Teams
Forum List
Toggle sidebar

Cyclones RPI

I don't think the RPI is going away. It's a simple clean formula that needs lots of games to be played to have accuracy. It is never used by itself as a number. It simply provides a clean baseline for evaluating schedules. As mentioned the RPI is usually used along with other metrics to bracket schedules at the end of the season. Nothing more than that. The CBS team RPI comparison tool is an example of that use. Teams can have good RPI's but can never completely hide from their schedules when bracketing (see Syracuse). RPI shouldn't be the only tool used and it's a good thing that the committee is formally including other metrics to evaluate teams.

At the end of conference tournaments, most schedule metrics and power ratings (which RPI is not) indicate similar evaluations for strong tourney teams. When seeding and examining fringe teams, all tools should be used to see where conflicts arise between metrics and identify where schedules may differ according to what the committee prioritizes as benchmarks.
 
  • Like
  • Agree
  • Useful
Reactions: cyputz, cyclones500 and Cyclonepride
EarthIsMan said:
Cyclones500, the NCAA selection committee uses a variety of metrics and criteria, but it does appear that the RPI use among committee members is waning. Also, the NCAA is not interested in rebuilding their own metric, but wanting to use other existing metrics. Statistically, RPI is just very poor and does not have logic.

For example, it is possible to have a team win against a very lowly ranked RPI team, and the winning team's RPI gets worse. Conversely, it is possible for a team to lose to very highly ranked RPI team, and the losing team's RPI improves.

Using solely Massey, Sagarin, KenPom, etc. have their problems too, but they do have much more statistical rigor and logic.

CyTwins, My thought's are not original, but mainly shaped from this article:
View attachment 52062
Click to expand...

That's also the beauty of RPI though. Look at it this way, if The committee didn't use RPI the Big 12 is probably capped as a 5 bid league. As a fan of the Big 12 you do NOT want to see the NCAA stop using RPI.
 
  • Agree
Reactions: isutrevman and cyputz
BillBrasky4Cy said:
That's also the beauty of RPI though. Look at it this way, if The committee didn't use RPI the Big 12 is probably capped as a 5 bid league. As a fan of the Big 12 you do NOT want to see the NCAA stop using RPI.
Click to expand...

Simply, I am a fan of having the best teams play in the tournament.....and ISU.

While I would not be an advocate of exactly this either, if the 68 top KenPom were chosen today for the NCAA field all 10 B12 would make the tournament.
 
  • Like
Reactions: Doc
EarthIsMan said:
Simply, I am a fan of having the best teams play in the tournament.....and ISU.

While I would not be an advocate of exactly this either, if the 68 top KenPom were chosen today for the NCAA field all 10 B12 would make the tournament.
Click to expand...

Don't forget, though, you have to factor in the 32 auto-bids, too, so all 68 wouldn't be at-large. There is some overlap for larger leagues (top teams would get in regardless of automatic), but typically there are still 22 or so 1-bid leagues. That leaves 46-ish -- So Big 12 in this case still would get 8 (OSU and ISU missing cutoff).

Sorry, didn't intend to nitpick your post, just discussion/observation.
 
KKramer said:
Well to be fair, take a look at the RPIs of the teams Iowa State lost to last year:

Gonzaga - 6
Cincinnati - 12
Iowa - 83
Kansas - 4
West Virginia - 20
Texas - 156
TCU - 52
Baylor - 9
Vanderbilt - 42

Take out the Texas loss and all of those losses are to tourney teams or bubble teams with good to great RPIs. Iowa State had great wins that amplified it, but that's why they got the 5 seed with 10 losses.
Click to expand...
I was just pointing out to the poster that I responded to that 10+ losses doesn't mean you're automatically sliding in as a bubble team. I don't see ISU getting to 11-12 wins in Big 12 play but if they did they'd absolutely be better than a bubble team.
 
  • Agree
Reactions: CyJeans, isutrevman and EarthIsMan
VeloClone said:
Getting in with 10+ losses is not rare.

That is 10 teams that had better than a 10 seed with 10 or more losses so it isn't rare.
Click to expand...

Yes; thanks for corroborating my post: "..We all can come up with examples of P5 teams with 10+ losses getting in as 10-12 seeds..."
 
randomfan44 said:
I was just pointing out to the poster that I responded to that 10+ losses doesn't mean you're automatically sliding in as a bubble team. I don't see ISU getting to 11-12 wins in Big 12 play but if they did they'd absolutely be better than a bubble team.
Click to expand...

I agree, if we get 12 Big 10 wins, Prohm is COY
 
Go2Guy said:
Yes; thanks for corroborating my post: "..We all can come up with examples of P5 teams with 10+ losses getting in as 10-12 seeds..."
Click to expand...

I refuted not corroborated. You missed this part of my post apparently:

"That is 10 teams that had better than a 10 seed with 10 or more losses so it isn't rare."

And you apparently also missed the 10 teams from last year (more than 1/7 of the field last year) I listed that had 10 or more losses and had a seed better than the 10-12 seed you put as your threshold. Those numbers in front of the teams are their seeds so the numbers from 5 to 9 are not teams "getting in as 10-12 seeds...."
 
  • Funny
Reactions: Gunnerclone
EarthIsMan said:
Simply, I am a fan of having the best teams play in the tournament.....and ISU.

While I would not be an advocate of exactly this either, if the 68 top KenPom were chosen today for the NCAA field all 10 B12 would make the tournament.
Click to expand...

I agree in principal. However, KenPom literally doesn't care who wins. Winning matters, and when looking at who should make the tourney I think it should be who has the best record against the best schedule. RPI is better for that part IMO.
 
  • Disagree
Reactions: FinalFourCy
BryceC said:
I agree in principal. However, KenPom literally doesn't care who wins. Winning matters, and when looking at who should make the tourney I think it should be who has the best record against the best schedule. RPI is better for that part IMO.
Click to expand...
My main problem is that the mathematically the RPI is quite arbitrary.

RPI does quite well with the top 5-10% of 351 teams, but it does pretty terribly at the margins when trying to determine the at-large teams in the field of 68.

The RPI also benefits the teams that are able to construct strong non-conference games and those teams happen to be in strong conferences.

Betting markets, KenPom, Massey, Sagarin, etc. are all better prognosticators of the best teams and who ultimately wins in the end.
 
  • Agree
Reactions: rholtgraves
EarthIsMan said:
My main problem is that the mathematically the RPI is quite arbitrary.

RPI does quite well with the top 5-10% of 351 teams, but it does pretty terribly at the margins when trying to determine the at-large teams in the field of 68.

The RPI also benefits the teams that are able to construct strong non-conference games and those teams happen to be in strong conferences.

Betting markets, KenPom, Massey, Sagarin, etc. are all better prognosticators of the best teams and who ultimately wins in the end.
Click to expand...

I agree. But if we're talking about a ranking, and resumes... I think KenPom is problematic. Again, if we're just looking at resumes, I'd like KPI the most, look into that one.

What would KenPom rank a team that was 0-30, but played all road games against the top 30 teams and lost each one by 1? Probably incredibly high.
 
  • Like
Reactions: EarthIsMan
If SOS were actually a measure of opponent strength rather than opponents' W/L record (which is only loosely related, unless you want to argue all conferences are similar strength...), RPI wouldn't be so bad.
 
BryceC said:
I agree. But if we're talking about a ranking, and resumes... I think KenPom is problematic. Again, if we're just looking at resumes, I'd like KPI the most, look into that one.

What would KenPom rank a team that was 0-30, but played all road games against the top 30 teams and lost each one by 1? Probably incredibly high.
Click to expand...
Agreed, KPI is superior to RPI and a great combination of metrics. It is also clear the individual who schedules games for Michigan St. keeps the KPI ranking in mind ;)
 
Sigmapolis said:
View attachment 52139
Click to expand...

dumb-and-dumber-theres-a-chance-gif.gif
 
  • Agree
Reactions: cyhiphopp
Heck even 8-10 (lets argue we beat a ranked team in those 8 wins) plus a win in KC might get you to the play-in game.
 
You must log in or register to reply here.

Help Support Us

Become a patron
Top Bottom