Predictive rankings don't like us as much as the record would indicate

[Cusefan95]

Of course it wouldn't predict 80% for any particular game. But over the course of many games, it does predict well. Florida state on the football side had an ordinary record through part of 2022 but their SRS said they were good and they won every game the rest of that year and the next until the qb got hurt

Can you mathematically define what is considered “well” for predictive accuracy, and demonstrate how the various analytics meet that criteria?

Almost inevitably, people point out anecdotal evidence like you just did - which I find incredibly ironic considering the topic.

 

[LAorange44]

we are an inefficient basketball team and that’s what hurts us, in addition to the big losses. We rank 13/15 on offense efficiency in conference.

When efficiency and margin of victory are not considered, like in the colley rankings, we are #26. Hope the committee focuses on the W’s and L’s and not how the game was played. The efficiency rankings really disfavor a young team learning how to play early in a season.

 

[Fly Rodder]

we are an inefficient basketball team and that’s what hurts us, in addition to the big losses. We rank 13/15 on offense efficiency in conference.

When efficiency and margin of victory are not considered, like in the colley rankings, we are #26. Hope the committee focuses on the W’s and L’s and not how the game was played. The efficiency rankings really disfavor a young team learning how to play early in a season.

Still need to beat “good” teams and there are only a few of those left.

 

[DCUNITED13]

As a famous SU Grad

 

[Eric15]

Depends if you use metrics or eye test. But optically I agree with you and if FSU can keep playing well it should take care of itself either number wise or with the committee (if we are actually there in the end)

JN, you’re smarter than me at this stuff. Why is there such a massive discrepancy between our RPI and NET? Our RPI is like 15 right now which obviously seems irrationally high.

 

[Cusefan0307]

Ironically we have no bad losses, we’ve just lost bad

But the bad teams we played we didn't blow out so that kills us.

 

[Millhouse]

Can you mathematically define what is considered “well” for predictive accuracy, and demonstrate how the various analytics meet that criteria?

Almost inevitably, people point out anecdotal evidence like you just did - which I find incredibly ironic considering the topic.

It was an example that I picked because it was a whole big argument on the football board

I will get right on it for you going through way back machine finding every teams SRS and how well it predicted the future just for you, to hell with my family, random internet guy demands more

 

[Cusefan0307]

JN, you’re smarter than me at this stuff. Why is there such a massive discrepancy between our RPI and NET? Our RPI is like 15 right now which obviously seems irrationally high.

Not JN, but 75% of the RPI formula is not how you play but how your opponents and their opponents perform. That was why it was so controversial in the first place. Only 25% is based on team performance.

 

[Cusefan0307]

If they used NET instead of RPI in 2007 we would have got in the tournament. We were better than about 10-12 at large teams according to KenPom and SRS including 4 seeded Virginia, 7 seeded BC, 6 seeded Vanderbilt and 7 seeded Nevada. What a friggin hose job that year was.

The RPI was so much worse. Teams just picked out of thin air based on terrible metrics.

 

[Cusefan95]

It was an example that I picked because it was a whole big argument on the football board

I will get right on it for you going through way back machine finding every teams SRS and how well it predicted the future just for you, to hell with my family, random internet guy demands more

That’s kinda my point, chief. You (and everyone else) effectively take it on faith that the results are “predictive” (whatever that means). Statistics are great if you know what the inputs are - if you don’t know what the inputs are, you’re just assuming this isn’t the equivalent of seeking advice from Oracle at Delphi.

 

[Cusefan95]

in 2021-22 season the vegas favorite won 76% of the time

I don’t think that really gives info at a granular enough level to be useful. What I’d be interested in is what percentage of teams favored by a point won, 2 pts won, 3 pts won, etc in a graphical form.

Even better if it can take variance into account - if a team is favored by 1 and wins by 20, I don’t think that is a successfully model. I’d want a model which is accurate say 60% of the time at picking the winner AND within something like 4 points of predicted margin.

I question if a lot of these models rack up “easy wins” on games like UConn vs UAlbany which drive up percentage of correct predictions, then use anecdotal examples to “sell” the model. I just haven’t seen the proof that games between relatively evenly matched teams get predicted correctly at better than a coin flip.

 

[NorSyr66]

Not every record is built the same even if you have a better schedule. Seems pretty simple. ST. John’s, Clemson, Bama all have better wins than us. We have sht the bed against every good team and we didn’t exactly blow out some of the mid majors we played either(Cornell/Colgate/New Hampshire/Canisius).

I agree with your post but it did get me to think about why maybe we did not blow out some of the mid majors. Early in the season Red was playing 11 players with all kinds of combinations trying to find the right five to use in a variety of game situations. Simply stated, it was a young team with a new coach. I feel we are getting the better results now from the earlier experimentation. As many have said, continue to win and the rest will take care of itself.

 

[jncuse]

JN, you’re smarter than me at this stuff. Why is there such a massive discrepancy between our RPI and NET? Our RPI is like 15 right now which obviously seems irrationally high.

They are both outliers, and both rankings are probably showing some of the weakness they both have. But at a very high level we are getting crushed by Margin in NET, and we are getting boosted by being "deemed" to have the 2nd best schedule in the country by RPI. As you will see below those are two factors that the respective formulas might be over-sensitive too.

-------

#1) I'll start with the NET - our NET is bad for two reasons. You are probably aware of this as it has been documented all over the place
a) We didn't blow out bad teams as much as we should have. Canisius by 12, Niagara by 12, Colgate by 4, New Hampshire by 11. Other teams better than us in the NET crushed better than us. (This factor is totally irrelevant under RPI -- a win is a win)
b) When we lose, we lose bad. Against bad for a metric that considers margin (for RPI a loss is a loss)

#2) Now for the RPI. We are deemed to have the second best schedule under RPI. At a high level a 13-6 record when you have 2nd best schedule is pretty darn good. Hence we are #15.
The RPI Formula is quite simple - 25% of your "score" is your winning%, 50% of your "score" is the winning % of your opponents, and 25% of your "score" is the winning % of their opponents. You can see a lot of the components are SOS based (75%) - sometimes it would cause strange things. You can also see the formula is purely W, and L based -- margin doesn't matter.

So why is our SOS so high? You have to remember that RPI could be gamed (intentionally or unintentionally). Play teams at home that you should be able to beat but are still OK (If I remember the sweet spot was 100-170 at home, and anything below 300 if on the road-- avoid playing the really bad (sub 300) teams at home. Sometimes teams were helped by it, or hurt by it, purely unintentional.
- We played some really good teams in OOC
- We had that "unintentionally" really good schedule for RPI purposes .. note these are teams RPI's not NETS -- Cornell #39, Colgate #133, New Hampshire #147...we only had one really bad game #280, but it was not in the 300's either.

As far as I can see it the key to our RPI
- Our SOS is so high.
- We lost to generally really good teams, which the RPI is forgiving on.
- We win almost all our games against teams that are moderately good (Oregon, LSU, Pitt, BC, Miami, NC St, Cornell)

 

[Eric15]

If they used NET instead of RPI in 2007 we would have got in the tournament. We were better than about 10-12 at large teams according to KenPom and SRS including 4 seeded Virginia, 7 seeded BC, 6 seeded Vanderbilt and 7 seeded Nevada. What a friggin hose job that year was.

The RPI was so much worse. Teams just picked out of thin air based on terrible metrics.

If you don't mind me asking, what about seasons that we just barely missed (1997, 2002, 2008), and then seasons where we got in without much room to spare (2016, 2018)? If NET was used in those seasons instead of RPI, would that have flipped it the other way?

 

[CardiacCuse]

St Johns 13-7
Slightly worse schedule
14th in SRS

Clemson 13-6 worse record
31st

We're getting killed for games where we quit. Maybe that's something that should be punished.

Florida State worse schedule 12-8, ranked higher

i feel like maybe part of it is the NCAA is strongly trying to weed out teams that might quit in NCAA tournament games (and thus tank the ratings for such a game)

*just a theory

 

[CardiacCuse]

The thing that is saving Clemson is going on the road and beating Alabama (at the end of the season that might be this entire conferences best non-conference win). They also beat South Carolina who is tied for 3rd in the SEC at 17-3 and beat a 15-5 Big 12 opponent in TCU at a neutral site. Without those they would be dead in the water based on their sub-par ACC performance so far.

so beating 14-6 Alabama is saving 13-6 Clemson ...and causing both teams to have much higher rankings than 14-6 Syracuse??

ok, got it.

now it makes so much sense. <<sarcasm>>

 

[br801]

i feel like maybe part of it is the NCAA is strongly trying to weed out teams that might quit in NCAA tournament games (and thus tank the ratings for such a game)

*just a theory

I don't think teams "quitting" in the NCAAT would have much of an impact on ratings, as viewers would likely just move to another game now that most are readily accessible.

 

[Cusefan0307]

If you don't mind me asking, what about seasons that we just barely missed (1997, 2002, 2008), and then seasons where we got in without much room to spare (2016, 2018)? If NET was used in those seasons instead of RPI, would that have flipped it the other way?

16 and 18 we would have been in comfortably with NET I believe.

I have always thought BC and St. John’s getting in over us in 2002 was crap, but that’s was when the last 10 games were also heavily weighted and we completely sht the bed down the stretch. Still by a rough glance we had at least 4 quad one wins in Wake, Notre Dame 2x, and Sparty. Under the current system where all wins are equal no matter when they happen we definitely get in.

I always thought the last 10 games being weighted more than the rest of the season was garbage because the schedule was always backloaded in the BE to create good TV matchups.

 

[Cusefan0307]

so beating 14-6 Alabama is saving 13-6 Clemson ...and causing both teams to have much higher rankings than 14-6 Syracuse??

ok, got it.

now it makes so much sense. <<sarcasm>>

You aren’t even trying to get it. Both of those teams have significantly better wins than us. Just look at their schedules.

 

[CardiacCuse]

we are an inefficient basketball team and that’s what hurts us, in addition to the big losses. We rank 13/15 on offense efficiency in conference.

When efficiency and margin of victory are not considered, like in the colley rankings, we are #26. Hope the committee focuses on the W’s and L’s and not how the game was played. The efficiency rankings really disfavor a young team learning how to play early in a season.

i think this team would be like a really smart college kid that slacks off and does just enough to get by...

meanwhile, the teacher's pets that work really hard but arent even "smarter" are the ones that "look good" in general

this is a young team that kinda just plays to the score...eases up with leads, collapses when down big...makes their metrics look worse than they are, imo...that and not being to beat upper tier teams...

 

[CardiacCuse]

You aren’t even trying to get it. Both of those teams have significantly better wins than us. Just look at their schedules.

YOU are in denial

clemsons biggest win is alabama lol

i did look at their schedule...its weak...theyve lost to: memphis, miami, va tach, ga tech, unc and duke...and yet...they are way higher than the cuse???

they have a few decent wins but no one really all that good.alabama, pitt, tcu...? who else? that really screams waay better than the orange? please explain.

alabama lost to: ohio state, clemson, purdue, arizona, creighton, and tennessee

they beat: LSU, Oregon (hmm familiar), Missouri, South Carolina...and just got their best win of the season in auburn a few days ago...they havent beaten any team that is a legit final 4 contender, imo

both teams are decent...and arguably deserve to be ranked a bit higher than the orange...but in a simialr tier, imo

Makes ZERO sense for alabama to be 7th, Clemson to be 28th, and SU to be 78th

absolutely zero.

 

[Eric15]

YOU are in denial

clemsons biggest win is alabama lol

i did look at their schedule...its weak...theyve lost to: memphis, miami, va tach, ga tech, unc and duke...and yet...they are way higher than the cuse???

they have a few decent wins but no one really all that good.alabama, pitt, tcu...? who else? that really screams waay better than the orange? please explain.

alabama lost to: ohio state, clemson, purdue, arizona, creighton, and tennessee

they beat: LSU, Oregon (hmm familiar), Missouri, South Carolina...and just got their best win of the season in auburn a few days ago...they havent beaten any team that is a legit final 4 contender, imo

both teams are decent...and arguably deserve to be ranked a bit higher than the orange...but in a simialr tier, imo

Makes ZERO sense for alabama to be 7th, Clemson to be 28th, and SU to be 78th

absolutely zero.

Clemson has three Quad 1 wins to our one. And their losses were very close -- to Memphis by 2 points, to Duke by 1 point, to Ga Tech by 3 in 2OT. We have been blown out in all six of our losses which works against us because NET factors in scoring margin.

 

[Cusefan0307]

YOU are in denial

clemsons biggest win is alabama lol

i did look at their schedule...its weak...theyve lost to: memphis, miami, va tach, ga tech, unc and duke...and yet...they are way higher than the cuse???

they have a few decent wins but no one really all that good.alabama, pitt, tcu...? who else? that really screams waay better than the orange? please explain.

alabama lost to: ohio state, clemson, purdue, arizona, creighton, and tennessee

they beat: LSU, Oregon (hmm familiar), Missouri, South Carolina...and just got their best win of the season in auburn a few days ago...they havent beaten any team that is a legit final 4 contender, imo

both teams are decent...and arguably deserve to be ranked a bit higher than the orange...but in a simialr tier, imo

Makes ZERO sense for alabama to be 7th, Clemson to be 28th, and SU to be 78th

absolutely zero.

Clemson beat TCU, Bama, and South Carolina. Alabama beat South Carolina, Auburn, and Mississippi ST.

All 6 of those wins are better than anything we have done. We also sht the bed against teams we should have blown out. They did not.

Beating Pitt and Oregon isn’t enough.

 

[SUFan44]

Since all these ratings are so highly correlated with each other, I'll just look at basketball reference

14-6 against the 13th toughest schedule should be good enough. 13-6 if you toss d2

The computers hate that we lost by 22 on average.
We win by 11 on avg (tossing charminaid)

It would be a shame if this year is decided by our effort at the end of bad losses

Tennessee comparable schedule 15-4 record. Their srs ranking is 8th. We are 72nd

Thank you for this great board

It is because we aren't very good. I think our record will even itself out. We have no really good wins. The BC road loss is a bad loss even though the metrics say otherwise.

We will finish somewhere in the range of 17-12.

 

[Cstate]

YOU are in denial

clemsons biggest win is alabama lol

i did look at their schedule...its weak...theyve lost to: memphis, miami, va tach, ga tech, unc and duke...and yet...they are way higher than the cuse???

they have a few decent wins but no one really all that good.alabama, pitt, tcu...? who else? that really screams waay better than the orange? please explain.

alabama lost to: ohio state, clemson, purdue, arizona, creighton, and tennessee

they beat: LSU, Oregon (hmm familiar), Missouri, South Carolina...and just got their best win of the season in auburn a few days ago...they havent beaten any team that is a legit final 4 contender, imo

both teams are decent...and arguably deserve to be ranked a bit higher than the orange...but in a simialr tier, imo

Makes ZERO sense for alabama to be 7th, Clemson to be 28th, and SU to be 78th

absolutely zero.

South Carolina just beat #5 Tennessee (who beat us pretty well I recall) last night and we lost to BC. Still thinks it makes absolutely zero sense where us and Clemson are seeded in the NET?