VCU and A10 advanced metrics/rankings (NET, KenPom, Torvik)

BracketForecast

Top Member
May 9, 2011
4,456
5,142
N Mollen 110% spot on. I was about to state metrics are useless until at least halfway through.
Then you'll be happy to know that this thread doesn't disappear on January 1st. In the meantime, some of us like to break down information because we're excited basketball is almost back. If you don't like it, read another thread until you deem it worthy of your clicks.
 

Ramcounter

Top Member
Dec 7, 2011
4,368
5,197
This makes no sense to me. He "counts" freshmen if they are in the "top 30 or so," presumably even if they are playing behind an all American and will see few minutes, but a top 50 who will start is not "counted?" And how does he calculate the value of the top 30 recruit (a status that itself is almost 100% subjective)?

He's certainly as entitled as anyone else to project how he thinks teams will perform, but giving it the patina of math seems to me to be fraudulent.

BTW, he had us as #126 preseason last year and we ended up at #46. The latter number, based on actual performance and played games, by VCU and all of the other 351 D1 teams, has some value. The former number is just a WAG just like the AP top 25, but masquerading as science.
I actually took screenshots of his 4 factor rankings last year because I was curious how they changed over the course of the year. Before the Hampton game he had the Offense Adjusted Efficiency ranked 156 and the Defense Adjusted Efficiency ranked 103. The final Offense Adjusted Efficiency was ranked 177 and the final Defense Adjusted Efficiency was ranked 7th. I don’t think many people saw VCU having a top 10 defense. That said, I think it was pretty clear by December his preseason defensive ranking was off based on how many games went way short of his predicted score.

Outside of the top 30 players, how many freshman have such an impact that they would change a point spread if they couldn’t play in a game? Obviously not saying his methodology is perfect but IMHO it is logical.
 

N Mollen

Top Member
Insider
Jun 5, 2012
18,725
56,762
I actually took screenshots of his 4 factor rankings last year because I was curious how they changed over the course of the year. Before the Hampton game he had the Offense Adjusted Efficiency ranked 156 and the Defense Adjusted Efficiency ranked 103. The final Offense Adjusted Efficiency was ranked 177 and the final Defense Adjusted Efficiency was ranked 7th. I don’t think many people saw VCU having a top 10 defense. That said, I think it was pretty clear by December his preseason defensive ranking was off based on how many games went way short of his predicted score.

Outside of the top 30 players, how many freshman have such an impact that they would change a point spread if they couldn’t play in a game? Obviously not saying his methodology is perfect but IMHO it is logical.
I think you've sort of proved my point. Anyone who looked at our 17-18 results before the 18-19 season began would have been at least as pessimistic about our ability to play D as Kenpom, and in fact, that is why we got picked 7th in the conference etc. I would have said the same thing, but largely because I was afraid Rhoades was in over his head in terms of coaching defense. Boy was I wrong.

But given the departure of some players, improved health by others, new arrivals . . . there was simply no way data could give you anything like an accurate prediction of the outcome for the season. So the supposed precision embedded in a preseason Kenpom rating was mere illusion.

As for counting some recruits but not others, a top 30 recruit on Duke's roster would be riding the pine behind a top 10 from last year and a top 5 in his own class. A top 40 on VCU or Dayton would play a lot and early, and would likely have far more impact on outcomes (good or bad) than the higher rated kid on Duke or Kentucky or even UNC.

I like and admire Kenpom's work. I spend a ton of time over the course of the season poring over it. But I think that at this point, the missing variables are so profound that it makes it meaningless -- much less meaningful than a guess by someone like Andy Katz -- but simultaneously suggests by its methodology that it is precise and driven by objective criteria. It isn't.
 

Ramcounter

Top Member
Dec 7, 2011
4,368
5,197
I think you've sort of proved my point. Anyone who looked at our 17-18 results before the 18-19 season began would have been at least as pessimistic about our ability to play D as Kenpom, and in fact, that is why we got picked 7th in the conference etc. I would have said the same thing, but largely because I was afraid Rhoades was in over his head in terms of coaching defense. Boy was I wrong.

But given the departure of some players, improved health by others, new arrivals . . . there was simply no way data could give you anything like an accurate prediction of the outcome for the season. So the supposed precision embedded in a preseason Kenpom rating was mere illusion.

As for counting some recruits but not others, a top 30 recruit on Duke's roster would be riding the pine behind a top 10 from last year and a top 5 in his own class. A top 40 on VCU or Dayton would play a lot and early, and would likely have far more impact on outcomes (good or bad) than the higher rated kid on Duke or Kentucky or even UNC.

I like and admire Kenpom's work. I spend a ton of time over the course of the season poring over it. But I think that at this point, the missing variables are so profound that it makes it meaningless -- much less meaningful than a guess by someone like Andy Katz -- but simultaneously suggests by its methodology that it is precise and driven by objective criteria. It isn't.
I guess it depends on what one’s definition of accurate is. Unless you have Biff Tannen’s sports almanac isn’t it all ultimately just a prediction?
 
  • Like
Reactions: RAM7889

Havoc2Ut2

Top Member
Nov 15, 2014
12,170
19,262
I think you've sort of proved my point. Anyone who looked at our 17-18 results before the 18-19 season began would have been at least as pessimistic about our ability to play D as Kenpom, and in fact, that is why we got picked 7th in the conference etc. I would have said the same thing, but largely because I was afraid Rhoades was in over his head in terms of coaching defense. Boy was I wrong.

But given the departure of some players, improved health by others, new arrivals . . . there was simply no way data could give you anything like an accurate prediction of the outcome for the season. So the supposed precision embedded in a preseason Kenpom rating was mere illusion.

As for counting some recruits but not others, a top 30 recruit on Duke's roster would be riding the pine behind a top 10 from last year and a top 5 in his own class. A top 40 on VCU or Dayton would play a lot and early, and would likely have far more impact on outcomes (good or bad) than the higher rated kid on Duke or Kentucky or even UNC.

I like and admire Kenpom's work. I spend a ton of time over the course of the season poring over it. But I think that at this point, the missing variables are so profound that it makes it meaningless -- much less meaningful than a guess by someone like Andy Katz -- but simultaneously suggests by its methodology that it is precise and driven by objective criteria. It isn't.
So, I guess us being in preseason top 30's voted on by a bunch of guys and gals means nothing too. There are few predictive tools out there better than KP. His model gets it right more often than not at this point preseason. Yes, there are a bunch of outliers. I'd expect that given there are over 350 teams in the pool.
 

N Mollen

Top Member
Insider
Jun 5, 2012
18,725
56,762
So, I guess us being in preseason top 30's voted on by a bunch of guys and gals means nothing too. There are few predictive tools out there better than KP. His model gets it right more often than not at this point preseason. Yes, there are a bunch of outliers. I'd expect that given there are over 350 teams in the pool.
As I said above, IMHO until there are data available, data approaches aren't really useful. Pick-your-analyst can take account of an incoming freshman's skills (having watched them play or talked to those who have), consider the coach's scheme, analyze the remainder of the team and project how the freshman will fit in. The analyst can look at new coaching situations, surgeries over the off-season, and rumored growth by last year's players. So at this stage I'd say that a guess (and it is nothing more than that) by Goodman or Katz or Rothstein has more value to me than a data-set that is so incomplete.

After a few weeks of playing actual games, Kenpom becomes the more useful tool, and by the end of the season, the polls and analysts guesswork is of no value to me. At that point Kenpom (and similar data-drive approaches) become vastly superior
 

GuardTheArc

Top Member
Dec 4, 2014
524
815
As I said above, IMHO until there are data available, data approaches aren't really useful. Pick-your-analyst can take account of an incoming freshman's skills (having watched them play or talked to those who have), consider the coach's scheme, analyze the remainder of the team and project how the freshman will fit in. The analyst can look at new coaching situations, surgeries over the off-season, and rumored growth by last year's players. So at this stage I'd say that a guess (and it is nothing more than that) by Goodman or Katz or Rothstein has more value to me than a data-set that is so incomplete.

After a few weeks of playing actual games, Kenpom becomes the more useful tool, and by the end of the season, the polls and analysts guesswork is of no value to me. At that point Kenpom (and similar data-drive approaches) become vastly superior
Regarding this and all threads, you are not required to read them, approve of them, or respond to them.

You can go do other things with your life that you find more important/valuable.

The KP data in early Oct is merely a starting point. Most folks understand this.
 
  • Like
Reactions: BracketForecast
May 3, 2012
1,747
1,745
As I said above, IMHO until there are data available, data approaches aren't really useful. Pick-your-analyst can take account of an incoming freshman's skills (having watched them play or talked to those who have), consider the coach's scheme, analyze the remainder of the team and project how the freshman will fit in. The analyst can look at new coaching situations, surgeries over the off-season, and rumored growth by last year's players. So at this stage I'd say that a guess (and it is nothing more than that) by Goodman or Katz or Rothstein has more value to me than a data-set that is so incomplete.

After a few weeks of playing actual games, Kenpom becomes the more useful tool, and by the end of the season, the polls and analysts guesswork is of no value to me. At that point Kenpom (and similar data-drive approaches) become vastly superior
Kenpom, I think, even says it's not accurate until it compiles enough data for patterns to emerge. It's usually the very early part of conference play until it's useful.
Early on, it's like filling up with gas at the top of a hill and driving home to the bottom of the hill. Your car registers 70 mpg. You know that's not going to be the actual eventual reading. So, IF we played a D III school in exhibition and won by 50, none of the metrics would mean anything.
The only thing I see in his initial ranking is that he expects our defense to stay where it has been in the past.
 
  • Like
Reactions: Havoc2Ut2

RESRam

Insider
Apr 21, 2009
1,032
834
Wasn’t sure where to post this and I’ll confess upfront to not being objective about UR, but lots of folks on their boards are expecting them to make the NCAA tournament as an at large or at least go to NIT. Maybe it’s just me, but does that seem delusional to anyone else? They still have Mooney, thankfully, as their coach.
 
Last edited:

vcubob

Insider
Apr 19, 2009
1,690
1,529
Wasn’t sure where to post this and I’ll confess upfront to not being objective about UR, but lots of folks on their boards are expecting them to make the NCAA tournament as an at large or at least go to NIT. Maybe it’s just me, but does that seem delusional to anyone else? They still have Mooney as their coach.
If they stay heathy a top 4 A10 finish and NIT bid is "not out of the question".
 

GuardTheArc

Top Member
Dec 4, 2014
524
815
Wasn’t sure where to post this and I’ll confess upfront to not being objective about UR, but lots of folks on their boards are expecting them to make the NCAA tournament as an at large or at least go to NIT. Maybe it’s just me, but does that seem delusional to anyone else? They still have Mooney as their coach.
Anything is possible (like winning the A-10 tourney) but in order for Richmond to dance as an at-large, they’ll likely need to do a few historical things:

1) Improve by about 11-12 wins and 150-175 spots in the KP defensive rankings and overall rankings (i.e. make historical leaps in efficiency) with essentially the same roster
2) Accomplish the above without a senior in their rotation, nearly unheard of for a non-P6
3) Leapfrog one or several of VCU, Davidson, Dayton, and Rhody — all of which have nearly if not all of their team back from last yr, including several seniors each.
 

VCU Heel

Top Member
Aug 7, 2012
10,280
10,587
Wasn’t sure where to post this and I’ll confess upfront to not being objective about UR, but lots of folks on their boards are expecting them to make the NCAA tournament as an at large or at least go to NIT. Maybe it’s just me, but does that seem delusional to anyone else? They still have Mooney, thankfully, as their coach.
Delusional or not, every team and fanbase should expect that kind of success before the season. As a fan, why even bother watching if you expect to lose?
 
  • Like
Reactions: BaNgMyPrOgRaM