Nice comparison of UVA and UK
- Thread starter Dutycat
- Start date
Yep, that was my comment. I saw a SOS sheep in the comments saying UVA had clearly played the tougher schedule because kenpom said..... I really hate this SOS nonsense. It is totally broken.
Really think UVA would follow the Columbia game plan. Methodically run their offense late into the shot clock and apply a very good defense. They worry me the most right now with Kansas and Duke as 2/2a. I see Kansas getting better and would have serious revenge motive while Duke has the shooters to go off like they did against Wisky.
The overall SOS from something like KenPom matters a lot when you're talking about things like X/game or X/possession, because you're comparing how a team has done on each possession/game/whatever, and therefore you must consider the quality of each and every opponent (and each opponent's Off/Def rating).
Where ratings that factor in every team do not matter in SOS is when talking about how a team has fared record wise against its competition. Anyone that follows the turd-ish RPI system that the NCAA relies on knows that it's better for RPI SOS to play all middling teams rather than several great teams and several sub-300 ranked teams. That holds true, to some degree, for any comprehensive system of measuring SOS. Yet we all know that, at least for elite teams, the risk of loss to any team below a certain threshold (say, top 40-ish) is essentially nil. So what really matters to a team's record is how many (realistically) losable games they played and how they fared in those games. So a lower SOS team could have a more impressive record by virtue of playing more elite teams. This is because a W-L record is a binary measure of what happened.
TL
R - different measures of SOS for different things.The next stretch of games for Virginia will tell the tale. They do play good d and are a cohesive unit.
I have a hard time believing a team we beat by 30-plus has made such a monumental turnaround that they could beat us just two or three months later.
I do agree about Virginia and the way they would play us, but they're going to run the gauntlet here in a week or so. They'll come out of that stretch with a couple of losses, and the narrative will change.
But I do think a lot of times people weight college basketball losses way too severely. Wisconsin is still a super-dangerous team, but they've been off the radar since the loss to Duke. People have this weird habit in this sport of just forgetting about teams when they lose, and I think in a way the early-season loss can actually be a positive.
So I went to Ken Pomeroy's site, and got the ranking of all of UK's (18) and UVA's (17) opponents.
There schedules aren't overwhelmingly different. But as we all believed, UK's schedule is stronger. I'm pretty sure if I did a non-parametric analysis on the data, it would say UK has the stronger schedule, but that result would likely not be statistically significant.
UK UVA
Median Opp Rank 65 103
Mean Opp Rank 108 110.6
Top 20 Opps 4 3
Top 50 Opps 6 6
Top 100 Opps 12 8
Top 150 Opps 13 13
Opps ranked 200-300 2 0
Opps ranked 300+ 1 2
There schedules aren't overwhelmingly different. But as we all believed, UK's schedule is stronger. I'm pretty sure if I did a non-parametric analysis on the data, it would say UK has the stronger schedule, but that result would likely not be statistically significant.
UK UVA
Median Opp Rank 65 103
Mean Opp Rank 108 110.6
Top 20 Opps 4 3
Top 50 Opps 6 6
Top 100 Opps 12 8
Top 150 Opps 13 13
Opps ranked 200-300 2 0
Opps ranked 300+ 1 2
Here's where I'd argue with you -- all of our best games, the ones where we have put up the most gaudy numbers, are the BEST teams we have played.Originally posted by Big_Blue79:
The overall SOS from something like KenPom matters a lot when you're talking about things like X/game or X/possession, because you're comparing how a team has done on each possession/game/whatever, and therefore you must consider the quality of each and every opponent (and each opponent's Off/Def rating).
Where ratings that factor in every team do not matter in SOS is when talking about how a team has fared record wise against its competition. Anyone that follows the turd-ish RPI system that the NCAA relies on knows that it's better for RPI SOS to play all middling teams rather than several great teams and several sub-300 ranked teams. That holds true, to some degree, for any comprehensive system of measuring SOS. Yet we all know that, at least for elite teams, the risk of loss to any team below a certain threshold (say, top 40-ish) is essentially nil. So what really matters to a team's record is how many (realistically) losable games they played and how they fared in those games. So a lower SOS team could have a more impressive record by virtue of playing more elite teams. This is because a W-L record is a binary measure of what happened.
TL
So what's important? What we are capable of doing to good teams, or what happens when we take a night off against bad teams?
^ a little of both. Against the best teams, sure UK is motivated and what not, but it's still small sample size theater. Was Kentucky not motivated to shoot well against Texas? Or was that just a bad shooting night? I think expanding the sample size, smartly, can give a better impression of a team's true performance level.Originally posted by Seth C:
Here's where I'd argue with you -- all of our best games, the ones where we have put up the most gaudy numbers, are the BEST teams we have played.Originally posted by Big_Blue79:
The overall SOS from something like KenPom matters a lot when you're talking about things like X/game or X/possession, because you're comparing how a team has done on each possession/game/whatever, and therefore you must consider the quality of each and every opponent (and each opponent's Off/Def rating).
Where ratings that factor in every team do not matter in SOS is when talking about how a team has fared record wise against its competition. Anyone that follows the turd-ish RPI system that the NCAA relies on knows that it's better for RPI SOS to play all middling teams rather than several great teams and several sub-300 ranked teams. That holds true, to some degree, for any comprehensive system of measuring SOS. Yet we all know that, at least for elite teams, the risk of loss to any team below a certain threshold (say, top 40-ish) is essentially nil. So what really matters to a team's record is how many (realistically) losable games they played and how they fared in those games. So a lower SOS team could have a more impressive record by virtue of playing more elite teams. This is because a W-L record is a binary measure of what happened.
TL
So what's important? What we are capable of doing to good teams, or what happens when we take a night off against bad teams?
If your argument is really "what we are capable of doing to good teams," then you're just answering a different question: what team is the best at its best. But we all know that basketball, and sports, do not work like that, especially in a 6 round single elimination tournament.
Also, your use of "all of our best games" is wrong. Is Missouri an awesome opponent? Grand Canyon? Montana State? UT-Arlington? UCLA (I'm serious, they're just not that good)? No, of course not, but that doesn't make the numbers therein meaningless, just less meaningful than if they were put up against better teams like Louisville, Kansas, and Texas. A good site like KenPom can account for some, but not all, of that and try to reduce the statistical noise that comes from only looking at a small sample of games.
I don't think anyone's saying that UK's best wins aren't better than UVA's best wins, but if I was comparing the rate stats (like offensive and defensive efficiency) I would also consider overall SOS.
The question becomes, in the context of the article, why is he stating that Kentucky "has clearly played an easier schedule"? The only meaning one can take is that it is there to suggest Kentucky's numbers are inflated by their "clearly" easier schedule. Then the question becomes, is that true?Originally posted by Big_Blue79:
^ a little of both. Against the best teams, sure UK is motivated and what not, but it's still small sample size theater. Was Kentucky not motivated to shoot well against Texas? Or was that just a bad shooting night? I think expanding the sample size, smartly, can give a better impression of a team's true performance level.Originally posted by Seth C:
Here's where I'd argue with you -- all of our best games, the ones where we have put up the most gaudy numbers, are the BEST teams we have played.Originally posted by Big_Blue79:
The overall SOS from something like KenPom matters a lot when you're talking about things like X/game or X/possession, because you're comparing how a team has done on each possession/game/whatever, and therefore you must consider the quality of each and every opponent (and each opponent's Off/Def rating).
Where ratings that factor in every team do not matter in SOS is when talking about how a team has fared record wise against its competition. Anyone that follows the turd-ish RPI system that the NCAA relies on knows that it's better for RPI SOS to play all middling teams rather than several great teams and several sub-300 ranked teams. That holds true, to some degree, for any comprehensive system of measuring SOS. Yet we all know that, at least for elite teams, the risk of loss to any team below a certain threshold (say, top 40-ish) is essentially nil. So what really matters to a team's record is how many (realistically) losable games they played and how they fared in those games. So a lower SOS team could have a more impressive record by virtue of playing more elite teams. This is because a W-L record is a binary measure of what happened.
TL
So what's important? What we are capable of doing to good teams, or what happens when we take a night off against bad teams?
If your argument is really "what we are capable of doing to good teams," then you're just answering a different question: what team is the best at its best. But we all know that basketball, and sports, do not work like that, especially in a 6 round single elimination tournament.
Also, your use of "all of our best games" is wrong. Is Missouri an awesome opponent? Grand Canyon? Montana State? UT-Arlington? UCLA (I'm serious, they're just not that good)? No, of course not, but that doesn't make the numbers therein meaningless, just less meaningful than if they were put up against better teams like Louisville, Kansas, and Texas. A good site like KenPom can account for some, but not all, of that and try to reduce the statistical noise that comes from only looking at a small sample of games.
I don't think anyone's saying that UK's best wins aren't better than UVA's best wins, but if I was comparing the rate stats (like offensive and defensive efficiency) I would also consider overall SOS.
Are Kentucky's stats against their strongest opponents bringing their numbers down, or up? If those games are bringing the numbers UP then the statement becomes irrelevant. Conversely, let's look at UVA. Are their best performances, even if there are perhaps fewer of them, against the bottom of their competition? If so, that's far more important than their average SOS.
But perhaps most importantly, if one team has the better mean opponent average, the vastly superior median opponent (because they have played FAR more great teams), has played more top 20 opponents, the same top 50 opponents, more top 100 opponents, and the same top 150 opponents, would a wise man use a metric that says the OTHER team has played a "tougher" schedule, apparently based on only one thing, them playing one fewer game against teams in the top 200-300? I don't think so.
Realistically, I don't think the author's statement was based on any factual data. I think he decided to throw in a statement based purely on conjecture and his opinion in to a piece that was supposed to be about numbers.
This post was edited on 1/22 10:34 AM by Seth C
I don't. It would be an obvious UK in 5. It would be a sweep, but our guys would get bored and lose one at their place we should not.
Gotcha. I was commenting generally and in response to the "SOS sheep" in the comments thing. Yeah, that comment seemed out of sorts. The only thing I think he could point to is that UVA has had better road wins. IIRC, UK has Louisville, TA&M, and Alabama, and UVA has Maryland (although w/out Wells), ND, I think someone else. But yeah, that's an unsupported comment.
Cupcake city is about over for UVA.
They're getting ready to face
Duke
@UNC
Louisville
@NCState
in four consecutive games.
The GA Tech, Va Tech, Boston College deal is about over with.
We'll see where they stand after that stretch. If they are still undefeated then I'll have more respect for them. Until then, sorry.
This post was edited on 1/22 1:49 PM by yabbadabbadoo
They're getting ready to face
Duke
@UNC
Louisville
@NCState
in four consecutive games.
The GA Tech, Va Tech, Boston College deal is about over with.
We'll see where they stand after that stretch. If they are still undefeated then I'll have more respect for them. Until then, sorry.
This post was edited on 1/22 1:49 PM by yabbadabbadoo
IMO, SOS should just be how many teams you played that are in the Top 100. Just leave cupcakes out of the equation altogether. I don't even like using the team's own ranking as a comparison. Who cares if Team X, who is 150th played all their games against teams between 60-99, but nobody above 60. It might be a super tough schedule for them, but not when compared to a Top 20 team.
I agree with this.
And it's why the RPI is flawed: the weight given to the poorer teams on the schedule can really anchor down the RPI.
Generally the only thing that matters when you're trying to gauge the worth of a team before the NCAA Tournament is how many good teams have they beaten?
You can read and anaylize stats on offense and defense all you want. Ive watched UVA play about 4 times this year and all i can say is UK is clearly the better team, deeper, athletic, taller, better players and id say coached. ANYTHING could happen in a one game setting but IF we come to play against them and our defense is at its best and were scoring like we can we will easily beat them. If we could play a 7 game series with them they would probably win 1 and take us to the wire in another but we would win 4-1 or 4-2.
I love Justin Anderson's game. I remember wanting him really badly as a senior in high school. Dunks with authority for a perimeter player. And he seems to have gotten much bigger as I remember watching him in High School and he was listed at 6'5 and maybe 200 pounds and now he's almost 230 and 6'6. Killer highlight reels too. I think we'd take them but somebody would have to shut him down. Poythress would usually be more than enough but no one other than WCS could really effectively cover him. He'd get 20 but we'd stop everybody else so I think we'd take them 5 out of 7 games
Your probably right, but if the Zags lose a game they will fall quite a bit I'd think. They play St Mary's tonight (13-4/4-1) though it is at home. That's probably the only team left on their schedule who can beat them. They've got a return game with them as well. Either way it probably ends up with Arizona and Gonzaga the top two seeds in the west. Of course 'Zona has a tough game tonight as well at Stanford.
UVA is really good but i think UK is much better when they play at the level they played at for all the big name teams they played this season . I think UK would win something like 69-59 .