Arizona, at 9-5 is 13th at Kenpom......best wins are Cincinnati and West Virgina.
We have FOUR wins better than those and are 13-3. 21st.
We have FOUR wins better than those and are 13-3. 21st.
Computer models are stupid
- Thread starter silentsaregolden2.0
- Start date
Our defense is killing us, that's why.
Losing by 20 to Ohio State is a lead boot to the nuts of our Kenpom ranking as well as a few other middling performances.
Last edited:
HOU is 12-3 has zero Quad 1 wins and is 4th in the NET??? How???
We're 18th with 4 Quad 1 wins and all losses are Quad 1.
WTF???
Even so, Ken Pom has always been silly and way over rated
It seems hard for some people to grasp, but to put it simply, KenPom doesn’t account for wins and losses. It’s simply a measure of how efficiently a team scores and how well they prevent the other team from scoring on a per possession (not per game) basis. In the long run, those things tend to correlate really well with winning, but in small sample sizes, you’ll can get teams whose record either underperforms (Arizona) or outperforms (us) their underlying data. Like with most statistical models, you’ll also occasionally get a few straight up outliers. No one model is perfect, but it’s interesting data if you know what it is and isn’t doing.
Yep.
If we lost to Gonzaga by a point and only lost to Ohio State on a buzzer beater, we'd be ranked higher in Kenpom despite having a worse record.
We haven't really "gained" in ranking during conference play despite 2 high quality wins because we lost by such a large margin to Georgia. We are 2-1 in conference, but we have been outscored overall. We could have lost to Florida and MSU by a point, but beat Georgia by 5 and be higher ranked in Kenpom.
But UK is 4-0 against top 14 teams. Who cares what the defense is ranked if UK keeps winning?
I will say Houston being top 5 in NET while we're 18..is...puzzling. Yes, our losses are all to "unranked teams"..but they're all Quad 1. While we had 3 quad 1 wins to their 0 coming into the night..(now 4 but it hasn't been updated yet.)
That's a head scratcher.
That's a head scratcher.
KenPom isn't really measuring your resume. Unfortunately, Kentucky's defense sucks and that's why they took a hit on KenPom. Their defense has been crushed ever since the Ohio State loss.
Fanta said tonight that Alabama got to the FF last year with a defense at 111 I think, and Miami the year before I think with a defense at 99. Now, neither one made the final game, but it does show that can be overcome. You just can't have an off-night offensively.
This is correct. If we have to have a guy go 7-10 from three, hit 50% from threes and scored 90+ points to eek out wins, that means our defense is weak and that is not a formula for continued success.
Now, having said that, it was a helluva win !!
Last edited:
Texas A&M is an elite defensive team. Maybe the best in the SEC.
Alabama is an elite offensive team. The best in the country.
Going to be an interesting week at Rupp.
Alabama is an elite offensive team. The best in the country.
Going to be an interesting week at Rupp.
They are better than nothing for some purposes and they can definitely expand your insight into upcoming games if you know how to be really responsible with the information they hold. Which is probably incredibly rare.
But yeah as straight predictors lol. We are 3 and 0 against teams those models said should beat us.
…and 0 and 3 against some key teams they said we should beat. Which in our case makes me wish Kenpom would start accounting for officiating staff.
If only UK and Arizona had a common opponent. Hmmm
What a joke.
What a joke.
Last edited:
Because it can lead to 20 point whoopings by inferior teams. Our defense needs to improve or we are going to really be hurting on off nights.
And yet I think no team has ever won the title and been ranked outside of both off/def T10 on there. Could be wrong on that, but it's not the pile of dog crap you think it is.
Believe it’s top 30 or so. It obviously is somewhat of a garbage metric as shown above in a few instances. Those are some egregious examples of terrible rankings by a few teams. It’s clearly flawed.
We all get that the defense needs to get better. Everything on any team needs to improve if you want to advance in March.
But fans are paranoid about the defense while treating those issues like we're the only team in America that has them.
There is no other team in the country that has as many quality wins as we do.
Or option B. play mediocre offensively and defensively and pull it out at the end (Duke game).
Im not doubting that. Im just concerned we aren’t in the same Tier as Auburn and Ewetee just yet. Maybe it’s just because this conference is insanely hard right now but I just don’t feel quite at ease. I feel like we can hang with anyone ONLY when we hit threes. Which isnt awful mind you, I just don’t want to see our season end in March because we couldn’t buy a bucket again.
As long as we win, I’m cool with Kenpom rankings. We are stacked right now with Quad 1 wins.
Its like the weatherman looking at his computers and radars but never looking out of the freaking window.
No. KenPomPom is a pace measurement. He doesn't count offensive rebounds as a new possession. So his basketball numbers are flawed. If you want to know, or care, how fast a game was played without watching it then KenPomPom is your nerd number to go by.
Auburn is taking a significant step back with Broome injured for a few weeks, if not longer.
Tennessee crapped the bed against the second real team they've faced this season.
I'm not ready to anoint either SEC King yet.
When you score 90 pts a game your defense is going to give up more points. I’ll take wins and realize that all of these models are just metrics.
And the net also has Michigan 9th! With a quad 2 loss and 2 quad 1 wins. What’s more infuriating is the committee uses this for seeding. I know people have mixed feeling about bilas but I love how he’s always called these models out.
Honesty couldn’t put it any better than that statement right there. It’s like the eye test doesn’t even matter when it should be the most important.
Once upon a time, there was a team that was sitting at 6-5 with their best win through 11 games coming from Indiana St. Most people wrote this team off.
This team then went on to make the final four only losing to eventual Champion Uconn.
Some things are more predictable than others.
Stats can be manipulated to get any result you seek. Stats are flawed to a degree and it usually depends on what it is your looking for.
The one model that seems to be more closely geared to actually playing the game is Aike's.
There will updates to Aike's model and I scratch my head (looking back, before our game against Ohio St, I asked him why he had them so high) asking myself what his he seeing in this team or that team and if you give it a week or two, his model will be very close to being spot on.
Shout out to Aike and his work.
The one model that seems to be more closely geared to actually playing the game is Aike's.
There will updates to Aike's model and I scratch my head (looking back, before our game against Ohio St, I asked him why he had them so high) asking myself what his he seeing in this team or that team and if you give it a week or two, his model will be very close to being spot on.
Shout out to Aike and his work.
All the efficiency stats are modeled to how the game is actually played.
Efficiency stats are based on four factors. Both on offense and defense.
1) Shooting %
2) Turnover %
3) Rebounding %
4) Free Throw Rate
Those four things almost completely explain the difference in efficiency between teams.
I cannot think of anything more important to the actual game of basketball than those four things above. If you do those things well, your a very good team and highly ranked. The more of those things you do, the better.
When it's all said and done, this isn't rocket science.
Shoot better than your opponent.
If you cannot do that, you better find ways to generate more shots whether that be from forcing more turnovers than giving up or rebounding more of your misses than the opponent rebounds of theirs.
Or just be Georgia and bulldoze your way to the line and shoot 400 free throws a game lol.
Shoot better than your opponent.
If you cannot do that, you better find ways to generate more shots whether that be from forcing more turnovers than giving up or rebounding more of your misses than the opponent rebounds of theirs.
Or just be Georgia and bulldoze your way to the line and shoot 400 free throws a game lol.
One of the issues I have with a lot of systems is a lack of accountability. Maybe they check behind the scenes I don't know but most of the time predictions are made, the games happen and there's no review (at least publicly) on how the system did.
What I like about Kenpom is he publishes the model diagnostics. The system is well calibrated.
Break the games down into bins and when he says teams have a 55% chance of winning, they do in fact win 55% of the time. 65%....65% etc etc.
Nothing is ever going to be perfect or even relatively close to perfect. There's so much random variation in everything. You can only get so far.
But I think what we've gotten to with these systems, is about as good as we might get.
What I like about Kenpom is he publishes the model diagnostics. The system is well calibrated.
Break the games down into bins and when he says teams have a 55% chance of winning, they do in fact win 55% of the time. 65%....65% etc etc.
Nothing is ever going to be perfect or even relatively close to perfect. There's so much random variation in everything. You can only get so far.
But I think what we've gotten to with these systems, is about as good as we might get.
My thoughts on stats come from my days coaching high school ball.
Coaches would use stats and have no idea on what they were actually reading, how to use stats.
Take batting averages; coaches would rave about this kid who is tearing git up, hitting over.500 and be in awe.
I took teams, broke them in groups of A, B, C. then throw out all the C team games, , do the same with the B teams, then study only the games against A competition,. The take that a step further and see who was pitching that day, what type of pitcher she was, etc.
I would break it down so I knew the true stats of a team and player.
I would know what pitch your player swung on, how many times she hit foul balls to one side or the other.
Does Ken Pom take into account the foul disparity in our game with Georgia? When your players have fouls on them forcing them to sit on the bench, do you really get a good take from that game?
Our game last night was just as physical; the biggest difference I saw was the class of the Miss St players after a hard foul reaching out to UK players to give them a hand up.
But like you said, just shoot the ball better and all will be OK.
I guess it's baked in to the 4th factor.
This probably isn't the most popular take in the world because everyone quotes FT disparity (in games where there's a large amount for the other team, but never quote when we have that advantage lol).
But getting to the line and conversely not fouling is a skill. It's not to say the refs don't have some bias (usually to the home team). But that bias is baked already into the home court advantage adjustment IMO.
Did UGA get to the line because the refs were bias against UK or because they are just good at getting to the line. They went there 37 times just the very next game vs Oklahoma.
But yeah I agree. I mean it's one of those things that cannot completely be accounted for. Your main guy gets two quick fouls in the first half and has to sit the rest of the half. That clearly has an effect on the outcome of the game. So definitely can't measure everything.
I don't think Tennessee is the juggernaut so many have been billing them as. They got housed by 30 and it wasn't that close.
They also struggled with Texas. These aren't traits of some great team.
I think we split with UT, but if we play like we did last night, we could get them in Knoxville.
We have 5 legit scorers, the best on-ball defender in the game, one of the best defensive bigs and Oweh will win his matchup more times than not.
Last night proved a lot to me, MSU is legit and they wanted this one more than any other game. That was a massive win.
Out of those 4 stats, only 1 has anything to do with defense (rebounding), the rest is all about offense, so if that's the case, ehy is our defensive efficiency rating weighing us down so much?
I think there has to be a 5th element that puts more weight on ppg allowed per possession, right?
