NET prediction of Final Four

For discussion of Vanderbilt Commodores men's basketball games and recruiting.

Moderators: kerrigjl, BrentVU, jfgogold, NateSY, KarenYates, Vandyman74, roanoke, VandyWhit

User avatar
charlestonalum
Fleet Admiral
Posts: 13165
Joined: Sat Nov 19, 2016 5:37 am
Location: Charleston, SC
Has thanked: 101 times
Been thanked: 81 times
Contact:

NET prediction of Final Four

Post by charlestonalum »

just curious what the net ranking of the teams in the final four was?


Labradore
Lieutenant
Posts: 154
Joined: Sun Dec 04, 2016 8:27 am
Been thanked: 9 times

Re: NET prediction of Final Four

Post by Labradore »

Ummm… it would appear that the NET rankings are remarkably on target, with Miami interestingly being the only outlier / Cinderella of the teams remaining. The NET says the other teams are doing exactly what their season performance would have predicted.

I think I hate the NET… but I also have wanted teams to be rewarded for being better and not simply given favor for seed and tournament entry based on blue blood status. Honestly, though it’s painful to say it, Vanderbilt should have beaten the teams it lost to prior to the sec season. We probably got the outcome we deserved. The NET, by all appearances, seems stunningly fair. (Though I can’t explain the Vols #4 ranking; that’s nonsense).

https://www.ncaa.com/rankings/basketbal ... t-rankings
User avatar
charlestonalum
Fleet Admiral
Posts: 13165
Joined: Sat Nov 19, 2016 5:37 am
Location: Charleston, SC
Has thanked: 101 times
Been thanked: 81 times
Contact:

Re: NET prediction of Final Four

Post by charlestonalum »

thanks and I see by NET we should have lost by a lot to UAB
User avatar
mathguy
Rear Admiral
Posts: 2046
Joined: Sat Nov 19, 2016 1:27 pm
Has thanked: 44 times
Been thanked: 31 times

Re: NET prediction of Final Four

Post by mathguy »

Labradore wrote: Sun Mar 26, 2023 9:10 am Ummm… it would appear that the NET rankings are remarkably on target, with Miami interestingly being the only outlier / Cinderella of the teams remaining. The NET says the other teams are doing exactly what their season performance would have predicted.

I think I hate the NET… but I also have wanted teams to be rewarded for being better and not simply given favor for seed and tournament entry based on blue blood status. Honestly, though it’s painful to say it, Vanderbilt should have beaten the teams it lost to prior to the sec season. We probably got the outcome we deserved. The NET, by all appearances, seems stunningly fair. (Though I can’t explain the Vols #4 ranking; that’s nonsense).

https://www.ncaa.com/rankings/basketbal ... t-rankings
So here's the ultimate debate. The NET is supposed to be a "predictive measure". So it's supposed to be rating teams based on who we think would beat who. This is DIFFERENT than rating teams based on who has the best resume or who is most deserving based on wins and losses.

So which would people rather see? Selection based more on predictive criteria? Or selection based more on traditional/reactive criteria?

I'd prefer the latter myself.
Obvious
Captain
Posts: 592
Joined: Sat Feb 04, 2023 7:51 pm
Has thanked: 34 times
Been thanked: 24 times

Re: NET prediction of Final Four

Post by Obvious »

UConn - 8
FAU - 13
SDSU - 14
Miami - 35

Texas, Gonzaga, Creighton all ranked pretty high in the NET. Kansas state was 24.
UltimateVUFan
Vice Admiral
Posts: 3143
Joined: Fri Dec 23, 2016 9:25 am
Has thanked: 163 times
Been thanked: 94 times

Re: NET prediction of Final Four

Post by UltimateVUFan »

The problem with the NET as the basis for NCAAT selection is that it doesn’t take into account who you beat and the quantitative value of your record. There was literally no conceivable way to argue that MS State (or Pitt, for that matter) deserved to be in the play-in game over VU.

Even UK’s resume by NET wins/losses was not altogether dissimilar from VU’s. Both had 2 quad 3 losses and 1 quad 4 loss. UK fared better in Q1/2 win-loss record at 12-8 vs. VU’s 10-11, though they suffered a Q2 loss, which VU avoided. VU also played 16 Q1 games vs. UK’s 13, so the tougher schedule can partially account for the slight discrepancy in the Q1/Q2 variation. Yet, at the end of the day, UK sits with a NET of 26 while VU is at 81. Shouldn’t the fact that a team actually won the games (i.e., 2 vs. UK and a whole lot more than other SEC teams taken in the tourney) matter more than whether or not they theoretically SHOULD win the games???
I’m sorry. The NET is a crock.

EDIT: Heck, even Florida - THREE AND FIFTEEN (3-15) in Q1/2 games Florida - YES, the Florida that VU mopped the floor with TWICE - sits at #62 in the NET. Nearly 20 spots ahead of VU. Results apparently mean next to nothing.
Jason94
Admiral
Posts: 6121
Joined: Wed Nov 23, 2016 10:15 am
Has thanked: 2 times
Been thanked: 56 times

Re: NET prediction of Final Four

Post by Jason94 »

UltimateVUFan wrote: Sun Mar 26, 2023 9:10 pm The problem with the NET as the basis for NCAAT selection is that it doesn’t take into account who you beat and the quantitative value of your record. There was literally no conceivable way to argue that MS State (or Pitt, for that matter) deserved to be in the play-in game over VU.

Even UK’s resume by NET wins/losses was not altogether dissimilar from VU’s. Both had 2 quad 3 losses and 1 quad 4 loss. UK fared better in Q1/2 win-loss record at 12-8 vs. VU’s 10-11, though they suffered a Q2 loss, which VU avoided. VU also played 16 Q1 games vs. UK’s 13, so the tougher schedule can partially account for the slight discrepancy in the Q1/Q2 variation. Yet, at the end of the day, UK sits with a NET of 26 while VU is at 81. Shouldn’t the fact that a team actually won the games (i.e., 2 vs. UK and a whole lot more than other SEC teams taken in the tourney) matter more than whether or not they theoretically SHOULD win the games???
I’m sorry. The NET is a crock.

EDIT: Heck, even Florida - THREE AND FIFTEEN (3-15) in Q1/2 games Florida - YES, the Florida that VU mopped the floor with TWICE - sits at #62 in the NET. Nearly 20 spots ahead of VU. Results apparently mean next to nothing.
How many times does it have to be explained before you understand what the NET is? FAU prior to the tournament had less quality wins (6-2 vs 10-11) and played a lot fewer quality opponents than we did, so are we better than them? Was the NET wrong there? There are no extra points for wins in the NET - as someone accurately pointed out, a team could theoretically play a schedule of only the top 5 teams on the road and loss 30 times by a single point and be considered a top 5 team by the NET. Conversely, a team could play a schedule of the worst 5 teams all at home, win each game by 1 point and be considered one of the worst teams by the NET.

I don't understand how people can look at a part of the season for a team, and compare that to another part of the season of a different team, and say that a rating system that looks at all of the games is wrong. If we take our best 15 games and compare them to another team's not best 15 games, then we will look better most of the time.

I mean, are you intentionally trying to misunderstand what the NET does? Why compare W/L records when this something the NET specifically is not doing? I've yet to see someone who criticizes the NET explain how we determine relative strength of opponents for all 350+ D1 teams. How do you know how good UF's schedule was? Did you watch all 30+ games that UF played and watched each of the 30+ games that their opponents played? It is embarrassing to see a bunch of Vandy fans/grads throw out really uninformed statements because the NET captured how we did early on. I get that we are mad at how the NET is being used, but it is a tool. Don't act like November and December didn't happen or get mad because the NET doesn't also ignore it.

Note that while the RPI had us at 39th, it also had UConn at 26th prior to the tournament. I don't see how only 13 spot could separate us and UCONN.
Go Vandy!
Rear Admiral
Posts: 2249
Joined: Wed Aug 22, 2018 9:02 pm
Has thanked: 312 times
Been thanked: 47 times

Re: NET prediction of Final Four

Post by Go Vandy! »

...So you're saying we should have been in? ;)
User avatar
mathguy
Rear Admiral
Posts: 2046
Joined: Sat Nov 19, 2016 1:27 pm
Has thanked: 44 times
Been thanked: 31 times

Re: NET prediction of Final Four

Post by mathguy »

Jason94 wrote: Sun Mar 26, 2023 10:30 pm Why compare W/L records when this something the NET specifically is not doing?
Because, flat out, this is something that the NET *should* be doing if the committee is going to use it as a selection criteria.

I understand the math. I understand the statistics. I understand the inherent difficulties in trying to objectively determine quality of play with limited common opponents.

But at the end of the day - you play the game to win the game. That does not that a 28-2 team that feasted on cupcakes is necessarily better than a 19-11 team from a tough conference ... but yeah, wins need to be considered. If the selection committee looks at Q1 wins and Q4 losses, but their computerized rating system *doesn't* ... than what good is the rating system for the established criteria?

What other options are there? Could we get a 365-day rolling ELO rating on teams? Could we have something that factors in margin of victory, but caps it out (since I'm pretty sure the Bama loss probably cost us about 12 spots in the NET which, IMO, is blatantly unfair)? It was imperfect, but the RPI served well for years. You could have a rating that does something like count your Q1 games for 40% of your rating, your Q2 games for 30%, your Q3 games for 20% and your Q4 games for 10% ... because let's be honest, the most important part of your resume is how you perform against the best teams, not how you perform against the worst. Then, for the small conference teams (that IMO always get the shaft) you aren't looking and giving some BS excuse like "well they only had 2 Q1 wins" (when they can't get on anyone's schedule) ... instead you can say ... hey! they were 2-1 with a .667 winning percentage in Q1 games (which BTW describes FAU this year). In contrast, Oral Roberts (46 spots ahead of VU in the NET) was a whopping 0-4 in Q1 games (and 1-4 in Q1/2 games). So looking at that we *know* that this year's version of ORU seems unable to win big games against good teams. So why not penalize them for that?

So yeah ... there are lots of options about what we *could* do instead of the NET ... but instead we apparently defer to a computer system that, in your description, specifically does not consider one of the major selection criteria the committee is asked to look at!!
UltimateVUFan
Vice Admiral
Posts: 3143
Joined: Fri Dec 23, 2016 9:25 am
Has thanked: 163 times
Been thanked: 94 times

Re: NET prediction of Final Four

Post by UltimateVUFan »

Jason94 wrote: Sun Mar 26, 2023 10:30 pm How many times does it have to be explained before you understand what the NET is?

I don't understand how people can look at a part of the season for a team, and compare that to another part of the season of a different team, and say that a rating system that looks at all of the games is wrong. If we take our best 15 games and compare them to another team's not best 15 games, then we will look better most of the time.

I mean, are you intentionally trying to misunderstand what the NET does? Why compare W/L records when this something the NET specifically is not doing? I've yet to see someone who criticizes the NET explain how we determine relative strength of opponents for all 350+ D1 teams. How do you know how good UF's schedule was? Did you watch all 30+ games that UF played and watched each of the 30+ games that their opponents played? It is embarrassing to see a bunch of Vandy fans/grads throw out really uninformed statements because the NET captured how we did early on. I get that we are mad at how the NET is being used, but it is a tool. Don't act like November and December didn't happen or get mad because the NET doesn't also ignore it.
I understand the NET. I know it both calculates results based on relative offensive/defensive efficiency AND serves as a "predictive" metric (theoretically). I was being marginally facetious in saying that "Results apparently mean next to nothing." To elaborate on that claim: the NET would predict that #81 VU on a neutral court and road court has very little chance of beating #26 UK. Yet, both came to pass. Likewise, VU would not be predicted to beat #62 (!?!?!?! Again, 3-15 in Q1/Q2 games) Florida on their home court. They did so pretty much "going away" (which I believe were roughly your exact words about that win back in a post when we fans were excited about the improbable late season tear VU went on).

I have made clear numerous times in my posts that I simply disagree with the NET as a basis for NCAAT selection. I think it was exceedingly obvious that VU was a better team than several SEC squads that made the cut ahead of them. For this season that is clearly a moot point, but moving forward, I don't think it is unreasonable to consider the possibility that a more well-rounded approach to selection be considered, especially when it comes to sending x number of teams from y conference. Conference rankings should bear some weight in the discussion if the committee knows a certain number of teams from that conference are worthy of inclusion based on the strength of that conference. You know, maybe using some common sense evaluation of how the games actually played out instead of relying on what a computer metric predicts might happen.

Lastly, I highlighted some passages above for your consideration. Condescending a fellow VU fan really isn't a very good look. I have remarked on many occasions that I value your perspective and generally enjoy the content of your posts. However, there really is no reason to think that you are the smartest man on the board. Consequently, the types of comments made above are exactly the types of remarks that cause many non-Vanderbilt fans to think a bit poorly of us as pompous or arrogant. Just throwing it out there. And again, I am totally sincere in saying that I appreciate your contributions to the board, but you might consider a little humility at times.
Obvious
Captain
Posts: 592
Joined: Sat Feb 04, 2023 7:51 pm
Has thanked: 34 times
Been thanked: 24 times

Re: NET prediction of Final Four

Post by Obvious »

UltimateVUFan wrote: Mon Mar 27, 2023 5:38 pm
Jason94 wrote: Sun Mar 26, 2023 10:30 pm How many times does it have to be explained before you understand what the NET is?

I don't understand how people can look at a part of the season for a team, and compare that to another part of the season of a different team, and say that a rating system that looks at all of the games is wrong. If we take our best 15 games and compare them to another team's not best 15 games, then we will look better most of the time.

I mean, are you intentionally trying to misunderstand what the NET does? Why compare W/L records when this something the NET specifically is not doing? I've yet to see someone who criticizes the NET explain how we determine relative strength of opponents for all 350+ D1 teams. How do you know how good UF's schedule was? Did you watch all 30+ games that UF played and watched each of the 30+ games that their opponents played? It is embarrassing to see a bunch of Vandy fans/grads throw out really uninformed statements because the NET captured how we did early on. I get that we are mad at how the NET is being used, but it is a tool. Don't act like November and December didn't happen or get mad because the NET doesn't also ignore it.
I understand the NET. I know it both calculates results based on relative offensive/defensive efficiency AND serves as a "predictive" metric (theoretically). I was being marginally facetious in saying that "Results apparently mean next to nothing." To elaborate on that claim: the NET would predict that #81 VU on a neutral court and road court has very little chance of beating #26 UK. Yet, both came to pass. Likewise, VU would not be predicted to beat #62 (!?!?!?! Again, 3-15 in Q1/Q2 games) Florida on their home court. They did so pretty much "going away" (which I believe were roughly your exact words about that win back in a post when we fans were excited about the improbable late season tear VU went on).

I have made clear numerous times in my posts that I simply disagree with the NET as a basis for NCAAT selection. I think it was exceedingly obvious that VU was a better team than several SEC squads that made the cut ahead of them. For this season that is clearly a moot point, but moving forward, I don't think it is unreasonable to consider the possibility that a more well-rounded approach to selection be considered, especially when it comes to sending x number of teams from y conference. Conference rankings should bear some weight in the discussion if the committee knows a certain number of teams from that conference are worthy of inclusion based on the strength of that conference. You know, maybe using some common sense evaluation of how the games actually played out instead of relying on what a computer metric predicts might happen.

Lastly, I highlighted some passages above for your consideration. Condescending a fellow VU fan really isn't a very good look. I have remarked on many occasions that I value your perspective and generally enjoy the content of your posts. However, there really is no reason to think that you are the smartest man on the board. Consequently, the types of comments made above are exactly the types of remarks that cause many non-Vanderbilt fans to think a bit poorly of us as pompous or arrogant. Just throwing it out there. And again, I am totally sincere in saying that I appreciate your contributions to the board, but you might consider a little humility at times.
I learned in research that statistics can be interpreted anyway you want them to be. Ten or so men on the selection committee look at some stats and determine the field of 68. Do they have a background in basketball? Maybe. Do I like the NET? No. It is okay for others to dislike the NET? Yea. Will it change anything? No. Does anything we say on here mean anything? Not really. But it’s cool to talk to you all about Vanderbilt basketball.
Obvious
Captain
Posts: 592
Joined: Sat Feb 04, 2023 7:51 pm
Has thanked: 34 times
Been thanked: 24 times

Re: NET prediction of Final Four

Post by Obvious »

Jason I think you’re one of those statistics people and you want to use it to make yourself look smarter than everyone. I talked about Ezra being a good defender based on what I saw. You gave me some statistics as to why he’s not really a good defender. Stackhouse and sec analysts think he’s a tough, relentless defender. You also admitted that you didn’t see the game that I was referring to but had a long statistical post about defense. You only want to discuss the NET probably because you can look at stats. Let people gripe about the NET. We were definitely one of the top 68 teams. I accept that the NET had us rated differently based on November and December games.
Locked Previous topicNext topic