Current Pomeroy Rankings
- Thread starter Smashcat
- Start date
Used to think KenPom was easily the best model and maybe it is, but clearly it is flawed. UL would have to lose to Southwest Polytechnic Institute State by 50 to drop below us.
Kenpom is supposed to adjust for Strength of Schedule, but it's pretty obvious his adjustments are insufficient.
Kenpom is only the best ranking method when it has UK rated highly. Otherwise it's terrible.
Massey and Sagarin have UL higher than us as well.
I choose to think of this simplistically: generally speaking, margin of victory is all important with computer models, even if those wins come against weak opponents. That explains a lot. Don't tell me if my theory is wrong, it causes some confusion.
Our 3 losses(who we lost to) and their subsequent records are enough to validate a rating system evaluation of us. Wins you should get(no matter how bad the team you play might be) trump losses you shouldn't have. Even if that is the case, this year in college basketball is so void of really good teams that any system will be more flawed than usual, this will include season rankings, NCAAT/NIT seedings or anything else that applies a number value to a team.
Look at the games in the last 2 days, Ala runs SC out of the gym, Clemson beats Duke, Iowa beats Mich St for a second time . No one or no system has any more idea than the hard core fan here which teams are good and which are pretenders.
Look at the games in the last 2 days, Ala runs SC out of the gym, Clemson beats Duke, Iowa beats Mich St for a second time . No one or no system has any more idea than the hard core fan here which teams are good and which are pretenders.
Last edited:
I do not put stock into computer ranking. March rolls around all of that goes out the window and whoever shows up to play that day wins. That's all that matters in the end.
Computer rankings are useful when referring to rating teams after the season is over. MId season they're pretty useless. (especially when UK is ranked low)
UL is not a top 10 team, they're going to get destroyed in Feb and will drop out of the top 25 by tourney time. Every win they've got is fools gold, take a look at their Feb schedule if you haven't already, based on everything I've seen so far not many wins there.
Every year there are a few teams that look better in the computers than in real life, and end up top 10/20 with people scratching their heads. Wisconsin has traditionally been one of them. KenPom has addressed this as one of the biggest flaws in his system, but one he doesn't know a way to compensate for.
What did you just state?
Do you know or don't you...Come on be a UK homer like me. I keep a pair of Blue tinted glasses in my pocket at all times.I guess these services have to have some output. Not sure I agree but it's fun watching the debate....
They might as well draw teams out of a computerized hat, then tell us how good they areWhat did you just state?
I guess these services have to have some output. Not sure I agree but it's fun watching the debate....
Anyone else having trouble typing posts, sometimes there is a delay or it misses keystrokes completely?
Yep that's your world ran internet.....catching all your keystrokes is tough business...
The one thing I think most people miss is just how close these teams are in terms of talent.
The difference between the number 25th team in Sagarin Baylor and say 10th ranked Virginia is a mere 4 points.
The further down you go the less the difference becomes. There's about a 4.3 point difference between the 50th ranked team and 20th ranked team
With all the teams bunched together, teams switch rankings far more than your typical once a week AP or coaches poll.
The difference between the number 25th team in Sagarin Baylor and say 10th ranked Virginia is a mere 4 points.
The further down you go the less the difference becomes. There's about a 4.3 point difference between the 50th ranked team and 20th ranked team
With all the teams bunched together, teams switch rankings far more than your typical once a week AP or coaches poll.
http://www.masseyratings.com/cb/compare.htm
This site is nice as it looks at a ton of rankings out there
This site is nice as it looks at a ton of rankings out there
Oh wait I'm on one,,, okay I hate computer generated rankings
If only this were settled on the court .
I never pay attention to these rankings , it's strictly numbers with no accounting for the human variable . Like for instance UK , there is motivational situation from a couple of players that produces wildly different results depending on their frame of mind . UK could be 21 and 1 on any given day . This applies to other teams as well .
![[laughing]](http://l.yimg.com/j/assets/img/emoticons/classic/laugh.r191677.gif "Laughing [laughing]")
![[thumb2]](https://ct.yimg.com/mr/uploads/891/1684387.png "Thumbs Up 2 [thumb2]")
That's precisely why UK is rated lower. A team that plays like the 15th ranked team every single game will win more games during a season than a team that plays like the number one team one game and then the sixty-fourth ranked team the next. The computer ratings are reflecting that.
UK fans (not necessarily directed at you personally) that complain about our ranking see our best performances and think that's who we are and where we should be ranked. However, that mindset ignores how badly we've looked at times. The truth is that the team is an average, so to speak, of the two. If the team gets more consistent, then they will rise in the rankings.
You made good points but it still makes no sense to have a teams you beat ranked higher . Even if you lost other games the one true comparable is head to head matchup . I already stated I don't care about the rankings but they are wildly flawed . The computer does not compensate for motivation , it cannot tell when you are inspired to play or not . The only thing it can do is crunch numbers and give an average but there are teams close to UK who's ceiling is nowhere near capable of winning the tournament . It also doesn't factor in potential growth of a team playing elite freshman , which will be significantly more than an older team . I don't know but I wonder where kenpom had UK ranked going into the tournament for 13-14 .
What your talking about is the game by game aspect of this. Obviously there's a ton of variables that go into the outcome of a game.
These rankings or any ranking system human or computer for that matter, it's all based on averages. So when you see an offensive efficiency rating of 113.9 that means that on average against an average defense, UK would be expected to score 1.14 points per possession. On a game by game basis, they could score 1.30 against this team or 1.00. It can vary greatly. There is no system in the world that's going to account for that. Sports is sports. There's always going to be varying performances that are outside explanation. The goal of these systems is to get as close as possible to the underlying talent levels. You never going to get there 100% of the way tho.
The ranking system tho is based on what happened. So IMO it factors the human variable as well. If UK doesn't do as expected, they drop. I like Kenpom because the offense and defense efficiency numbers are what they are. You can't make them up. If UK plays a game and scores 68 points in 68 possessions, they have scored 1.00 points per possession. You can't make that up. It is what it is. The underlying process with Kenpom is sound. You can make a case that the adjusted figures are inaccurate (maybe he's adjusting too much or too litte).......you can also argue that maybe the MOV cap isn't accurate enough. Or the fact that he does take into account what happened last year and that gets weighed less as the current season goes on. So there's different aspects you might agree or disagree with, the as a whole the system is sound.
But the thing about that is the UL game was just that.........ONE GAME. And that game was IN RUPP. We won by 2 points. Home court advantage is worth 4 points. So the system really sees it as had we played UL on a neutral court, they would have won by 2.
You have to factor home/away and also have to factor one game. You have to take the entire season into account.
You have to factor home/away and also have to factor one game. You have to take the entire season into account.
Teams with talent
1 point better win 54%
2 points better win 56%
3 points better win 60%
Those numbers are estimates.
So say Team A was 3 points better than Team B and they play on a neutral court. Team B still wins that game 4 outta 10 times. This is just based on pure chance.
That's why you can't just look and say Team A beat Team B therefore they are better.
1 point better win 54%
2 points better win 56%
3 points better win 60%
Those numbers are estimates.
So say Team A was 3 points better than Team B and they play on a neutral court. Team B still wins that game 4 outta 10 times. This is just based on pure chance.
That's why you can't just look and say Team A beat Team B therefore they are better.
If it makes no sense for a team you've beaten to have a higher ranking, then that means the following statement MUST also be true. The better team ALWAYS wins every game. Do you believe that to be the case? It would seem from your comment about motivation, that you don't.
We just beat both UL and Duke , beat Duke handily but I'm supposed to believe the computer says they're better than us ? No
Whether you think they're better than us or not is up to you. That wasn't the question though.
I learned something today. I never realized they could design software to determine how a team. looks when they win or lose. Technology amazes me daily.
It's factored in tho.
What difference does it make if the team played poorly because they couldn't shoot well one night or the team played poorly because they all got dumped by their girlfriends?
So yeah the computer (or any poll for that matter) doesn't see the WHY, they just know that something happened and the rankings are adjusted accordingly.
