Not that it matters, but just curious re: Computer Polls

[HagginHall1999]

How are we currently ranked #26 in Kenpom?

In the same poll we have losses to #29 and #31 (OSU/Clemson) and wins over #2 and #6 (Duke/Gonzaga).

It just doesn't make much sense.

 

[B.B.H.]

How are we currently ranked #26 in Kenpom?

In the same poll we have losses to #29 and #31 (OSU/Clemson) and wins over #2 and #6 (Duke/Gonzaga).

It just doesn't make much sense.

Margins have a big impact. Most of our recent wins have very low margins and our most recent loss was 30 point lower margin than expected. Also our defensive efficiency has tanked thanks to Louisville game and Ohio State. Plenty of time to improve it but much work to be done

 

[UKBB4Ever]

KenPomPom uses fake numbers.


It’s a pace metric. Nothing else.

 

[NociHTTP]

This doesn't back anything up with numbers, but my eye test says that unless we're going to stick to the offense guru's plan (the guy Pope brought with him to teach this stuff), the team really isn't very good. When they run that offense and then put the defensive clamp on when the other team panics, they're an excellent team. Otherwise, it's a poor defensive team with guys who can't create their own shots, are afraid of contact and are blocked more by the rim than they are by the opponent.

 

[bbncal02]

How are we currently ranked #26 in Kenpom?

In the same poll we have losses to #29 and #31 (OSU/Clemson) and wins over #2 and #6 (Duke/Gonzaga).

It just doesn't make much sense.

That blowout loss to OSU killed us. We were #12 in Kenpom before that. Point being is, T10 teams don't get blown out on a neutral court by middling squads. It's a Q1 loss but you can't lose that game by 20 and hope to maintain good metrics.

The other part is while the wins over Duke and Zaga are good, we barely won them. That doesn't help you as much metric wise.

Best thing we can do is come out Tuesday and beat the tar out of Brown and get ready for the hell that is conference play.

 

[The_Answer1313]

Yeah as most have said.........humans tend to put way too much emphasis on wins/losses but really margin of victory is a better predictor for future success than actual W/L record.

The fact we got blown out by OSU is what caused a massive drop.

Now the good news.

The committee while they use NET, don't strictly go by NET when seeding teams. It's very much still resume based. So the wins vs Duke and Gonzaga are still on the board and still will weigh very highly despite the fact those were close games that could have very well ended differently. So 26th in Kenpom but I don't think anyone is projecting us a 7 seed. Bracket Matrix has us avging a 3 seed currently.

The bad news is if we don't get better, in all likelihood we won't perform to the seeding that we get. And that's where the computer metrics are important.

I really wish the rest of Kenpom wasn't behind a paywall. People pay too much to the ranking but not nearly enough to the wealth of information on the team pages that provide information on exactly WHY teams are ranked where they are. Similar info can be found for free on Bart Tovik.

 

[The_Answer1313]

Also, there's a reason why all these computer metrics follow basically a similar methodology.

Efficiency can be explained 95% of the way through four factors:

1) Shooting
2) Turnovers
3) Rebounding
4) Getting to the line.

Factor in a fifth factor of Free Throw % and you are probably 99.99% of the way to explain offensive and defensive efficiency.

The methodology ......is sound.

Kenpom model diagnostics for 2024 season:

Win probability forecasts​

Predicted Range	W-L	Win%	Exp Win%
50-60%	765-623	55.1	54.9
60-70%	867-505	63.2	64.9
70-80%	901-295	75.3	74.8
80-90%	816-136	85.7	84.8
90-95%	367-30	92.4	92.3
95-98%	239-12	95.2	96.5
98-99%	84-2	97.7	98.5
99-100%	84-0	100.0	99.5

As you can see, the system is calibrated well. If the system gives you a 85% of winning a game, those teams typically win 85% and so on.

 

[HagginHall1999]

That blowout loss to OSU killed us. We were #12 in Kenpom before that. Point being is, T10 teams don't get blown out on a neutral court by middling squads. It's a Q1 loss but you can't lose that game by 20 and hope to maintain good metrics.

The other part is while the wins over Duke and Zaga are good, we barely won them. That doesn't help you as much metric wise.

Best thing we can do is come out Tuesday and beat the tar out of Brown and get ready for the hell that is conference play.

Understand blowout loss and barely winning but vs Duke/Zags, assuming we were not favored....seems like that would matter more.

Also, despite their mediocre record, OSU still just 2-3 spots behind us in KP.

 

[ukcatz12]

How are we currently ranked #26 in Kenpom?

In the same poll we have losses to #29 and #31 (OSU/Clemson) and wins over #2 and #6 (Duke/Gonzaga).

It just doesn't make much sense.

Kenpom isn't looking at resume, it's just an efficiency formula. I don't believe it even takes wins and losses into account. It's just looking at how efficient a team is on offense and defense and adjusts it based on the competition.

 

[The_Answer1313]

Another thing we need to remember is we started out 43rd in Kenpom (which he admit on his blog felt rather low).

By the Duke game we were 31st. After Duke we jumped 9 spots to 22nd. So we did move up quite a bit in that one.

We didn't move up nearly as much after Gonzaga but we also had just lost the previous game to Clemson at that point.

Another thing is instead of focusing on the actual rank, I'd focus more on the actual efficiency margin number. The gaps between the teams in different positions are not the same. So you can make huge jumps after certain games and maybe not so big a jump in others if you are that far behind.

 

[The_Answer1313]

Kenpom isn't looking at resume, it's just an efficiency formula. I don't believe it even takes wins and losses into account. It's just looking at how efficient a team is on offense and defense and adjusts it based on the competition.

It's basically shifting things. The won/loss sets that line at 0 whereas Kenpom coming into the game there's an expectation. If you perform better than that expectation, system will assume you are better than where you currently are (and conversely your opponent is worse).

If you are expected to win by 5 and you win by 10, your Efficiency Margin will go up and thus your ranking. If you win but it's by less than expected, you might move down.

The OSU drop was actually TWO things. The fact we lost by 20 but also the fact we were predicted to win by 6. That's a 26 point swing there.

 

[The_Answer1313]

Teams are just so closely bunched together too.

UK is 26th with an EM of 20.68
Uconn is 15th with an EM of 22.30

They would basically be a 1 point favorite on a neutral court. That's how close we are to being 15th instead of 26th.

 

[bbncal02]

Understand blowout loss and barely winning but vs Duke/Zags, assuming we were not favored....seems like that would matter more.

Also, despite their mediocre record, OSU still just 2-3 spots behind us in KP.

Computer polls don't care about rankings per se, it's what you do against those rankings. We were favored against OSU, but got torpedoed. That's a huge negative metric wise, even if the eye test goes, "Well, just a bad game." Same for UofL. Computer doesn't care it was an emotional rivalry game. We gave up a lot of points to them.

It's not who you beat or lose to . It's HOW you win or lose those games. And let's be honest. We haven't looked good since WKU. Even if we've won those games.

 

[TucsonCat]

So wait. We're 26, Ohio State is 29, and Clemson is 31?

Its funny. I always see posts outraged or befuddled why we are ranked behind teams we beat.
Never seen one ask why we are ranked above teams we lost to.

 

[The_Answer1313]

So wait. We're 26, Ohio State is 29, and Clemson is 31?

Its funny. I always see posts outraged or befuddled why we are ranked behind teams we beat.
Never seen one ask why we are ranked above teams we lost to.

Haha yeah good point lol.

Also if you look at other boards I wonder if OSU and Clemson fans don't look at it and think to themselves "wait a second we beat UK why are they ranked ahead of us" lol

 

[JonathanW]

How are we currently ranked #26 in Kenpom?

In the same poll we have losses to #29 and #31 (OSU/Clemson) and wins over #2 and #6 (Duke/Gonzaga).

It just doesn't make much sense.

Don't look at it in a vacuum. Other teams have good wins and losses too. As a whole, our schedule has been weaker than many other top teams. There are also other factors. For example, we not only lost to OSU, but we lost on a neutral court (not on the road), AND we lost by 20. That # may be a bigger factor than you think. The Clemson loss, on the road, in OT, doesn't probably hurt us any.

 

[Aike]

Unless something has changed, preseason ratings still factor in for another 3 weeks or so.

So that initial 40 something rating is still weighted into the overall. We should probably be in the teens based on this season alone.

 

[B.B.H.]

Unless something has changed, preseason ratings still factor in for another 3 weeks or so.

So that initial 40 something rating is still weighted into the overall. We should probably be in the teens based on this season alone.

Torvik has us 15 and his metric doesn’t factor preseason projections at this point, or as far as I’ve heard. We need to win some games and win some of them big to move up.

 

[Aike]

Torvik has us 15 and his metric doesn’t factor preseason projections at this point, or as far as I’ve heard. We need to win some games and win some of them big to move up.

I’m almost positive I said we should be in the teens based on this season alone. You even quoted me. 15 is in the teens, right?

Also disagree that we need to win games big to move up. Our SOS going forward is brutal. Winning at all will take care of everything. Not that it will be easy.

 

[B.B.H.]

I’m almost positive I said we should be in the teens based on this season alone. You even quoted me. 15 is in the teens, right?

Also disagree that we need to win games big to move up. Our SOS going forward is brutal. Winning at all will take care of everything. Not that it will be easy.

Oh I wasn’t critiquing you, I was just showing what the numbers look like sans KenPom early season bake-in. And i think it is true that if we win enough of our games in SEC we will be fine, but I expect the best case scenario is 13-5. If you end up closer to expectation, such as 11-7, 10-8 or so you can really boost yourself with big margins in some of the wins.

 

[Aike]

Oh I wasn’t critiquing you, I was just showing what the numbers look like sans KenPom early season bake-in. And i think it is true that if we win enough of our games in SEC we will be fine, but I expect the best case scenario is 13-5. If you end up closer to expectation, such as 11-7, 10-8 or so you can really boost yourself with big margins in some of the wins.

Sorry, I misunderstood. I run my own numbers, so I had a pretty good idea of where we “should” be, analytically. Not that there would ever be consensus.

Somewhere around the teens feels about right to me. Should caveat that this field is pretty deep. In the teens this year would equate to top 10 two years ago.

 

[B.B.H.]

Sorry, I misunderstood. I run my own numbers, so I had a pretty good idea of where we “should” be, analytically. Not that there would ever be consensus.

Somewhere around the teens feels about right to me. Should caveat that this field is pretty deep. In the teens this year would equate to top 10 two years ago.

All good, my brother. I know we all hope to see Kentucky near the top by season’s end