Pitt's Capel goes after Big 12 and efficiency ratings

[auror]

Any of these metrics are useful in “bands.” The problem is people use them as if they were precise.

There is no evidence that the committee uses the NET to be precise. Check this post and see where a team's actual NET ranking lies:

 

[Husky68]

I'd like to see someone do a retro analysis of last 25 years NCAA tournament results and see what factors were in play for teams that won or exceeded expectations. Think an important factor would be having UConn on the front of the jersey.

 

[businesslawyer]

The Big 12 clearly has a scheduling philosophy that it is advising its teams to follow, and it is working. Are you saying you are smarter than the statisticians at the Big 12? You should call them and tell them they are wrong.

Is there a single point you can try to make without overbeating it to death? This year, the XII’s OOC strength was low. Last year it was high and guess what — the computers still showed them as the strongest conference. So maybe there was a conference wide move to abandon what was working for them and try something else, and maybe — just maybe — this year their schedules just happened to work out to be easier.

But I know your seeing more than one possible right answer will never happen so ….

 

[kobe]

 

[McLovin]

Cap MOV for any Q3 & 4 game at 15 points. Eliminate neutral site games outside of the MTE's / select events (Champion's Classic, Jimmy V, etc.) and require each team to be designated "home" or "away" for NET purposes, even if Baylor wants to play Purdue in Sioux Falls, SD. Put more weight on non-conference Q1 & Q2 wins.

Wouldn't solve all the issues with gaming the NET, but it would make it harder and also make basketball a better product (by incentivizing more non-con high major home & home series).

 

[pj]

The Big 12 clearly has a scheduling philosophy that it is advising its teams to follow, and it is working. Are you saying you are smarter than the statisticians at the Big 12? You should call them and tell them they are wrong.

Isn't it the Big Ten's scheduling philosophy that is the best? Annually they get a lot of teams in and with high seeds, but underperform. The Big 12 does about as expected.

 

[kobe]

big 12 goes to 20 conference games next year so basically you can remove 28 buy games from the conference overall's ooc.

wonder what loser conferences will whine about then?

 

[Vowelguy]

1) the NET is one of about a dozen factors the committee uses
2) if you think the NET is so undervaluing your team, then show me a rating that where that’s not the case

 

[nelsonmuntz]

Is there a single point you can try to make without overbearing it to death? This year, the XII’s OOC strength was low. Last year it was high and guess what — the computers still showed them as the strongest conference. So maybe there was a conference wide move to abandon what was working for them and try something else, and maybe — just maybe — this year their schedules just happened to work out to be easier.

But I know you’re seeing more than one possible right answer will never happen so ….

I didn't realize there were so many Big 12 fanboys on this site.

It is not a coincidence when every school from a conference schedules the same way and then runs the score up on their bad opponents. BYU is the posterboy for this. They have 3 wins over teams in the top 167 of college basketball, and are 8-7 in conference, yet are 10 in the NET. But they are not the only one.

 

[auror]

I'd like to see someone do a retro analysis of last 25 years NCAA tournament results and see what factors were in play for teams that won or exceeded expectations. Think an important factor would be having UConn on the front of the jersey.

Every time this is done, it's basically just "Be very good at basketball on both ends of the court." The style doesn't necessarily matter, but offense is a little more important than defense. Playing fast is a little better than playing slow. Having a deep run NCAA experienced coach is better than not but not the end of the world (see Hurley). Having a good seed is generally important. Having both good guards and good bigs.

 

[ecasadoSBU]

The biggest issue with the NET (for me) is that the formula isn't public.

When something is guarded as a secret it can easily lead to abuse because someone will always find out the formula and then there is asymmetry of information which leads to unfair advantages.

The formula needs to be public, transparent. If everyone knows the formula then the benefits are negated.

The beauty of the RPI was the transparency in its formula and how easy everyone could access it. If everyone tried to game it, the effects of gaming it would end up being neutralized and that's fair game. The RPI was simple and it worked for so long. So why change it?

At the end of the day - there is no perfect way to select 68 teams. If perfection isn't possible then objectively nothing makes the NET any better than the RPI. Its just another selection tool. The only thing we ended up losing was transparency and gave the keys of the castle to the NCAA.

A program would get mad that the RPI left them out. But if a team can't make a cut of 68 they are more to blame than any given selection criteria

The conspiracy kitty in me tells me the NCAA ultimately had bad intentions with the NET. Something gives me the impression that the end goal was to get more Power-6 teams into the NCAA with reduced backlash by reducing the transparency.

 

[auror]

At the end of the day - there is no perfect way to select 68 teams. If perfection isn't possible then objectively nothing makes the NET any better than the RPI. Its just another selection tool. The only thing we ended up losing was transparency and gave the keys of the castle to the NCAA.

A program would get mad that the RPI left them out. But if a team can't make a cut of 68 they are more to blame than any given selection criteria

The conspiracy kitty in me tells me the NCAA ultimately had bad intentions with the NET. Something gives me the impression that the end goal was to get more Power-6 teams into the NCAA with reduced backlash by reducing the transparency.

Just because a tool isn't perfect, doesn't mean it can't be better. That's laughable. Would you rather use a spoon or a knife to open a can of beans instead of a can opener.

There was no conspiracy in getting rid of the RPI. People were just consistently mad it spit out results like Dayton being 6 ranks better than us this year and us being 10th in general this year. If your algorithm doesn't have us as at least a top 5 or 6 team this year, start over from the drawing board.

It affected more than just selection, also seeding. Some teams got some awfully tough draws that they should not have due to the RPI.

 

[storrsroars]

There was a team in the seventies. I was a team member the 76 -77 season. Practices and home games were at Wright Tech and we played the other branches, Norwalk Tech and CC and some other community colleges. I disagree with the athleticism comment.

You sure you have your years right? I was at Stamford Branch from fall '74-fall '76. Don't recall a thing about a hoops team starting up before I left for Storrs. FTR, I was 6'2" and I don't recall meeting anyone there taller than I was.

 

[EROCK CONN]

Capel goes after Big 12 scheduling and metrics, and I agree with him. I don't agree with a lot of what he said, especially his ACC comments, but overall it was a good rant.

I always thought efficiency ratings were stupid because a conference doing what the Big 12 did this year was completely predictable. There was nothing wrong with the RPI.

If a tree falls in the forest ……

 

[cullaville]

You sure you have your years right? I was at Stamford Branch from fall '74-fall '76. Don't recall a thing about a hoops team starting up before I left for Storrs. FTR, I was 6'2" and I don't recall meeting anyone there taller than I was.

I graduated from Storrs in 79 and went to the Stamford Branch on Scofieldtown Road my freshman and so[homore years and was a team member my sophomore year which would have been the 76 - 77 school year If i remember correctly we had twopr three players in the 6'5 range. Practices and games were held at Wright Tech, Games were against the other UCONN branches and community colleges. The team had been in existence prior to that becasue players from past teams would come in for scrimmages.

I do remember the baseball team as well

 

[ecasadoSBU]

Just because a tool isn't perfect, doesn't mean it can't be better. That's laughable. Would you rather use a spoon or a knife to open a can of beans instead of a can opener.

There was no conspiracy in getting rid of the RPI. People were just consistently mad it spit out results like Dayton being 6 ranks better than us this year and us being 10th in general this year. If your algorithm doesn't have us as at least a top 5 or 6 team this year, start over from the drawing board.

It affected more than just selection, also seeding. Some teams got some awfully tough draws that they should not have due to the RPI.

Fair point. I kinda didn't express my thought too well there.

At the end of the day the bottom line for me is that If both tools are imperfect selection tools then I will always choose the transparent, public formula that levels the playing field for everyone with no information asymmetry. Take the control away from the NCAA and a few power (who can easily leak it for money) brokers and return it back to everyone!

The public till this day doesn't know the exact formula behind the NCAA Evaluation Tool (NET). That opens the possibility of a ton of abuse and corruption. That's why we are having this conversation right now. The conspiracies and accusations of abuse will continue to happen until the NCAA opens up and displays the formula to everyone.

 

[willie99]

I say let Vegas tell us the best 32 at large teams, don't have to reinvent the wheel

As for any other tweak, there should be some process to enable the best schools from non power conferences to participate more. An 18-15 or so team should not can an at-large invite. Especially when their conference exits early year after year after year

 

[Hey Adrien!]

This made me chuckle:



Pitt hasn't played Duquesne, an A10 school in Pittsburgh, since 2018. Over that time, the Dukes have three 18+ win seasons and .500+ conference records.

 

[storrsroars]

1) the NET is one of about a dozen factors the committee uses
2) if you think the NET is so undervaluing your team, then show me a rating that where that’s not the case

Remember that NET also determines who's a Q1, Q2, etc. And that is influenced by having numerous wins over Q3/Q4.

 

[Scoe]

If you can’t play your way in then cry your way in. I’m not a numbers guy but my eyes tell me that the ACC sucks.

 

[kobe]

Remember that NET also determines who's a Q1, Q2, etc. And that is influenced by having numerous wins over Q3/Q4.

pitt has 14 Q3/Q4 wins which is more than anyone in the big 12 and their net is still 44.

maybe if you're a good team you have a good net and if you stink your net stinks too?

 

[storrsroars]

pitt has 14 Q3/Q4 wins which is more than anyone in the big 12 and their net is still 44.

maybe if you're a good team you have a good net and if you stink your net stinks too?

You'll get no argument from me that Pitt played a soft OOC sked. What I will say is that I think based on the past six weeks of play, Pitt would be a much more interesting tourney field addition than a bunch of B1G, B12, or even a couple of other ACC and BE teams around the bubble or projected for 10-11 seeds.