RPI rank and likelihood of at-large NCAA bid

[OrlandoCuse]

Our RPI ranking has been fairly volatile lately with each win or loss. We were hovering around 40 after the win streak, but we plummeted to the mid 60's after losses to UL and Pitt. We're currently at 56 after yesterday's win.

I was curious about how RPI ranks of P5 teams correlate to at-large bid likelihood (probability). I analyzed the RPI ratings since 2011 (when the field expanded to 68 teams). I isolated only P5 conferences (included the Big East up to 2013, dropped thereafter). P5 teams on probation were dropped as well. Total number of P5 teams receiving at-large bids vs total number of P5 teams eligible for at-large bid.

Here's the results:
RPI 41-50: 23 eligible P5 teams, 21 received at-large bids = 91% likelihood
RPI 51-60: 21 eligible P5 teams, 13 received at-large bids = 62% likelihood
RPI 61-70: 17 eligible P5 teams, 4 received at-large bids = 24% likelihood
RPI 71-80: 15 eligible P5 teams, 0 received bid = 0% likelihood

So we're currently squarely on the bubble, in the 60% likelihood range. If we want to feel comfortable about our chances we need to move our RPI into the 40's. A win at UNC would be huge (loss at the #9 RPI team shouldn't hurt us too much), but the game at FSU is now super important.

If we head into Selection Sunday with an RPI above 60 I don't like our chances.

 

[HoustonCuse]

Great analysis !

 

[Chris02M]

despite our rpi we are in a good spot to get in with no bad losses

 

[Dave85]

Our RPI ranking has been fairly volatile lately with each win or loss. We were hovering around 40 after the win streak, but we plummeted to the mid 60's after losses to UL and Pitt. We're currently at 56 after yesterday's win.

I was curious about how RPI ranks of P5 teams correlate to at-large bid likelihood (probability). I analyzed the RPI ratings since 2011 (when the field expanded to 68 teams). I isolated only P5 conferences (included the Big East up to 2013, dropped thereafter). P5 teams on probation were dropped as well. Total number of P5 teams receiving at-large bids vs total number of P5 teams eligible for at-large bid.

Here's the results:
RPI 41-50: 23 eligible P5 teams, 21 received at-large bids = 91% likelihood
RPI 51-60: 21 eligible P5 teams, 13 received at-large bids = 62% likelihood
RPI 61-70: 17 eligible P5 teams, 4 received at-large bids = 24% likelihood
RPI 71-80: 15 eligible P5 teams, 0 received bid = 0% likelihood

So we're currently squarely on the bubble, in the 60% likelihood range. If we want to feel comfortable about our chances we need to move our RPI into the 40's. A win at UNC would be huge (loss at the #9 RPI team shouldn't hurt us too much), but the game at FSU is now super important.

If we head into Selection Sunday with an RPI above 60 I don't like our chances.

Really great post! Thanks!

 

[stlorange]

"despite our rpi we are in a good spot to get in with no bad losses" ... uh... St. Johns was beyond bad loss...

 

[SgtCuse]

despite our rpi we are in a good spot to get in with no bad losses

St John's wasnt a bad loss...it was atrocious.

 

[NineOneSeven]

despite our rpi we are in a good spot to get in with no bad losses

We have one of the worst losses in history this year

 

[jncuse]

Thanks. Very good research. Confirms my obserbed view (which I had never validated) as I have tracked our projected RPI this year. I appreciate it for my future analysis.

I think somebody could misinterpret your results to suggest that the RPI number in itself drives selection which is not really accurate. The RPI is a "look at my metrics" number. It's hard for a P5 school to have a sub 50 RPI and not have some really good key metrics. That is probably the bigger reason it's nearly 100% at that point.

If you are in the 50's or close to it like we are this year, you will have some warts and some good. And that is how they select those teams - that is where our top 50 wins help us.

I just wanted to end up in the 40-60 range, and it seemed based on forecast our RPI at 9-9 would be right around the late 50's. It will end up there or low 60's. My view this year was this.

RPI of 40-60. You will get a full resume search, you have good chances to get in. The good will not be missed (and we have some good top 50 wins)
RPI 60+ --You probably have a few warts that are causing that, and they start to gain importance.
RPI 70+ - They will usually find something smelly to eliminate you right away.

Here is where our RPI projects to if we go 1-1 the next game. It does not really matter if its UNC or FSU as it's a road game against somebody we already played. If we go 1-1 the next 2 games, we will be sub 50 entering tourney week, and I can't see how we will be on the teams on the bubble line entering tourney week -- we will have to much good on our resume for a team with that RPI.

If we lose those 2 games we will be right around 60 heading into tourney week. Although there will be a few more teams with 60's getting in this year probably due to SMU and Louisville creating spots... it also depends on bubble buster.

As of now I don't see a P5 team with a projected RPI under 60 that is not in.

 

[dasher]

despite our rpi we are in a good spot to get in with no bad losses

St john's was a good loss?

 

[pearl31]

quick thoughts...
(as jn states above) RPI is an indicator and a sort of road map to roughly group teams - it's not the all-encompassing tool to seed teams that many make it out to be...
There are teams in the top 40 rpi right now that will not get in unless it's their conf auto bid: Princeton, Ark. Little Rock, Bonnies, Tulsa (a couple of these could possibly slip in, but unlikely)
I predict there will be at least 3-4 teams with 60+ rpi ranking that get an at-large bid - teams like Gonzaga, Fla St, Washington, San Diego St, for example, all have just as good or better chance than the top-40 teams listed...
finally, lots of games still to be played including CONFERENCE TOURNAMENTS

 

[jncuse]

While St. John's was beyond bad, the good for us is that we have limited that number to just 1. We have the "quality" but other teams have more quantity which I think is important.

A few teams have no bad losses which is really helping them (Uconn and Gonzaga) in terms of their current status.

 

[pearl31]

"despite our rpi we are in a good spot to get in with no bad losses" ... uh... St. Johns was beyond bad loss...

he menas from here out

 

[jncuse]

quick thoughts...
(as jn states above) RPI is an indicator and a sort of road map to roughly group teams - it's not the all-encompassing tool to seed teams that many make it out to be...
There are teams in the top 40 rpi right now that will not get in unless it's their conf auto bid: Princeton, Ark. Little Rock, Bonnies, Tulsa (a couple of these could possibly slip in, but unlikely)
I predict there will be at least 3-4 teams with 60+ rpi ranking that get an at-large bid - teams like Gonzaga, Fla St, Washington, San Diego St, for example, all have just as good or better chance than the top-40 teams listed...
finally, lots of games still to be played including CONFERENCE TOURNAMENTS

I agree with everything you said Pearl. RPI is not really that relevant for the lower tier mid-majors.

The OP however isolated his analysis to P5 schools only and limited his conclusions to them. The % would be higher for the better mid-majors like the AAC, A-10. MWC (in some recent years) but nowhere near as high as the P5. The P5 schools just give you the opportunity to get more good on the resume.

 

[Brooky03]

Not to beat a dead horse, but the Boeheim suspension should help us, too. Might bump us up a tier.

 

[Chris02M]

We have one of the worst losses in history this year

i mean no new bad losses

 

[jdubs30]

While St. John's was beyond bad, the good for us is that we have limited that number to just 1. We have the "quality" but other teams have more quantity which I think is important.

A few teams have no bad losses which is really helping them (Uconn and Gonzaga) in terms of their current status.

I've said this before as well. It's such a bad loss it's almost easy to ignore.

We have 3 losses against teams with an RPI worse than 48, they are -- GTown (101), Clemson (114) and St. John's (225). How ommittee consider that our 3 worst losses all came within a month of each other when our coach wasn't on the sidelines?

 

[Knicks411]

I've said this before as well. It's such a bad loss it's almost easy to ignore.

We have 3 losses against teams with an RPI worse than 48, they are -- GTown (101), Clemson (114) and St. John's (225). How ommittee consider that our 3 worst losses all came within a month of each other when our coach wasn't on the sidelines?

It's amazing to think that if you add up the RPI of their 2nd and third worst losses it's still lower than st. john's RPI.

 

[Chris02M]

for me i only consider it a bad loss if there below 125

 

[jncuse]

I've said this before as well. It's such a bad loss it's almost easy to ignore.

We have 3 losses against teams with an RPI worse than 48, they are -- GTown (101), Clemson (114) and St. John's (225). How ommittee consider that our 3 worst losses all came within a month of each other when our coach wasn't on the sidelines?

I didn't think we had any other sub 100 RPI losses, but I guess 2 have fell back into that area. Georgetown started to suck again and Clemson has stagnated, to cause 2 more sub 100 losses. The lost to Clemson at home could be deemed a bad loss

I'm not saying that 100% the coach situation will not be adjusted, but my opinion is that it will not be. And I don't think it will matter in the end.

 

[jncuse]

for me i only consider it a bad loss if there below 125

The official number for the statistic is #100 or less. But just like not all top 50 wins are the same, not all bad losses are the same. The closer you are to the bubble line the more carefully they will dissect each game -- I don't think a road loss against a team a bit over 100 would be used against us. Not as convinced about Clemson since it was at home.

 

[Knicks411]

I didn't think we had any other sub 100 RPI losses, but I guess 2 have fell back into that area. Georgetown started to suck again and Clemson has stagnated, to cause 2 more sub 100 losses. The lost to Clemson at home could be deemed a bad loss

I'm not saying that 100% the coach situation will not be adjusted, but my opinion is that it will not be. And I don't think it will matter in the end.

My guess is if it comes down to us and another team at the end to make the field, and they believe we have roughly the same resume, the coaching thing might be a tiebreaker.

 

[jncuse]

Not to beat a dead horse, but the Boeheim suspension should help us, too. Might bump us up a tier.

It probably won't if we are on the bubble line.

1) While considered, I never thought it would be adjusted based on history. But opinion, either way, probably won't matter because of #2 below

2) If we end up on the bubble discussion we end the season going 1-4, 2-4 or 2-5. It will look more like this team is just inconsistent rather then anything else.

 

[pearl31]

Our RPI ranking has been fairly volatile lately with each win or loss. We were hovering around 40 after the win streak, but we plummeted to the mid 60's after losses to UL and Pitt. We're currently at 56 after yesterday's win.

I was curious about how RPI ranks of P5 teams correlate to at-large bid likelihood (probability). I analyzed the RPI ratings since 2011 (when the field expanded to 68 teams). I isolated only P5 conferences (included the Big East up to 2013, dropped thereafter). P5 teams on probation were dropped as well. Total number of P5 teams receiving at-large bids vs total number of P5 teams eligible for at-large bid.

Here's the results:
RPI 41-50: 23 eligible P5 teams, 21 received at-large bids = 91% likelihood
RPI 51-60: 21 eligible P5 teams, 13 received at-large bids = 62% likelihood
RPI 61-70: 17 eligible P5 teams, 4 received at-large bids = 24% likelihood
RPI 71-80: 15 eligible P5 teams, 0 received bid = 0% likelihood

So we're currently squarely on the bubble, in the 60% likelihood range. If we want to feel comfortable about our chances we need to move our RPI into the 40's. A win at UNC would be huge (loss at the #9 RPI team shouldn't hurt us too much), but the game at FSU is now super important.

If we head into Selection Sunday with an RPI above 60 I don't like our chances.

Those are pretty revelatory #s, but...
I'd like to see how many of those RPI 51-70 teams that were left out also had 5 top 50 wins and 3 top 25 wins - I'm guessing not one of them did... and it would also be good to look closely at top-50 rpi teams from a p5 that were left out (i.e., the 9% of your first line of %s) - I'll bet their resumes just do not cut it whatsoever - in other words, there are teams every year that "game" the rpi but don't have the meat and potatoes to back it up

 

[OrangeDW]

Those are pretty revelatory #s, but...
I'd like to see how many of those RPI 51-70 teams that were left out also had 5 top 50 wins and 3 top 25 wins - I'm guessing not one of them did... and it would also be good to look closely at top-50 rpi teams from a p5 that were left out (i.e., the 9% of your first line of %s) - I'll bet their resumes just do not cut it whatsoever - in other words, there are teams every year that "game" the rpi but don't have the meat and potatoes to back it up

I've been paying attention to Florida all year and they would seem to be a good example of this. They have one good win all year, against WVU. Other than that, they lost to everybody good that they played (Purdue, Miami, Michigan St. Texas A&M, Kentucky). They're 17-12 overall and 8-8 in the mediocre SEC. Yet their RPI is better than ours.

 

[OrlandoCuse]

Those are pretty revelatory #s, but...
I'd like to see how many of those RPI 51-70 teams that were left out also had 5 top 50 wins and 3 top 25 wins - I'm guessing not one of them did... and it would also be good to look closely at top-50 rpi teams from a p5 that were left out (i.e., the 9% of your first line of %s) - I'll bet their resumes just do not cut it whatsoever - in other words, there are teams every year that "game" the rpi but don't have the meat and potatoes to back it up

No doubt that quality wins (or lack thereof) is the delta between making it in and not when teams are between the 40-70 RPI range.