More NET rankings shenanigans

[Go Budaw]

Before the game yesterday, we were one spot back of Auburn in the NET rankings and 27th overall. Now after beating them, we are two spots back of them and 29th overall. I think this formula might be even more 17ed than RPI. The good news is we are still well on the good side of the bubble, but who knows what the remaining games will yield or what they will do to our ranking.

 

[Apoplectic_Objurgator]

Before the game yesterday, we were one spot back of Auburn in the NET rankings and 27th overall. Now after beating them, we are two spots back of them and 29th overall. I think this formula might be even more 17ed than RPI. The good news is we are still well on the good side of the bubble, but who knows what the remaining games will yield or what they will do to our ranking.

how big of a deal was it that we didnt try to win by ten - did we get outsmarted at the end

 

[msstate7]

We really needed Clemson to not collapse yesterday

 

[QuaoarsKing]

[tweet]1067431341975777280[/tweet]

 

[SixtInning]

Is the size of the red blocks supposed to indicate how much that component is weighed? So something like 40-20-20-10-10?

And is the location of the game not addressed twice? For that matter, the winner of a game is factored in 3 times, so that definitely favors teams with a bunch of wins against crap schedules. Strength of schedule was 75 pct of the RPI, and it seems waaay less than that in the NET.

 

[Go Budaw]

how big of a deal was it that we didnt try to win by ten - did we get outsmarted at the end

I think the difference between winning by 8 and 10 is negligible. There are no bonus points for winning by 10, thats just where they cap the “margin of victory” component of the algorhithm, which only represents a small part of one out of the 4 components. It shouldn’t make a huge difference.

 

[Go Budaw]

[tweet]1067431341975777280[/tweet]

That’s pretty telling. For reference, the old RPI has us at 23 now, with Auburn at 33. We are almost exactly at our current national rank. I think I’d like the old RPI back and get rid of this NET crap.

 

[patdog]

That’s exactly what I thought when I saw the factors (as far as I know they’ve never actually released the formula). It would be hard to come up with something worse than RPI, but I think they’ve done it. They should just use Sagarin rankings & be done with it.

 

[Dawgzilla]

I agree. I think people got confused when this first came out. You don't get bonus points for winning by 10. 10 is just the maximum number for margin of victory...so if you win by 20, it only counts as winning by 10. At least, that's the way I understand it.

I was watching on TV, and the announcers made it sound as though Pearl and Howland had signaled each other or something that the game was over. Auburn didn't foul, and in exchange we didn't try to score.

 

[Seinfeld]

I’m still trying to wrap my head around the Net Efficiency portion. I get that you can’t turn these algorithms into calculus, but if you’re only going to use a handful of stats, why in the world would you base the entire formula on FG attempts, FT attempts, and offensive rebounds?

 

[VegasDawg13]

I agree. I think people got confused when this first came out. You don't get bonus points for winning by 10. 10 is just the maximum number for margin of victory...so if you win by 20, it only counts as winning by 10. At least, that's the way I understand it.

This is true only for one of the factors used. Another factor is point differential (I assume adjusted for opponent), so every point counts a little bit. It's just that they count a little more up to +10.

 

[Mayor of Little London]

Hurts us this year but what if it came along in time for last years season, we would have made the tournament for sure.

 

[VegasDawg13]

I’m still trying to wrap my head around the Net Efficiency portion. I get that you can’t turn these algorithms into calculus, but if you’re only going to use a handful of stats, why in the world would you base the entire formula on FG attempts, FT attempts, and offensive rebounds?

Efficiency is essentially points per possession.

However, "Possessions" isn't an official stat. So any efficiency rating has to define "Possessions" using other stats. Any FG attempt is a possession unless you get your own rebound (+ FGA - Offensive Rebounds). A turnover represents a possession (+ TO). Lastly, getting fouled and going to the line represents a possession. However, because some FT's are part of an and-1, and sometimes you get 3 FT's on a single possession, they determined that 2 FT attempts are the equivalent of .95 FG Attempts (+ (.475*FTA)).

So, the end result is:

Offensive Efficiency = Points / (FGA + TO + (.475 * FTA))
OR
Offensive Efficiency = Points / Possession

Do the opposite for defense, and you have your Defensive Efficiency. Subtract your defensive rating from your offensive rating, and you have you Net Efficiency.

While this may seem needlessly complicated, it is a good way to determine how good a team's offense or defense is while controlling for pace. What's stupid about the NCAA's version is that it doesn't account for opponent.

 

[DAWG61]

Before the game yesterday, we were one spot back of Auburn in the NET rankings and 27th overall. Now after beating them, we are two spots back of them and 29th overall. I think this formula might be even more 17ed than RPI. The good news is we are still well on the good side of the bubble, but who knows what the remaining games will yield or what they will do to our ranking.

You're looking at the rankings backwards. This week we are 27th and Auburn is 26th. Last week we were 29th and Auburn was 27th so we moved up 2 spots but strangely Auburn moved up one spot despite losing 3 games in a row. Wofford is 31st and Cincinnati 23rd so we have 3 wins over top 31 teams. We are probably something like a 6/7th seed right now.

 

[Go Budaw]

You're looking at the rankings backwards. This week we are 27th and Auburn is 26th. Last week we were 29th and Auburn was 27th so we moved up 2 spots but strangely Auburn moved up one spot despite losing 3 games in a row. Wofford is 31st and Cincinnati 23rd so we have 3 wins over top 31 teams. We are probably something like a 6/7th seed right now.

I looked at it before Auburn and it said “through games played on 1/25/18”, then the next day it said “through games played 1/26/18” and there was more separation. Now evidently it has changed again after only a sparse number of games. No matter how you slice it, Auburn not being penalized at all for losing (the opposite, in fact) is complete horse ****. They moved up after losing to a lower ranked team. The NET rankings are so inferior to the RPI that its ridiculous.

 

[DawgInTheCold]

Net rankings seem to be fairly efficient, well, more efficient than rpi