This Week in the ACC - 1/24-1/30

[CameronDuke]

Yep, they are pulling a 2021 Duke right now: losing all of their close games (2-5 in games decided by 5 points or less) so far. KenPom has them at #353 (out of 358) in the nation in "luck" (which is just the difference in actual win% from expected win%).

Yeah, they are an outlier for sure. The next highest 9 loss team on Kenpom is Utah State at 71.

[Norman Pfyster]

Yep, they are pulling a 2021 Duke right now: losing all of their close games (2-5 in games decided by 5 points or less) so far. KenPom has them at #353 (out of 358) in the nation in "luck" (which is just the difference in actual win% from expected win%).

A cynical way of understanding KenPom's "luck" factor is it is just a pathological refusal to consider that the model is wrong.

[CDu]

Yeah, they are an outlier for sure. The next highest 9 loss team on Kenpom is Utah State at 71.

Yeah, Torvik has them a little lower at #50, but still comfortably above the next 9-loss team. They have them as just slightly less of an outlier (#342).

Basically, the stats guys think VT has played like a ~7 loss team in aggregate, which makes sense. The computers think that VT should have at least one more win in close games. For whatever reason, Tech hasn't translated those generally dominant performances against teams outside the top-90 into more success.

But really, it's "just" the loss at BC and at home vs NC State that are anomalous. The rest is just that their schedule of competitive games has been skewed towards likely losses - either against very good teams, or on the road (or both).

I think they probably are more like a team in the 50s or 60s nationally, and probably a ~.500 ACC team in quality. But they overperformed in their out of conference schedule, which inflated their rating. They have the two head-scratching losses to NC State and @BC, which is the difference between their actual record and what their record "should" be given their overall quality.

[camion]

A cynical way of understanding KenPom's "luck" factor is it is just a pathological refusal to consider that the model is wrong.

Interesting that "pathological" includes "logical path." And following a logical path too far can be pathological.



Or maybe it's not really interesting.

[hallcity]

Chris Mack may be getting fired tomorrow. Duke plays Louisville on Saturday.

[JasonEvans]

Chris Mack may be getting fired tomorrow. Duke plays Louisville on Saturday.

I'd rather they wait a few days... Mack's handling of this team has been horribly uninspired thus far. I'd hate for them to suddenly feel free and relaxed and actually play up to their abilities.

[CameronDuke]

Current chances to make the NCAA Tournament:

VT: 10-9 (2-6 in ACC)

BPI: 33.1%
TeamRankings: 13.9%
Barttorvik: 5.2%

UVA: 12-8 (6-4 in ACC)

BPI: 0.7%
TeamRankings: 6.1%
Barttorvik: 1.8%

[Dr. Maturin]

Miami fired their football coach just before a certain game that, well, I'm not ready to talk about yet. It's only been 6.5 years. Hopefully Louisville won't get a similar "new coach boost" if such a thing even exists.

[tommy]

Yeah, Torvik has them a little lower at #50, but still comfortably above the next 9-loss team. They have them as just slightly less of an outlier (#342).

Basically, the stats guys think VT has played like a ~7 loss team in aggregate, which makes sense. The computers think that VT should have at least one more win in close games. For whatever reason, Tech hasn't translated those generally dominant performances against teams outside the top-90 into more success.

But really, it's "just" the loss at BC and at home vs NC State that are anomalous. The rest is just that their schedule of competitive games has been skewed towards likely losses - either against very good teams, or on the road (or both).

I think they probably are more like a team in the 50s or 60s nationally, and probably a ~.500 ACC team in quality. But they overperformed in their out of conference schedule, which inflated their rating. They have the two head-scratching losses to NC State and @BC, which is the difference between their actual record and what their record "should" be given their overall quality.

I respect KenPom and Torvik and the others at the top of this field. Their work is generally very good and I certainly appreciate getting the benefit of their work. In this case, though, the formula doesn't make so much sense, at least not to me. If what you're meaning by "overperforming" in their out of conference schedule you mean the Hokies blew out terrible teams by even more than they were expected to blow them out by, I think that is being overweighted. I mean really, should they really get so much extra credit for beating Cornell, Merrimack, Radford, St. Francis, and Maine by 30+ points, assuming they were, say, 20 point favorites? Who cares? What does that prove, given that those teams are all around 300 in the NET?

This team has a home win over Notre Dame and a neutral floor win over St. Bonaventure, both of which are Q2 wins. That's it. And some ugly losses. I just don't see how them being rated in the 50s makes any sense at all. Sure they may get there if they turn their season around and start winning, you know, games. But based on their actual performance on the floor so far, they stink.

[CDu]

Current chances to make the NCAA Tournament:

VT: 10-9 (2-6 in ACC)

BPI: 33.1%
TeamRankings: 13.9%
Barttorvik: 5.2%

UVA: 12-8 (6-4 in ACC)

BPI: 0.7%
TeamRankings: 6.1%
Barttorvik: 1.8%

This is largely a function of the metrics thinking VT is the better team, and thus they have a better chance of winning the ACC tourney.

Torvik puts both teams' at-large chances at 0.0%. They think VT and UVa will both end up 10-10 in conference, mainly due to UVa having a tough remaining schedule and VT having an easy one. From there, it's simply that they think VT has a slightly higher chance of winning the ACC tournament.

TeamRankings thinks VT has an 8.9% chance of winning the ACC tourney vs a 2.5% chance of UVa winning it. They are probably a bit overoptimistic about either team's chances at an at-large bid (~4-5% for both).

For what it's worth, I think you're misreading the BPI %s. That is their projected chance of getting to the second round of the NCAA tournament, not their chances of making the NCAA tournament. So they actually think VT's chances of making the tournament are MUCH higher than that (they have them as a ~9-10 seed). But BPI is crap to begin with, so I'd just ignore it. I think that Torvik and TeamRankings provide a nice upper and lower bound of probability of getting an at-large bid for both teams (0-5%).

[SavDukeGrad]

Miami fired their football coach just before a certain game that, well, I'm not ready to talk about yet. It's only been 6.5 years. Hopefully Louisville won't get a similar "new coach boost" if such a thing even exists.

That was the first thing I thought of, as well.

[OldPhiKap]

A cynical way of understanding KenPom's "luck" factor is it is just a pathological refusal to consider that the model is wrong.

Always speak truth to power.

[MChambers]

I'd rather they wait a few days... Mack's handling of this team has been horribly uninspired thus far. I'd hate for them to suddenly feel free and relaxed and actually play up to their abilities.

Hate that new coach effect. Probably win 2 or 3 games before settling back to mediocre.

[CDu]

I respect KenPom and Torvik and the others at the top of this field. Their work is generally very good and I certainly appreciate getting the benefit of their work. In this case, though, the formula doesn't make so much sense, at least not to me. If what you're meaning by "overperforming" in their out of conference schedule you mean the Hokies blew out terrible teams by even more than they were expected to blow them out by, I think that is being overweighted. I mean really, should they really get so much extra credit for beating Cornell, Merrimack, Radford, St. Francis, and Maine by 30+ points, assuming they were, say, 20 point favorites? Who cares? What does that prove, given that those teams are all around 300 in the NET?

I agree. Which is why I said that they're probably being overrated by those models. It's just really hard to address something like this, because the models are literally just taking per-possession results and adjusting for strength of opponent. There isn't really an easy way to adjust mathematically for "well, so and so overperformed against awful teams" without blowing up the model. There will always be outliers in any formula. You just hope that, by the end of the season, enough games are played to largely tease out that noise.

This team has a home win over Notre Dame and a neutral floor win over St. Bonaventure, both of which are Q2 wins. That's it. And some ugly losses. I just don't see how them being rated in the 50s makes any sense at all. Sure they may get there if they turn their season around and start winning, you know, games. But based on their actual performance on the floor so far, they stink.

Despite agreeing with you that they are overrated, I also don't think they're as bad as their record suggests either. They have a loss at home to NC State and the loss at BC as their bad losses. They got blown out by Wake, but Wake appears to be actually good. Other than that, they've lost almost all of their close games. If they were 4-4 in conference instead of 2-6 - and that could quite easily be the case - I don't think anyone would blink at a rating of around 50. It's just that they've been on the wrong bounce of the ball 2-3 extra times this year, and that has their record looking really rough.

Tech has played like the #92 team since their pre-conference games ended (interestingly, the top 7 in the ACC post-interconference play have been Duke, Miami, Wake, Notre Dame, UNC, and... NC State), so they have been dropping over the past 7 games (from a peak of #21, down to #50 right now). But their schedule is still much friendlier in the back half, so they are likely to see a decent boost in their record.

[DU82]

Miami fired their football coach just before a certain game that, well, I'm not ready to talk about yet. It's only been 6.5 years. Hopefully Louisville won't get a similar "new coach boost" if such a thing even exists.

Don’t squib it: kick it into the end zone.

Oops, wrong thread. 😁

[CDu]

Despite agreeing with you that they are overrated, I also don't think they're as bad as their record suggests either. They have a loss at home to NC State and the loss at BC as their bad losses. They got blown out by Wake, but Wake appears to be actually good. Other than that, they've lost almost all of their close games. If they were 4-4 in conference instead of 2-6 - and that could quite easily be the case - I don't think anyone would blink at a rating of around 50. It's just that they've been on the wrong bounce of the ball 2-3 extra times this year, and that has their record looking really rough.

Also worth noting that one of their two bad losses (the home loss to NC State) is beginning to look less bad - not unlike our loss at home to Miami. NC State has been playing like a top-50ish team since 2022 began. If they continue to move up, that Q3 home loss could potentially turn into a Q2 loss, and the Hokies' Q2 road win at State a Q1 win.

[flyingdutchdevil]

https://www.espn.com/mens-college-ba...ck-sources-say

[CameronDuke]

Down 52-45 at Pitt.

7:38 left.

[SavDukeGrad]

Pitt with the win. Good for Capel.

[OldPhiKap]

Pitt with the win. Good for Capel.

Yup. And a bad Southern Swing for Team Boeheim.