NCAA Basketball Drops RPI for Tournament Selection

[Still Eeyore]

http://www.espn.com/mens-college-bas...ing-system-rpi

I have no idea what the math of the new system looks like, though the NCAA's history doesn't make me very confident that it will be very good. Still, will be interesting to see whether this is a start of moving on from an RPI based system (which PWR is) in all sports.

[DDad16]

Re: NCAA Basketball Drops RPI for Tournament Selection

Too bad for the Engineers! I really thought they had a shot this year!

[TonyTheTiger20]

Re: NCAA Basketball Drops RPI for Tournament Selection

Oh that's good

[robertearle]

Re: NCAA Basketball Drops RPI for Tournament Selection

http://www.espn.com/mens-college-bas...ing-system-rpi

I have no idea what the math of the new system looks like, though the NCAA's history doesn't make me very confident that it will be very good. Still, will be interesting to see whether this is a start of moving on from an RPI based system (which PWR is) in all sports.

The short explanation (maybe too short) of the rankings they are going to use - Ken Pomeroy - is a calculation of offensive and defensive efficiency.

A Basketball game is a series of alternating ball possessions; team A has the ball, then team B gets the ball, then team A, etc etc. In an average game, each team might get 70 possessions over the course of the game. If team A averages one point per possession, and team B averages only 0.9 points per possession, team A wins 70-63.

Some teams like to play 'fastbreak' basketball, that over the course of a season might end up raising their possessions per game to 80 or 90 or 100. Other teams like to play slow, deliberate 'half court' offense that gets them to 50 possessions per game. If each averages one point per possession, the fastbreak team is averaging 90 points a game, while the slow team is averaging only 50 points a game.

But what if the slow 'half court team' is more efficient, averaging 1.1 points per possession, 55 points a game. And what if the fast team played the slower team? At a glance, one team is averaging 90 points a game, and the other only 55 points a game. Fast team is gonna win, no?

No, probably not. Because the game is going to be maybe a 70 possession game, faster than 'slow' would like, but slower than 'fast' would like. And assuming their point per possession efficiency stays the same, slow is going to score 77 with their 70 possessions, while fast is only going to score 70.

Same idea for defense: points per possession given up.

The Pomeroy ranking 'normalizes' each team's offensive and defensive 'per possession' efficiency, and then uses those numbers in a predictive calculation to rank the teams overall.

There are a bunch of other 'fudge factors, including home/away/neutral site, conference, others I forget. But the basic idea is this per possession efficiency.

I can't imagine that it would be applicable to hockey in any usable way.

[shelfit]

The short explanation (maybe too short) of the rankings they are going to use - Ken Pomeroy - is a calculation of offensive and defensive efficiency.

A Basketball game is a series of alternating ball possessions; team A has the ball, then team B gets the ball, then team A, etc etc. In an average game, each team might get 70 possessions over the course of the game. If team A averages one point per possession, and team B averages only 0.9 points per possession, team A wins 70-63.

Some teams like to play 'fastbreak' basketball, that over the course of a season might end up raising their possessions per game to 80 or 90 or 100. Other teams like to play slow, deliberate 'half court' offense that gets them to 50 possessions per game. If each averages one point per possession, the fastbreak team is averaging 90 points a game, while the slow team is averaging only 50 points a game.

But what if the slow 'half court team' is more efficient, averaging 1.1 points per possession, 55 points a game. And what if the fast team played the slower team? At a glance, one team is averaging 90 points a game, and the other only 55 points a game. Fast team is gonna win, no?

No, probably not. Because the game is going to be maybe a 70 possession game, faster than 'slow' would like, but slower than 'fast' would like. And assuming their point per possession efficiency stays the same, slow is going to score 77 with their 70 possessions, while fast is only going to score 70.

Same idea for defense: points per possession given up.

The Pomeroy ranking 'normalizes' each team's offensive and defensive 'per possession' efficiency, and then uses those numbers in a predictive calculation to rank the teams overall.

There are a bunch of other 'fudge factors, including home/away/neutral site, conference, others I forget. But the basic idea is this per possession efficiency.

I can't imagine that it would be applicable to hockey in any usable way.

Agreed, unusable for hockey especially given all the turnovers. I can only guess that there are many more changes of possession in hockey than in basketball. It would be nearly impossible to track that stat consistently or accurately enough in one game let alone for an entire season across every conference and at both the D1 and D3 levels.

[Still Eeyore]

Re: NCAA Basketball Drops RPI for Tournament Selection

I'm less interested in the question of whether they would try to switch hockey to this specific ratings method than I am whether it indicates that the NCAA is recognizing that they went down a mathematically nonsensical path and might start trying to figure out a more sensible approach for each sport. As I said, I'm skeptical that this is what it means, but it at least seems more likely than it did 24 hours ago.

[Timothy A]

Re: NCAA Basketball Drops RPI for Tournament Selection

I'm less interested in the question of whether they would try to switch hockey to this specific ratings method than I am whether it indicates that the NCAA is recognizing that they went down a mathematically nonsensical path and might start trying to figure out a more sensible approach for each sport. As I said, I'm skeptical that this is what it means, but it at least seems more likely than it did 24 hours ago.

Them going to a committee in FB was probably the first step away from math. The first step in women's hockey will be the ncaa using the USCHO poster's poll to count for half of the seeding weight for the ncaa tourney.

[D2D]

Re: NCAA Basketball Drops RPI for Tournament Selection

The first step in women's hockey will be the ncaa using the USCHO poster's poll to count for half of the seeding weight for the ncaa tourney.

If that should come to be I might have to make a few adjustments as to how I go about rating my favorite team. And other contending teams as well.

[TonyTheTiger20]

Re: NCAA Basketball Drops RPI for Tournament Selection

Wow if they're using Kenpom maybe they'd actually consider KRACH in the future. A man can dream!

[Timothy A]

Re: NCAA Basketball Drops RPI for Tournament Selection

If that should come to be I might have to make a few adjustments as to how I go about rating my favorite team. And other contending teams as well.

So you are saying your current rating system is WAY too nonobjective...I will remember that when the polls come out.

[FiveHoleFrenzy]

Re: NCAA Basketball Drops RPI for Tournament Selection

I will remember that when the polls come out.

Who are you kidding?...You can barely remember where you put your keys and wallet from last night.

[ARM]

Re: NCAA Basketball Drops RPI for Tournament Selection

You can barely remember where you put your keys and wallet from last night.

Keys? They let him drive???

[FiveHoleFrenzy]

Re: NCAA Basketball Drops RPI for Tournament Selection

Keys? They let him drive???

Wisconsin...

https://carbreathalyzerhelp.com/wisc...i-convictions/

But maybe it's getting better...

https://www.grievelaw.com/WisconsinOWI/LawChange

[Timothy A]

Re: NCAA Basketball Drops RPI for Tournament Selection

Who are you kidding?...You can barely remember where you put your keys and wallet from last night.

I leave the wallet and the FOB in the car, duh! Now the cell phone I lose all the time.....

[FiveHoleFrenzy]

Re: NCAA Basketball Drops RPI for Tournament Selection

Now the cell phone I lose all the time.....

That's just your subconscious telling you, you've got enough health risk with the enlarged liver that you don't need the cancer risk...