RIP RPI - New NCAA Basketball Ranking System
- Thread starter MadRU
- Start date
Dumb - this is the “ESPN” ranking now, based on made-up analytics that can’t be controlled. Do you think a player during the game thinks “hmm, this is a low-percentage shot, I better not take it so that it doesn’t decrease our defensive efficiency rating” blah blah. RPI is what it should be about, WIN and who did you WIN against and who did they WIN against. It’s sports!! And margin of victory is incredibly dumb to use in basketball where the final scores of close games (within 10 like they say) are so much dependent on free throws at the end and not indicative of the entire game. Dumb
Never was a huge fan of the RPI so don't mind trying something new. Kinda like margin of victory. I think it does matter and a decent barometer.
2nd best team in the nation last year in scoring margin: Villanova.
3rd best team in the nation last year in scoring margin: Gonzaga
2nd best team in the nation last year in scoring margin: Villanova.
3rd best team in the nation last year in scoring margin: Gonzaga
Last edited:
RPI is absolutely terrible. It rewards teams for beating bad opponents. Things like KenPom reward teams for actually being good at basketball.
Margin of victory is dumb to use? Really? As if losing a game on the road to the #1 team in the country by 1 point should just go down as an L and be a worse performance than beating a terrible team by 10 points at home?
Using RPI leads to the committee making terrible decisions. If they gave more weight to things like KenPom instead they would make better decisions. I also like how you claim this has anything to do with ESPN and "made up analytics that can't be controlled" as if being better than your opponent wasn't under your own team's control. And I'd hope a player coming down the court isn't worried about a bad shot impacting their team's efficiency margin, I'd hope he is worried about it impacting their chances of winning the game. Because that's what those efficiency margins actually predict, real wins and losses.
The RPI is like statistics for dummies. It's sounds smart and mathy but is really just pointless. It'd be akin to just filling the NCAA tournament with whatever schools just have the most wins in the regular season no matter the conference or schedule difficulty. I mean sure the RPI incorporates schedule strength, it just bases it on it's own measure which is by definition severely flawed.
Basketball simply comes down to trying to score when you have the ball and stopping your opponent from scoring when they have the ball. The better you are at those 2 simple things, the higher you will rank in efficiency metrics. It's quite simple and controllable. Also conveniently it predicts future success in wins and losses and nearly perfectly follows the Vegas line.
The only reason they got in is because the powers that be thought they could catch Steph Curry lightning in a Trae Young bottle.
All this new ratings system will do is put mediocre teams like this in the tournament for even stupider statistical reasons.
There is nothing wrong with the RPI if it is used as a tool and not the be-all and end-all of basketball. And in recent years I think we've moved closer to that.
I don't see any reason to ditch anything, and I see reasons to add more data. After all, if one team is dramatically better or worse in one of the rankings, that team might get a closer look and thus a more accurate evaluation.
I don't see any reason to ditch anything, and I see reasons to add more data. After all, if one team is dramatically better or worse in one of the rankings, that team might get a closer look and thus a more accurate evaluation.
Keep scoring margin out.
Capping at 10 is an attempt of keeping teams from running up score, but there isn’t much of a difference between a 2 point and 10 point victory. Take a 8 point margin and you have no idea if.......
1. Game was close and losing team fouled at end
2. Game went in to overtime
3. Game was 15 point margin near end and then losing teams 1st string made a run at the winning team 2nd string
The kenpom rankings I love only use margin of victory. It is great for fans to look at, but not be used for post season purposes. It is easily manipulated if it matters
Capping at 10 is an attempt of keeping teams from running up score, but there isn’t much of a difference between a 2 point and 10 point victory. Take a 8 point margin and you have no idea if.......
1. Game was close and losing team fouled at end
2. Game went in to overtime
3. Game was 15 point margin near end and then losing teams 1st string made a run at the winning team 2nd string
The kenpom rankings I love only use margin of victory. It is great for fans to look at, but not be used for post season purposes. It is easily manipulated if it matters
The real solution is to keep the RPI and peak at kenpom, but NEVER admit looking at it.
RPI also encourages teams to play on the road which is good.
RPI also encourages teams to play on the road which is good.
Scoring margin should be in. Say no to Caps! Firmly believe strength of schedule and margin of victory are the two most important factors in judging teams. Scoring margins are still averages like every other stat. Some games will be a blow out, some go to OT, some will be first team v. second although often second v second, some will be the results of poor foul shooting. It's still an average over 30 games. The best argument in favor of keep scoring average is last year. Who were the National Champs? Villanova. Who was 2nd in the entire country in scoring margin? Villanova. Case closed.
I think scoring margin should be a part, but I don't know necessarily how I'd do it.
- I think being tied at the end of regulation should count for something, even if you get blown out in OT
- I think that a loss margin of 3 or less should count for more than a loss by 4 or more, since it denotes that the game came down to the final shot
- I think that a single digit loss should count for more than a double digit loss, since it usually denotes a game that was fairly competitive (even if it came down to FTs over the final minutes and the lead ballooned to 8 or 9)
- I think that 20+ and 30+ point losses should count more against a team
- I think being tied at the end of regulation should count for something, even if you get blown out in OT
- I think that a loss margin of 3 or less should count for more than a loss by 4 or more, since it denotes that the game came down to the final shot
- I think that a single digit loss should count for more than a double digit loss, since it usually denotes a game that was fairly competitive (even if it came down to FTs over the final minutes and the lead ballooned to 8 or 9)
- I think that 20+ and 30+ point losses should count more against a team
scoring margins say everything...until they matter.
uncapped scoring margins would kill the integrity of the sport. cupcakes would be scheduled to win by 60
Perhaps. But 10 points is ridiculously low to know how good one team or how poor another is. Renders scoring margin for every good team a meaningless stat. Shouldn't there then be a cap for Offense efficiency against bad teams. Makes no sense so low.
The scoring margin is used in the context of the opponent. For example beating the #5 team on the road by 14 points is more impressive than beating the #275 team at home by 31 points.
The issue is they won by 10 and lost by ten in those 2 games. If you don't have a cap there is an incentive to run up the score. With a cap you hurt your ranking the minute you schedule the game.
In the end shouldn't everything be about whether you lost or won the game. When in any sport is it about putting the best teams in the playoffs. It is only about a W or L. Obviously this is different because unbalanced schedule and we need to identify how quality was the schedule. the simple RPI of 25% winning percentage 50% opponent's winning percentage 25% opponents opponents winning percentage + an adjustment for home/road is actually about as good as you are going to get.
And add scoring margins. Some years you may just be lucky. Team that dominates without running it up should be rewarded, like for instance, Villanova 2018. Ten points is a ridiculously low cap. Ten point game nobody is usually taking out starters and hardly running it up.
Make it like hockey. Both teams get a point for going to overtime, winner of overtime gets an additional point. A win in regulation is worth 2 points.
You put scoring margin as a criteria Rutgers plays at Nebraska this year there is a number Nebraska feels they have to beat RU by to hold serve. Nebraska isn’t just trying to win the game. They know they have to win by 15 to get ahead.
It's still an average scoring margin. If you are going to use scoring margin as a stat it makes no sense to cap it at 10. I am guessing for all the Top 40 teams that the cap numbers will be so similar that the stat will be rendered meaningless. Is Offense efficiency capped? Why not?
offensive efficiency isn't looked at by a committee as criteria for making a tournament.
I'd bet if one really wanted to take kenpom data and standardize it you could remove data 2 to 2.5 standard deviations above the mean and get a more accurate measure.
the bottom line is kenpom and/or scoring margin should never be knowingly used as a criteria for postseason. The integrity of the game would be lost.
The scoring margin cap is for each individual game, not a season long average being capped.
It might work to have scoring margin a factor in analyzing the strength of schedule. That way scoring margin can be used as a tool without teams having an incentive to manipulate it.
It might help if they factored in when team A played team B also. Team B might be highly rated, then an injury or dismissal to a player and all hey are no longer highly rated. Team A should get credit for beating a tougher version of team B.
Last Years Minnesota team comes to mind.
Last Years Minnesota team comes to mind.
I would think all the RPI, or whatever tool is used, is to help create a list of teams to evaluate and pit against each other
I highly doubt one team is selected over another either for inclusion in the tournament or seed based on RPI.
Probably more application in assessing their SOS. A blanketed how they did vs RPI 1-50
I highly doubt one team is selected over another either for inclusion in the tournament or seed based on RPI.
Probably more application in assessing their SOS. A blanketed how they did vs RPI 1-50
yes, they cap the margin in any game at 10 points. So if Villanova beats a team by 50 and DePaul beats the same team by 10, it would consider those 2 performances as equal.
That's a slight improvement on RPI, but not a ton.
I do like that...one possession win/loss stat.
They won't all be representative, but shows your mettle in close games
I hate it. Tells you nothing.
Every sport I see has a binary result. You either win the game or lose a game. Taking care of business shouldn't meaning winning a game by 10.
If the NCAAs were really concerned about getting the best at large teams in you would see teams in with losing records. Get credit for wins ONLY.
Hoston Astros happy. they are now 14 games, not 2, over the Oakland A's. Based on pythagorem standings Houston can coast providing they dont get blown out in too many games
http://www.baseball-reference.com/leagues/MLB-standings.shtml
http://www.baseball-reference.com/leagues/MLB-standings.shtml
I favor a blending of pure efficiency metrics (no scoring margin caps) with Strength of Record (who did you beat).
While a game has a "binary result", it isn't advanced math to understand a team winning by 50 points and playing their walkons for 10 minutes is more impressive than a team winning by 1 point on a last second desperation shot against the same opponent.
Too much luck involved in whether an individual game was a win or a loss to try to make anything too meaningful of it. I mean if your 80% FT shooting opponent decides to go 2/10 on FTs in the game and you win by 1 point, is that all credit to you or just a little bad luck on the FTs for your opponent? I mean between shots rimming in or out, refs making ticky tack calls, and other random little things that happen you don't get nearly as much meaningful information out of W or L as you'd like.
If all you go by is wins and losses, you end up letting in some teams that aren't very good and just happened to play a crappy schedule. And what does that incentivize? Teams scheduling the weakest opponents they can find. Fans benefit when the best teams play each other OOC and we should reward that.
Exactly! I don't get it, why even include it then?
