a chance?