PDA

View Full Version : RPI System is done



ZagNative
08-22-2018, 07:33 AM
From NCAA.com: (https://www.ncaa.com/news/basketball-men/article/2018-08-22/division-i-mens-basketball-committee-adopts-new-ranking)


The NCAA has developed a new ranking system to replace the RPI as the primary sorting tool for evaluating teams during the Division I menís basketball season. The new ranking system was approved in late July after months of consultation with the Division I Menís Basketball Committee, the National Association of Basketball Coaches, top basketball analytics experts and Google Cloud Professional Services.

The NCAA Evaluation Tool, which will be known as the NET, relies on game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency, and the quality of wins and losses. To make sense of team performance data, late-season games (including from the NCAA tournament) were used as test sets to develop a ranking model leveraging machine learning techniques. The model, which used team performance data to predict the outcome of games in test sets, was optimized until it was as accurate as possible. The resulting model is the one that will be used as the NET going forward.

More here. (https://www.ncaa.com/news/basketball-men/article/2018-08-22/division-i-mens-basketball-committee-adopts-new-ranking)

caldwellzag
08-22-2018, 07:34 AM
It's about time!

ZagsObserver
08-22-2018, 07:42 AM
Awesome! It should make the field slightly more diverse and fun.

As a consequence, however, there will be an incentive to blow teams out and play reserves less.

Zagdawg
08-22-2018, 07:49 AM
"The existing quadrant system still will be used on team sheets, with the NET replacing the Rating Percentage Index to sort games based on the opponentís ranking:

Quadrant 1: Home 1-30, Neutral 1-50, Away 1-75
Quadrant 2: Home 31-75, Neutral 51-100, Away 76-135
Quadrant 3: Home 76-160, Neutral 101-200, Away 135-240
Quadrant 4: Home 161-351, Neutral 201-351, Away 241-353"

gonzagafan62
08-22-2018, 07:50 AM
https://www.slipperstillfits.com/2018/8/22/17768858/college-basketball-ncaa-rankings-rpi-dead-net?utm_campaign=slipperstillfits&utm_content=entry&utm_medium=social&utm_source=facebook

gonzagafan62
08-22-2018, 07:51 AM
Please delete my thread lol. My phone didn’t pull this one up. But yes it’s about time

Shows that staying WCC was a good decision.

NotoriousZ
08-22-2018, 07:51 AM
Excellent news! The article also says theyíre keeping the quadrant system but using the NET instead of the RPI to rank the teams in the quadrants.

One nice feature they put into the NET was only going to 10 points when factoring scoring margins for each game played (to keep teams from running up the score on weaker opponents).

Edit: someone beat me to the quadrant news, but somebody else maybe didnít read the article yet about the scoring margin thing.

ZagsObserver
08-22-2018, 08:13 AM
Excellent news! The article also says they’re keeping the quadrant system but using the NET instead of the RPI to rank the teams in the quadrants.

One nice feature they put into the NET was only going to 10 points when factoring scoring margins for each game played (to keep teams from running up the score on weaker opponents).

Edit: someone beat me to the quadrant news, but somebody else maybe didn’t read the article yet about the scoring margin thing.

Ahhh, good to know they put a limit on scoring margin. I think 10 points is too low, however. 15-20 would be more meaningful.

NotoriousZ
08-22-2018, 08:19 AM
Ahhh, good to know they put a limit on scoring margin. I think 10 points is too low, however. 15-20 would be more meaningful.

I was thinking 15 myself, but I like what they did.

Kong-Kool-Aid
08-22-2018, 08:29 AM
It would be interesting to see how our "NET" ranking would have differed from our "RPI" ranking over the past few years.

Kong-Kool-Aid
08-22-2018, 08:45 AM
https://www.ncaa.com/news/basketball-men/article/2018-08-22/college-basketball-new-ranking-system-replacing-rpi

Mark is quoted in this piece. Apparently he was a consultant in these changes.


“Then Mark Few brought up an idea of using a composite of all the various prominent indices, some of which have predictive qualities that could help identify the best teams. After consulting with many of the designers of these other metrics, the NCAA began to develop its own index that would incorporate the most current evaluation measures. With an improved sorting tool, and a tighter definition of a quality win, the hope is we now have a more accurate selection and seeding procedure.”

23dpg
08-22-2018, 08:58 AM
This seems like a much more fair way of picking teams.

Wait, the NCAA got something right? Hmmm.

webspinnre
08-22-2018, 09:26 AM
https://www.ncaa.com/news/basketball-men/article/2018-08-22/college-basketball-new-ranking-system-replacing-rpi

Mark is quoted in this piece. Apparently he was a consultant in these changes.


“Then Mark Few brought up an idea of using a composite of all the various prominent indices, some of which have predictive qualities that could help identify the best teams. After consulting with many of the designers of these other metrics, the NCAA began to develop its own index that would incorporate the most current evaluation measures. With an improved sorting tool, and a tighter definition of a quality win, the hope is we now have a more accurate selection and seeding procedure.”

Well done, coach!

This sounds much more like the kenpom and sagarin ratings, which sounds promising. Certainly more useful than RPI, which was garbage.

Zagnificent
08-22-2018, 09:27 AM
Awesome! It should make the field slightly more diverse and fun.

As a consequence, however, there will be an incentive to blow teams out and play reserves less.

The margin of victory stops mattering at 10 points. There's no incentive to blow teams out.

ZagaholicPodcast
08-22-2018, 09:36 AM
If you're a Power 5/6 program, this can be a positive. This puts somewhat of a freeze on movement, if the movement for the sake of improving March positioning. If you're trying to protect money, you don't want a bunch of mid's trying to get on the life rafts.

bigblahla
08-22-2018, 11:02 AM
IMO, the quadrant system favors power conferences .... I believe if a team can't win 60% of their games in conference doesn't matter what their rating is, there will be more deserving teams, that don't fit the quadrant curve....

Just my opinion...

Go!! Zags!!!

mgadfly
08-22-2018, 11:21 AM
This is good (as the RPI was not a good metric to be relying on), but the Committee has shown a willingness to disregard the RPI when it benefited major conference teams but rely on it to exclude mid-majors. If NET is used in the same way (i.e., completely disregarded when it disagrees with the politics of a majority of the committee), I don't know if this is as helpful as it could be. MAYBE, since it is the NCAA's own metric, it will be given more weight and if SMC is at 42 and Syracuse is at 78, SMC gets in. But I'm going to wait to see how it plays out before I throw a parade.

Kong-Kool-Aid
08-22-2018, 11:28 AM
I'm guessing since this was in large part Fews idea, well it's going to be about as positive for Gonzaga as humanly possible.

soccerdud
08-22-2018, 01:09 PM
I'm guessing since this was in large part Fews idea, well it's going to be about as positive for Gonzaga as humanly possible.

I disagree on one point. I think that capping the margin of victory so low will hurt teams, like Gonzaga, that are significantly better than most or all of their conference mates. One of the big reasons that Gonzaga was ranked number 1 in kenpom a couple years ago, despite playing in the wcc, was that we had a historic margin of victory.

I imagine we will be good enough across other metrics to make up for it, but it certainly will disadvantage standout low- and mid-major teams.

I think a cap nearer 20 would have been better for evaluation purposes.

Robzagnut
08-22-2018, 01:14 PM
I disagree on one point. I think that capping the margin of victory so low will hurt teams, like Gonzaga, that are significantly better than most or all of their conference mates. One of the big reasons that Gonzaga was ranked number 1 in kenpom a couple years ago, despite playing in the wcc, was that we had a historic margin of victory.

I imagine we will be good enough across other metrics to make up for it, but it certainly will disadvantage standout low- and mid-major teams.

I think a cap nearer 20 would have been better for evaluation purposes.

I guess this mean we'll see more of the bench players at the 4:30 mark with a 25 point lead than at 2:00.

hockeyzag
08-22-2018, 03:05 PM
I disagree on one point. I think that capping the margin of victory so low will hurt teams, like Gonzaga, that are significantly better than most or all of their conference mates. One of the big reasons that Gonzaga was ranked number 1 in kenpom a couple years ago, despite playing in the wcc, was that we had a historic margin of victory.

I imagine we will be good enough across other metrics to make up for it, but it certainly will disadvantage standout low- and mid-major teams.

I think a cap nearer 20 would have been better for evaluation purposes.

Completely agree. Acting like a 10 point win is the same as a 20 point or more win doesn't accurately reflect the strength of a dominating team like Gonzaga.

kitzbuel
08-22-2018, 03:38 PM
The margin of victory stops mattering at 10 points. There's no incentive to blow teams out.Interesting to see how that will impact lines at sports books.

Sent from my XT1710-02 using Tapatalk

willandi
08-22-2018, 07:43 PM
IMO, the quadrant system favors power conferences .... I believe if a team can't win 60% of their games in conference doesn't matter what their rating is, there will be more deserving teams, that don't fit the quadrant curve....

Just my opinion...

Go!! Zags!!!

I agree, but probably would have put the bar at 50%.

ZagDad84
08-22-2018, 07:57 PM
IMO, the NET is only the first step.

Unless the committee shows a willingness to put their faith in the NET and use it and the other evaluation metrics to fairly evaluate the Mbb's teams to arrive at the best 68 teams, this NET is only lip service.

The Committee has shown little willingness to try and select the best 68 teams, they are much more likely to select the most wanted 68 teams. One of the larger pundits, Joey Brackets from ESPN, has his doubts until the committee shows otherwise. Here is his opinion of the NET.

ESPN Link: http://www.espn.com/mens-college-basketball/story/_/id/24446023/ncaa-tournament-selection-committee-ditches-rpi-march-madness-evaluation

ZagDad

TexasZagFan
08-22-2018, 08:08 PM
I agree, but probably would have put the bar at 50%.

I would also set the bar at 50%. An additional reform I'd like to see would limit payouts to conferences that place more than 50% of their teams in the Dance. Tournament selection is already skewed to the power conferences, and the new NET won't solve that problem...although it's an improvement.

Since conferences divide tournament shares among all the teams in their respective conferences, cap the "shares" to no more than 50%...i.e. let's say a conference with ten teams gets seven teams into the tournament. Shares would not be awarded to the sixth and seventh teams, that money would be allocated to one bid conferences.

I shed no tears for the power conferences, whose teams routinely play crap schedules OOC, rarely playing good midmajor teams. Those conferences already make tons of money from their TV contracts, and clean up during March.

DixieZag
08-23-2018, 04:53 AM
I disagree on one point. I think that capping the margin of victory so low will hurt teams, like Gonzaga, that are significantly better than most or all of their conference mates. One of the big reasons that Gonzaga was ranked number 1 in kenpom a couple years ago, despite playing in the wcc, was that we had a historic margin of victory.

I imagine we will be good enough across other metrics to make up for it, but it certainly will disadvantage standout low- and mid-major teams.

I think a cap nearer 20 would have been better for evaluation purposes.

Wow, good point.

Yes, 10 points is too low. We've seen games that were essentially 1-2 possession games that after a team starts fouling ends up being an 11 point victory (one of the many reasons I'll never bet on basketball or any other sport). But, 20 points really does define a dominating performance, while also cutting off any gain in running it up to 30 or 40, the really embarrassing numbers.

avid-zag-fan
08-23-2018, 06:03 AM
I think 10 points is perfect. that's 4 possessions.

LongIslandZagFan
08-23-2018, 06:50 AM
Completely agree. Acting like a 10 point win is the same as a 20 point or more win doesn't accurately reflect the strength of a dominating team like Gonzaga.

Just my 2 cents before I slink back into my hole... I have watched many close games that finish as a 10 pt win... really just based on free throws. This isn't even about it being about GU... just basic college hoops. 20 would be a better threshold.

Bogozags
08-23-2018, 08:17 AM
I would also set the bar at 50%. An additional reform I'd like to see would limit payouts to conferences that place more than 50% of their teams in the Dance. Tournament selection is already skewed to the power conferences, and the new NET won't solve that problem...although it's an improvement.

Since conferences divide tournament shares among all the teams in their respective conferences, cap the "shares" to no more than 50%...i.e. let's say a conference with ten teams gets seven teams into the tournament. Shares would not be awarded to the sixth and seventh teams, that money would be allocated to one bid conferences.

I shed no tears for the power conferences, whose teams routinely play crap schedules OOC, rarely playing good midmajor teams. Those conferences already make tons of money from their TV contracts, and clean up during March.

I have felt for years that if conference teams can't win at least 50% of their games, then they should not be in the Dance. Bilas and others feel those teams are better than non-Power 5 schools regardless of their record. I can't recall the name of the school out of NJ but a couple of years ago, they won several games against P5 schools but lost games in conference and didn't win the conference tourney so they were not invited...it is too bad but money talks and nothing else matters...P5's do not want competition coming from non-P5 schools and to restrict their financial opportunities "keeps them back on the farm" so to speak.

Yes this is a tangent but I wonder how many of you were aware that those non-BCS/P5 schools that were invited to the "BCS Bowls" had to divide monies from their appearances with the other non-BCS/P5 conferences...yes IMO, it is rigged...

kitzbuel
08-23-2018, 09:23 AM
I am confused as to how this can be predictive without capturing when games are played.


End-of-season performance has no bearing in the NET'S calculus or in the team sheets.
https://www.cbssports.com/college-basketball/news/the-ncaa-tournament-committee-will-select-and-seed-teams-using-a-new-ranking-system-after-killing-the-rpi/

This does not help teams that develop over the course of the season and IMO eliminates a source of 'eye test' data. I think you need to factor in outlier losses that can occur in conference tournaments, but teams that are trending up should be identified and seeded accordingly.

It does seem to run the risk of becoming a tool to justify Committee selections. It is hard for the NCAA to combat that criticism with a proprietary tool. Transparency would be the best way to deflect criticism, but time will tell.

https://www.washingtonpost.com/news/fancy-stats/wp/2018/08/22/rip-rpi-the-ncaa-has-a-new-ranking-system-for-college-basketball-teams/



Sent from my XT1710-02 using Tapatalk

maynard g krebs
08-23-2018, 10:15 AM
I can't recall the name of the school out of NJ but a couple of years ago, they won several games against P5 schools but lost games in conference and didn't win the conference tourney so they were not invited...it is too bad

Monmouth, with the spectacular 5'8 (or maybe smaller) Justin Robinson, and coached by King Rice, who I always thought must be named for a Creole recipe.

That was really a shame. That team was tough and dynamic and could easily have won a game or two, and Robinson could have been the talk of the first weekend. The committee needs to credit teams like this for their road/neutral wins (UCLA, USC, Georgetown, one other I forget) and realize that most good mid/low major teams will lose a couple of road games in conf.

bballbeachbum
08-23-2018, 04:00 PM
my thoughts follow those of mgadfly and ZagDad84
we'll see how it's implemented, what the attorney speak is as the committee justifies itself in its choices.
jury out, we'll see

JPtheBeasta
08-23-2018, 04:40 PM
A composite ranking system that uses many respected ones makes sense to me. Some sort of Kensagapomrin Power Index plus maybe some tea leaves, entrail readings, and Paul the Prognosticating Octopus from a few years ago would probably do the trick.