ADVERTISEMENT

What we learned from the selection committee...

not sure what to think of conference record, they say it does not matter but with ND and RU it may have.
I know you were very firm about conference record not mattering, but I think that is true to an extent. Committee won't care if you finished 6th and the 7th place team has a better resume.

Leaving out the 3rd place ACC team or 4th place B1G team is a little different and it impacts perception
 
  • Like
Reactions: ColonelRutgers
Didn't Palm call us officially the most annoying fanbase on Twitter the other day? If true then ya gotta think he may have actually left us off just to piss us off, for clicks and attention, etc. Sounds crazy I know since it messed with him predicting perfectly, but you never know with idiots like him.
Predict perfectly? Its not just about getting all 68 teams. Palm seeded only 32 of the 68 teams on the correct line. That is atrocious, and even worse when you realize that his information is the one being shared on CBS during the big games on the final weekend before Selection Sunday!

Correct Seed Percentage:
Palm (CBS) - 32/68 = 47%
Lunardi (ESPN) - 38/68 = 56%
DeCourcy (FOX) - 50/68 = 74%

DeCourcy by far had the best year of the three. Some above mentioned Brad Wachtel as well. He went 49/68 = 72% with exact seeding. He would be a great candidate to replace Palm on CBS.
 
what a learned is... short of winning (or AT LEAST winning 3)...conference tourneys mean little....

don't schedule cupcakes - because they don't help in any way (just play some pre-season or exempt games) and Do Good in conference...
 
what a learned is... short of winning (or AT LEAST winning 3)...conference tourneys mean little....

don't schedule cupcakes - because they don't help in any way (just play some pre-season or exempt games) and Do Good in conference...
I think they matter if you lose early to a bad team (they can knock you out) or obviously (if you win it all). Seems like the committee did not put as much stock in winning these games as it did in winning the regular season games.
 
unless you are close to the bubble like Indiana, Miami, BYU, SMU. Xavier, Wake were, it really does not matter. Otherwise if you are coming from way back you better win the whole thing
 
  • Like
Reactions: Scangg
BAC, some great summary takeaways.
in my opinion , despite 1 or 2 seeding errors the committee did a good job. Much much better than the last 3-4 or forever years. For years the seeding was truly terrible trying to set up TV matchups that made no sense. Playing mid majors against each other instead of the Power 6 against a mid major which is what the tourney was about and miss seeding teams all the time.

I do think the Committee from their interviews over the weekend advised they had most of the field selected except 6-7 teams. I really think this is a true and more honest way to select teams. You go through an entire season and finish whatever place and with whatever wins in your conference and it should matter. One team struggling down the stretch winning 4 games in 4 days should not wipe out a season. We earned our 4th seed with 12 out of 20 BIG 10 games , most against the top half of the conference (7-4). That meant something and identified a worthy team. I cannot justify Michigan ahead of us in any way because they finished 10-10 , 2 games behind us and did not play 1,23 two times each like we did . Yeah their out of conference was a tougher but that is a relative term. They got blown out by a very good Arizona team , blown out by a mediocre UNC team , almost blown out by UCF a non tourney decent team. They beat San Diego State , who got better later and was a tourney team and beat Buffalo , UnLV and Tarelton State , all non tourney teams. If you are hanging your hat on beating San Diego State and getting blown out by the 3 other losses , something is wrong.

If a team gets hot , then if they weren’t in or that close to getting in at the end of the regular season , then they had to win the conference tourney. Va Tech if they lost to Duke would have and should have been out and completed the 4 in 4 to earn a bid. Texas A& M did not. ( Although personally thought their wins in the SEC tourney were more impressive , a much better conference , and I thought there were in on Saturday and would have bumped someone ( Indiana , Wyoming , Notre Dame. ).

Also , the strength of conferences has to play a role. The Committee selected 9 BIG 10 teams and 5 ACC teams ( a gift ) . Rutgers finishing 4 th in conference with a 12 -8 record is light years ahead of Notre Dame finishing second at 15–5 in an absolutely putrid , probably the worst ACC ever . If the Big 10 was like terrible like in 2018 , I think, when Nebraska was 13-5 , but beat no one , maybe 1 of the top teams , they deserved to be left off , because the BiG 10 sucked that year, same as the ACC this year , with an okay overrated Duke and bubbles Notre Dame , UNC , Miami and Wake Forest at year end. The Committee has to value conference records and standings and who you beat, the top half , when evaluating teams from consensus better conferences. It did with Rutgers but my point is that it should have done it more and we should have been firmly in the field at the end of the regular season. We earned it .

Final takeaway is Pike should not decline any of these preseason tourneys , playing 3 games on a neutral floor or 2 games is not going to bite us in the ass if we take care of business in conference. If we didn’t , then we didn’t deserve a bid. It also would be better for our fan base and season ticket holders. For the life of me , Jimmy V was a Rutgers graduate. We should be 1 of the featured teams going forward every year. MSG is our second home and now that we are good again( took a long time ) all the bandwagon fans will fill the joint. Stop with the local round robin 3-4 game home schedule and play 3-4 not 6-7 against them. You do not know who is going to end up below 300 so do whatever you want. Plus the Committee can never say you played 6 teams in the 300’s again.
Love the job BAC has done and all the regulars that have made this a very vibrant message board that is followed closely by the media people but more importantly just basketball junkies shooting the breeze.
 
Predict perfectly? Its not just about getting all 68 teams. Palm seeded only 32 of the 68 teams on the correct line. That is atrocious, and even worse when you realize that his information is the one being shared on CBS during the big games on the final weekend before Selection Sunday!

Correct Seed Percentage:
Palm (CBS) - 32/68 = 47%
Lunardi (ESPN) - 38/68 = 56%
DeCourcy (FOX) - 50/68 = 74%

DeCourcy by far had the best year of the three. Some above mentioned Brad Wachtel as well. He went 49/68 = 72% with exact seeding. He would be a great candidate to replace Palm on CBS.
When I said predict perfectly I meant who gets in, not seeding etc. He is terrible, I know.
 
not sure what to think of conference record, they say it does not matter but with ND and RU it may have.
Also, I found it very telling that I saw an interview with next year's chairman (I think they're in the room the year before they take over) and he actually mentioned someone's conference record when discussing a specific team. They've been telling us forever that conference records don't matter but if you're following college basketball closely and watching and reading all year, if you don't know a bubble team's conference record then you're blind. It has to have some influence even if they don't normally admit it.
 
  • Like
Reactions: bac2therac
Can someone explain how play-in game seeing works. I thought the winner gets a 16 seed. We have RU/ND play for an 11, WYO/Indy at 12 and 2 other games for 16 seed.
 
  • Like
Reactions: mugrat86
One other point I forgot. Conference tournaments meaning almost nothing is a travesty . They shouldn’t change the pre set teams except when teams win the tourney have to be rewarded. Iowa was a 5 seed going in and should have ended a 4 at least and borderline 3 for winning the BIG 10. For them to remain a 5 is a joke. Why kill yourselves playing 3 games in 3 nights or 4 games in 4 nights to have no reward except to say you are BIG 10 tourney champs. Teams risk injury. That has to change. Contingency plans were always done and done this year with Davison losing.
Providence was a 4 in the first reveal and won a ton of close games and then lost a few games at the end and got blown out by Creighton by 25. They shouldn’t have remained a 4 since they did nothing to remain a 4 from the first reveal to the end of the tourneys. Duke another overrated team. Why no swap of Tennessee and Duke where Tennessee was a top 3 team before and won the more difficult SEC, as they should have been a 2.
So a bunch of other things need to be cleaned up.
 
I know you were very firm about conference record not mattering, but I think that is true to an extent. Committee won't care if you finished 6th and the 7th place team has a better resume.

Leaving out the 3rd place ACC team or 4th place B1G team is a little different and it impacts perception

This is what always made me so convinced we were a lock.

The NCAA would never leave out the 4th place team from the best conference.
 
  • Like
Reactions: mugrat86
I've been looking at the relationship between NET and seeding. The correlation between NET and seeding for the 50 1-12 teams is .817. It's higher for all teams, but that would include teams for which NET wouldn't have been a factor (AQs). .817 is pretty high, with NET accounting for roughly 2/3 of the variance in seeding. That is not to say that NET actually influenced the seedings in that fashion, but rather that is how strong the relationship is.

I graphed the scatterplot between NET and seeding below. As you can see, it's pretty much linear with a few moderate outliers. Houston (HO) got a substantially worse seed than it's NET and Wisky and Providence did better. Below this graph is one with all 68 teams in it (except for T A&M CC, for whom I cannot find a NET). Caveat emptor: I entered these data myself, so could be some key entry errors.




 
Last edited:
SkilletHead thinks like me.

Here is the difference between four rankings (NET, Bart's WAB, My Computer Ranking that I post in the "computer rankings" thread, and my "better NET") and the committee's rankings for everyone above the last at-large line:

TeamSeedNETWins Above Bubble (Bart)My Model AMy "Better NET"NET ErrorWAB ErrorMy Model Error"Better NET" Error
Gonzaga
1​
1​
8​
1​
2​
0​
7​
0​
1​
Arizona
2​
2​
2​
2​
1​
0​
0​
0​
1​
Kansas
3​
6​
1​
3​
3​
3​
2​
0​
0​
Baylor
4​
4​
3​
4​
4​
0​
1​
0​
0​
Auburn
5​
11​
5​
5​
8​
6​
0​
0​
3​
Kentucky
6​
5​
9​
7​
11​
1​
3​
1​
5​
Villanova
7​
8​
6​
8​
5​
1​
1​
1​
2​
Duke
8​
12​
11​
13​
17​
4​
3​
5​
9​
Wisconsin
9​
24​
10​
16​
12​
15​
1​
7​
3​
Tennessee
10​
7​
4​
6​
7​
3​
6​
4​
3​
Purdue
11​
13​
7​
11​
10​
2​
4​
0​
1​
Texas Tech
12​
9​
13​
9​
9​
3​
1​
3​
3​
UCLA
13​
10​
15​
10​
13​
3​
2​
3​
0​
Illinois
14​
15​
19​
19​
21​
1​
5​
5​
7​
Providence
15​
32​
14​
14​
6​
17​
1​
1​
9​
Arkansas
16​
20​
18​
18​
19​
4​
2​
2​
3​
UConn
17​
17​
26​
17​
14​
0​
9​
0​
3​
Houston
18​
3​
12​
12​
18​
15​
6​
6​
0​
Saint Mary's
19​
19​
20​
20​
20​
0​
1​
1​
1​
Iowa
20​
14​
16​
15​
24​
6​
4​
5​
4​
Alabama
21​
30​
34​
29​
35​
9​
13​
8​
14​
LSU
22​
18​
32​
23​
33​
4​
10​
1​
11​
Texas
23​
16​
28​
21​
22​
7​
5​
2​
1​
Colorado State
24​
28​
17​
22​
15​
4​
7​
2​
9​
USC
25​
35​
21​
25​
23​
10​
4​
0​
2​
Murray State
26​
21​
24​
36​
16​
5​
2​
10​
10​
Michigan State
27​
36​
25​
28​
30​
9​
2​
1​
3​
Ohio State
28​
26​
42​
35​
36​
2​
14​
7​
8​
Boise State
29​
29​
22​
26​
27​
0​
7​
3​
2​
North Carolina
30​
31​
23​
34​
38​
1​
7​
4​
8​
San Diego State
31​
25​
27​
24​
29​
6​
4​
7​
2​
Seton Hall
32​
37​
35​
27​
26​
5​
3​
5​
6​
Creighton
33​
55​
39​
31​
25​
22​
6​
2​
8​
TCU
34​
44​
29​
32​
32​
10​
5​
2​
2​
Marquette
35​
42​
52​
37​
28​
7​
17​
2​
7​
Memphis
36​
33​
43​
40​
48​
3​
7​
4​
12​
San Francisco
37​
22​
36​
30​
41​
15​
1​
7​
4​
Miami
38​
62​
37​
50​
46​
24​
1​
12​
8​
Loyola Chicago
39​
23​
38​
48​
61​
16​
1​
9​
22​
Davidson
40​
41​
33​
43​
43​
1​
7​
3​
3​
Iowa State
41​
49​
30​
33​
31​
8​
11​
8​
10​
Michigan
42​
34​
54​
42​
42​
8​
12​
0​
0​
Wyoming
43​
50​
31​
41​
34​
7​
12​
2​
9​
Rutgers
44​
77​
61​
63​
47​
33​
17​
19​
3​
Indiana
45​
38​
55​
45​
44​
7​
10​
0​
1​
Virginia Tech
46​
27​
48​
46​
57​
19​
2​
0​
11​
Notre Dame
47​
53​
45​
54​
51​
6​
2​
7​
4​
Mean Abs Error
7.06383​
5.276596​
3.638297872​
5.063829787​
 
All three of the other models predict committee seed much better than NET.

My computer rankings (which contain are margin of victory component and a pure W/L component) were off by an average of 3.6 spots for the committee's top 47.

My "Better NET" which is a pure W/L model was off by 5.1

Bart's WAB which is a pure W/L where the game difficulty depends on efficiency was off by 5.3

The NET, which the committee is supposedly using, was off by 7.1

Conclusions
(a) MOV / efficiency does matter to some degree
(b) They do not give a **** about the NET
 
Correlation between seed and model:

NET = 0.814 (slightly different from Skillet's # because I am using the actual seed # and not the 1-16 and also because I only included up through Notre Dame as that is the last team we can infer an overall ranking from the committee for)

WAB = 0.898
"Better NET" = 0.903
My original model = 0.942
 
  • Like
Reactions: SkilletHead2
Using solver in Excel I fit a linear combination of the four models (constraints each model weight between 0 and 1, sum of weights = 1)

The weights are:
NET 18.1%
WAB 12.3%
My Original Model 40.5%
Better NET 29.0%

This gets the average error to 3.2 spots and the correlation to 0.954
 
Rutgers was the biggest miss by every one of these models except for Better NET.

Actual seed 44
NET 77
WAB 61
My Original Model 63
Better NET 47
Combined Model Prediction 60.6
 
Bart's actual T-Ranketology has an avg error of 2.74 and a correlation of 0.968, outperforming even the in-sample performance of the fitted model. Which makes sense because none of the other ones are actual designed for this purpose.

If you let his model into the combined model the weighting of the actual NET drops to zero:

NET 0%
WAB 10.3%
My Original Model 0.2%
Better NET 18.1%
Bart T-Ranketology 71.5%

Average Error 2.02 spots, Correlation 0.982

Caveats:
- There are multicollinearity issues because Bart's T-Ranketology already uses NET and WAB as inputs
- We are probably pushing how many parameters we can fit on a data set of 47 teams without overfitting

It's interesting that my original model, which was the best performing and had the highest weight of the four before, drops to basically zero weight once Bart's T-Ranketology comes in.

It's also interesting that Better Net seems to add significant information to what he has.

If I exclude it from the fit I get:

NET 0.0%
WAB 13.1%
My Original Model 21.1%
Bart T-Ranketology 65.7%

Average Error 2.23 spots, Correlation 0.977
 
Sarahcuse is already dead.

Classic case of a coach hanging on too long.
I'm a Syracuse hater, but they were in the Sweet 16 last year. So it seems to me that SU and Boeheim are still relevant. But I do wonder if he's just playing out the string with his son and if the end is near.
 
again its one year and every year is somewhat different. There are some takeaways but also some things that need to be printed out and saved as it seems like twitter has given every amateur the right to bloviate about metrics, efficiency and what the committee actually values.

Rutgers got in with a bad net of 77 which is the worst in history. No it doesn't mean the net doesn't matter in the grand scheme of things but it does confirm that if a school has significant wins or a solid resume besides the NET ranking, the school will be selected. See Arizona State, St Johns, Michigan State in recent years with NETs in the 65-75 range.

Quality wins matter the most. Quad one wins matter alot. Wins vs the field really help a school. RU was 8-5. That a big time number. Very very hard to leave off and it was stunning to watch the metrics driven guys look at the number and just dismiss the wins out of hand. Some really bad takes on the Rutgers resume. RU was listed on just 1/3 of the brackets on the Matrix in the end and much of the commentary from them were focused solely on their losses and metrics and not wins. My number one tool in bracketology is wins vs the field. Q1 and 2 record is important but you have to dig deeper. I believe Wyoming was 11-6 in Q1/2 but still in first 4 games. All those wins are not the same. RU had high level Q1 wins as opposed to beating a bunch of 500 WCC/MWC schools. Wyoming only had 2 wins vs the field and thats why despite a gaudy record fell precipitously every week in bracketology from a projected 8 seed to last 4 in.

Quality wins matter more than bad losses. See Rutgers and then see surprising Dayton which came out of left field to be the last team out. Dayton had wins over Kansas, Miami and Va Tech. In league beat Davidson and Richmond. 5 wins vs the field. 5-0. They have 4 horrific losses, 3 in just one week to Lipscomb, Austin Peay, Umass Lowell, the other was La Salle. Those are hideous losses. Yet incredible out of conference wins for a high mid major. Rutgers overcame bad losses to Umass and Lafayette but had 8 wins vs the field.

The selection committee has an affinity for the Atlantic 10. Dayton just missing was twofold...it was the wins but it was also that the committee almost always is searching for that 2nd A10 team. I see it every year. I hate it because I think that having the Commish from the A10 on the committee has biased it a bit. Sometimes they seem to get a little bit of benefit of the doubt. I am glad a school like VCU lost and was nowhere near getting in because the metric and efficiency guys were pumping up a blank resume simply because their computer ratings were high

They say conference record and finish does not matter but it clearly did with North Carolina, 15-5 2nd place in the ACC. Inexplicably got in over Texas A&M (more on them later) with just a pedestrian 4-9 mark in Q1/2 and 4 wins vs the field and one of those was late addition Corpus Christi. They had a high end win over Kentucky sure but the other wins were just Miami and UNC. UNC AD on committee putting his thumb on things. I am sort of torn. I realize the ACC was down but IMO I do think conference record should be considered...look at RU...12-8. I think finishing 15-5 in the ACC does show they did not screw up too much against the weaker schools.

Conference tournaments mean less than we thought. Even I always fall into that trap. Tourneys are just one to three games during a season. It is body of work. It is hard to go from fringes of the bubble to into the field. We see it with Va Tech and Texas A&M. Va Tech did get in by winning the ACC AQ but if they had lost in the finals, they were not getting in. That seed of 11 and placement on the seedlist behind RU says it all. Texas A&M amassed two big quality wins and a nice won over bubble Florida but still didnt make it. At the end of the day A&M was too far back off the bubble. The committee looks like they had done alot of work heading into the final few days and had an idea of who they were putting in the field and the Aggies were not on the radar screen. Even with the wins they got, they still were a wobbly 4-10 vs Q1...do not get caught up in recency bias, body of work. That is not a good number at all. While they were 5-9 vs field, 6-9 if adding Corpus Christi, its okay but not overwhelming. They had some Q2 losses and their OOC sos was 257 whereas Notre Dame was 26. Those comparisons are important are the very end. I do feel that A&M probably deserved it over Notre Dame...and geez they beat them head to head but thats another thing we learned. Head to head really does not matter for selection or for seeding.

Sunday results mean almost nothing because the field is already set and besides a contingency bracket for a bid stealer the seeds are basically set. Iowa winning the Big 10 played no role as they were a 5 whether they won the tourney but in reality that should have immediately bumped them to a 4. Alot of talk about Tennessee not being a 2 seed, well dont expect a change on that day. Tennessee was also beating a team they didnt have in the field. Beating A&M wasnt going to push them up a seed.

Duke still gets favoritism. So on the other end of the argument, Duke really hasnt looked good in a weak conference even though they dominated the regular season. Losing to a team in the final who wasnt going to get in the field otherwise not a good look. Besides Tennessee one could argue Purdue over Duke as the last 2 too. UNC still gets favoritism. 8 seed which I got right because I knew they would the committee would overseed them. 3 wins vs tourney teams...alot of Q3 wins, if they didnt beat Duke they still would have got in with very little quality wins.

Big 10 also getting alot of favoritism. 9 schools out of 14 in. I mean agree with it but its alot. One can argue about spreading the wealth but with the ACC and Pac 12 being garbage this year. The Big 10 benefits from that.

Non conference wins matter alot...especially high end wins. Alabama just 19-13 and 9-9 in SEC play had a slew of high end wins. Notre Dame's win over Kentucky was a game changer. Non conference activity is a big factor in determining how your NET is going to end up and how your conference is going to do in the NET. That really helped the SEC and Mountain West out and you see it represented in their strong numbers. And while the NET isnt all that important its going to be important if you are a power 6 school and you have both a great net and great wins. In Iowas case they had a great net without the great wins. PS yes RU got but yes they need to schedule stronger OOC

15 losses is a bridge too far. See Oklahoma. Only one school ever with that many losses. You have to be extraordinary to get in. Oklahoma wasnt...they were just 4-12 vs the field. Perhaps if they beat Texas Tech and made the Big 12 final, 19-15 would be a new test for the committee but that did not happen. Michigan at 17-14 was just 3 above 500 which means they became the 5th school to make it in a non covid year less than 4 games above 500. If they were 17-15 they were not getting in. Record matters but its a tightrope walk. Michgan was also 7-12 vs the field. Better stuff than Oklahoma. You do see that schools with bulky loss totals like Michigan and RU will be put into the field if they have a significant amount of quality wins...

A big blow to metrics guys like Norlander, Parrish, Warriner, and the younger bracketologists. Efficiency and Predictive ratings are not how the field is selected. Many suspected this and it was confirmed. Old school wins out. The Kenpom, bpi, blah blah blah appear on the team sheet as numbers, nothing more. The committee is given no instruction to even consider them. If number looks out of whack or numbers in RU's case, the committee is going to see they are outliers given the meat of the resume. Personally when I evaluate, I do not even look at any of the 5-6 numbers like sor, bpi, kenpom, totally ignore them. The NET I do pay attention to but not until I have made my full evaluation. The committee knows full well stuff like Houston at 3, San Fran at 22, Va Tech at 27 , Rutgers at 77 does not fool anyone. What your resume says rather than the number matters more. I never want to see KenPom on the this board or his partner Bart Torvik...they are literally meaningless especially stuff like wins above the bubble..what does that even mean...someone please explain. its literally useless and so is Bart.

Bracketology should not begin in December and heat up in January. Bracketology should start in February and heat up by the end of the month. Alot of worthless brackets in January. Xavier was a projected 4 seed in January. Wake Forest was being talked about as a lock for 6 weeks. Both schools were talked about as locks even 2 weeks. Wyoming was a lock for weeks. Things can change in a hurry once the resumes are actually looked at rather than looking at a number or efficiency ratings. It was funny to watch how these schools dropped day after day and the big adjustments made.

Bracketologists have big egos including me. No one wants to be wrong. We all are. I need to do a better job on seeding, some of that is I have to get more in tune with how the committee sees things and not how I see things. Much to clean up there. Bracketologists are always learning.

Joe Lunardi is such a polarizing figure but I like him. He was the first. He is generally good at what he does. He knows his stuff. He is more consistent than most and he is old school. Jerry Palm is the worst guy that CBS can have. He is awkward not photogenic. In some ways I do like having a geek up there and it seems to be embracing his role as the Village Idiot but CBS is covering the NCAA tourney they should have a guy that give it to you straight. His bracketing the final week was awful. He was putting BYU in the field, like the only one on the matrix to do so. He was keeping Michigan, Indiana and RU out and then suddenly putting the former 2 in even though no one played. Yes we all scrub the field, but Jerry knew better, he knew what he was doing and he knew he was going to move at least one if not two Big 10 schools in. Historical data matters until it does not. Jerry knows this and will go with the odds over rational discussion about Rutgers.
Are you retired?
 
The flip side of that is that the computers said Virginia Tech was a strong team despite their record and then lo and behold they go and win the ACC tournament.
You could say the same the same about Iowa too. Their stellar NET made little sense until they finished on fire down the stretch and won the Big10 tournament. But that may just be the numbers working like a broken clock — now matter how shaky, they are still right twice a day. 😏
 
again its one year and every year is somewhat different. There are some takeaways but also some things that need to be printed out and saved as it seems like twitter has given every amateur the right to bloviate about metrics, efficiency and what the committee actually values.

Rutgers got in with a bad net of 77 which is the worst in history. No it doesn't mean the net doesn't matter in the grand scheme of things but it does confirm that if a school has significant wins or a solid resume besides the NET ranking, the school will be selected. See Arizona State, St Johns, Michigan State in recent years with NETs in the 65-75 range.

Quality wins matter the most. Quad one wins matter alot. Wins vs the field really help a school. RU was 8-5. That a big time number. Very very hard to leave off and it was stunning to watch the metrics driven guys look at the number and just dismiss the wins out of hand. Some really bad takes on the Rutgers resume. RU was listed on just 1/3 of the brackets on the Matrix in the end and much of the commentary from them were focused solely on their losses and metrics and not wins. My number one tool in bracketology is wins vs the field. Q1 and 2 record is important but you have to dig deeper. I believe Wyoming was 11-6 in Q1/2 but still in first 4 games. All those wins are not the same. RU had high level Q1 wins as opposed to beating a bunch of 500 WCC/MWC schools. Wyoming only had 2 wins vs the field and thats why despite a gaudy record fell precipitously every week in bracketology from a projected 8 seed to last 4 in.

Quality wins matter more than bad losses. See Rutgers and then see surprising Dayton which came out of left field to be the last team out. Dayton had wins over Kansas, Miami and Va Tech. In league beat Davidson and Richmond. 5 wins vs the field. 5-0. They have 4 horrific losses, 3 in just one week to Lipscomb, Austin Peay, Umass Lowell, the other was La Salle. Those are hideous losses. Yet incredible out of conference wins for a high mid major. Rutgers overcame bad losses to Umass and Lafayette but had 8 wins vs the field.

The selection committee has an affinity for the Atlantic 10. Dayton just missing was twofold...it was the wins but it was also that the committee almost always is searching for that 2nd A10 team. I see it every year. I hate it because I think that having the Commish from the A10 on the committee has biased it a bit. Sometimes they seem to get a little bit of benefit of the doubt. I am glad a school like VCU lost and was nowhere near getting in because the metric and efficiency guys were pumping up a blank resume simply because their computer ratings were high

They say conference record and finish does not matter but it clearly did with North Carolina, 15-5 2nd place in the ACC. Inexplicably got in over Texas A&M (more on them later) with just a pedestrian 4-9 mark in Q1/2 and 4 wins vs the field and one of those was late addition Corpus Christi. They had a high end win over Kentucky sure but the other wins were just Miami and UNC. UNC AD on committee putting his thumb on things. I am sort of torn. I realize the ACC was down but IMO I do think conference record should be considered...look at RU...12-8. I think finishing 15-5 in the ACC does show they did not screw up too much against the weaker schools.

Conference tournaments mean less than we thought. Even I always fall into that trap. Tourneys are just one to three games during a season. It is body of work. It is hard to go from fringes of the bubble to into the field. We see it with Va Tech and Texas A&M. Va Tech did get in by winning the ACC AQ but if they had lost in the finals, they were not getting in. That seed of 11 and placement on the seedlist behind RU says it all. Texas A&M amassed two big quality wins and a nice won over bubble Florida but still didnt make it. At the end of the day A&M was too far back off the bubble. The committee looks like they had done alot of work heading into the final few days and had an idea of who they were putting in the field and the Aggies were not on the radar screen. Even with the wins they got, they still were a wobbly 4-10 vs Q1...do not get caught up in recency bias, body of work. That is not a good number at all. While they were 5-9 vs field, 6-9 if adding Corpus Christi, its okay but not overwhelming. They had some Q2 losses and their OOC sos was 257 whereas Notre Dame was 26. Those comparisons are important are the very end. I do feel that A&M probably deserved it over Notre Dame...and geez they beat them head to head but thats another thing we learned. Head to head really does not matter for selection or for seeding.

Sunday results mean almost nothing because the field is already set and besides a contingency bracket for a bid stealer the seeds are basically set. Iowa winning the Big 10 played no role as they were a 5 whether they won the tourney but in reality that should have immediately bumped them to a 4. Alot of talk about Tennessee not being a 2 seed, well dont expect a change on that day. Tennessee was also beating a team they didnt have in the field. Beating A&M wasnt going to push them up a seed.

Duke still gets favoritism. So on the other end of the argument, Duke really hasnt looked good in a weak conference even though they dominated the regular season. Losing to a team in the final who wasnt going to get in the field otherwise not a good look. Besides Tennessee one could argue Purdue over Duke as the last 2 too. UNC still gets favoritism. 8 seed which I got right because I knew they would the committee would overseed them. 3 wins vs tourney teams...alot of Q3 wins, if they didnt beat Duke they still would have got in with very little quality wins.

Big 10 also getting alot of favoritism. 9 schools out of 14 in. I mean agree with it but its alot. One can argue about spreading the wealth but with the ACC and Pac 12 being garbage this year. The Big 10 benefits from that.

Non conference wins matter alot...especially high end wins. Alabama just 19-13 and 9-9 in SEC play had a slew of high end wins. Notre Dame's win over Kentucky was a game changer. Non conference activity is a big factor in determining how your NET is going to end up and how your conference is going to do in the NET. That really helped the SEC and Mountain West out and you see it represented in their strong numbers. And while the NET isnt all that important its going to be important if you are a power 6 school and you have both a great net and great wins. In Iowas case they had a great net without the great wins. PS yes RU got but yes they need to schedule stronger OOC

15 losses is a bridge too far. See Oklahoma. Only one school ever with that many losses. You have to be extraordinary to get in. Oklahoma wasnt...they were just 4-12 vs the field. Perhaps if they beat Texas Tech and made the Big 12 final, 19-15 would be a new test for the committee but that did not happen. Michigan at 17-14 was just 3 above 500 which means they became the 5th school to make it in a non covid year less than 4 games above 500. If they were 17-15 they were not getting in. Record matters but its a tightrope walk. Michgan was also 7-12 vs the field. Better stuff than Oklahoma. You do see that schools with bulky loss totals like Michigan and RU will be put into the field if they have a significant amount of quality wins...

A big blow to metrics guys like Norlander, Parrish, Warriner, and the younger bracketologists. Efficiency and Predictive ratings are not how the field is selected. Many suspected this and it was confirmed. Old school wins out. The Kenpom, bpi, blah blah blah appear on the team sheet as numbers, nothing more. The committee is given no instruction to even consider them. If number looks out of whack or numbers in RU's case, the committee is going to see they are outliers given the meat of the resume. Personally when I evaluate, I do not even look at any of the 5-6 numbers like sor, bpi, kenpom, totally ignore them. The NET I do pay attention to but not until I have made my full evaluation. The committee knows full well stuff like Houston at 3, San Fran at 22, Va Tech at 27 , Rutgers at 77 does not fool anyone. What your resume says rather than the number matters more. I never want to see KenPom on the this board or his partner Bart Torvik...they are literally meaningless especially stuff like wins above the bubble..what does that even mean...someone please explain. its literally useless and so is Bart.

Bracketology should not begin in December and heat up in January. Bracketology should start in February and heat up by the end of the month. Alot of worthless brackets in January. Xavier was a projected 4 seed in January. Wake Forest was being talked about as a lock for 6 weeks. Both schools were talked about as locks even 2 weeks. Wyoming was a lock for weeks. Things can change in a hurry once the resumes are actually looked at rather than looking at a number or efficiency ratings. It was funny to watch how these schools dropped day after day and the big adjustments made.

Bracketologists have big egos including me. No one wants to be wrong. We all are. I need to do a better job on seeding, some of that is I have to get more in tune with how the committee sees things and not how I see things. Much to clean up there. Bracketologists are always learning.

Joe Lunardi is such a polarizing figure but I like him. He was the first. He is generally good at what he does. He knows his stuff. He is more consistent than most and he is old school. Jerry Palm is the worst guy that CBS can have. He is awkward not photogenic. In some ways I do like having a geek up there and it seems to be embracing his role as the Village Idiot but CBS is covering the NCAA tourney they should have a guy that give it to you straight. His bracketing the final week was awful. He was putting BYU in the field, like the only one on the matrix to do so. He was keeping Michigan, Indiana and RU out and then suddenly putting the former 2 in even though no one played. Yes we all scrub the field, but Jerry knew better, he knew what he was doing and he knew he was going to move at least one if not two Big 10 schools in. Historical data matters until it does not. Jerry knows this and will go with the odds over rational discussion about Rutgers.

again its one year and every year is somewhat different. There are some takeaways but also some things that need to be printed out and saved as it seems like twitter has given every amateur the right to bloviate about metrics, efficiency and what the committee actually values.

Rutgers got in with a bad net of 77 which is the worst in history. No it doesn't mean the net doesn't matter in the grand scheme of things but it does confirm that if a school has significant wins or a solid resume besides the NET ranking, the school will be selected. See Arizona State, St Johns, Michigan State in recent years with NETs in the 65-75 range.

Quality wins matter the most. Quad one wins matter alot. Wins vs the field really help a school. RU was 8-5. That a big time number. Very very hard to leave off and it was stunning to watch the metrics driven guys look at the number and just dismiss the wins out of hand. Some really bad takes on the Rutgers resume. RU was listed on just 1/3 of the brackets on the Matrix in the end and much of the commentary from them were focused solely on their losses and metrics and not wins. My number one tool in bracketology is wins vs the field. Q1 and 2 record is important but you have to dig deeper. I believe Wyoming was 11-6 in Q1/2 but still in first 4 games. All those wins are not the same. RU had high level Q1 wins as opposed to beating a bunch of 500 WCC/MWC schools. Wyoming only had 2 wins vs the field and thats why despite a gaudy record fell precipitously every week in bracketology from a projected 8 seed to last 4 in.

Quality wins matter more than bad losses. See Rutgers and then see surprising Dayton which came out of left field to be the last team out. Dayton had wins over Kansas, Miami and Va Tech. In league beat Davidson and Richmond. 5 wins vs the field. 5-0. They have 4 horrific losses, 3 in just one week to Lipscomb, Austin Peay, Umass Lowell, the other was La Salle. Those are hideous losses. Yet incredible out of conference wins for a high mid major. Rutgers overcame bad losses to Umass and Lafayette but had 8 wins vs the field.

The selection committee has an affinity for the Atlantic 10. Dayton just missing was twofold...it was the wins but it was also that the committee almost always is searching for that 2nd A10 team. I see it every year. I hate it because I think that having the Commish from the A10 on the committee has biased it a bit. Sometimes they seem to get a little bit of benefit of the doubt. I am glad a school like VCU lost and was nowhere near getting in because the metric and efficiency guys were pumping up a blank resume simply because their computer ratings were high

They say conference record and finish does not matter but it clearly did with Notre Dame, 15-5 2nd place in the ACC. Inexplicably got in over Texas A&M (more on them later) with just a pedestrian 4-9 mark in Q1/2 and 4 wins vs the field and one of those was late addition Corpus Christi. They had a high end win over Kentucky sure but the other wins were just Miami and UNC. UNC AD on committee putting his thumb on things. I am sort of torn. I realize the ACC was down but IMO I do think conference record should be considered...look at RU...12-8. I think finishing 15-5 in the ACC does show they did not screw up too much against the weaker schools.

Conference tournaments mean less than we thought. Even I always fall into that trap. Tourneys are just one to three games during a season. It is body of work. It is hard to go from fringes of the bubble to into the field. We see it with Va Tech and Texas A&M. Va Tech did get in by winning the ACC AQ but if they had lost in the finals, they were not getting in. That seed of 11 and placement on the seedlist behind RU says it all. Texas A&M amassed two big quality wins and a nice won over bubble Florida but still didnt make it. At the end of the day A&M was too far back off the bubble. The committee looks like they had done alot of work heading into the final few days and had an idea of who they were putting in the field and the Aggies were not on the radar screen. Even with the wins they got, they still were a wobbly 4-10 vs Q1...do not get caught up in recency bias, body of work. That is not a good number at all. While they were 5-9 vs field, 6-9 if adding Corpus Christi, its okay but not overwhelming. They had some Q2 losses and their OOC sos was 257 whereas Notre Dame was 26. Those comparisons are important are the very end. I do feel that A&M probably deserved it over Notre Dame...and geez they beat them head to head but thats another thing we learned. Head to head really does not matter for selection or for seeding.

Sunday results mean almost nothing because the field is already set and besides a contingency bracket for a bid stealer the seeds are basically set. Iowa winning the Big 10 played no role as they were a 5 whether they won the tourney but in reality that should have immediately bumped them to a 4. Alot of talk about Tennessee not being a 2 seed, well dont expect a change on that day. Tennessee was also beating a team they didnt have in the field. Beating A&M wasnt going to push them up a seed.

Duke still gets favoritism. So on the other end of the argument, Duke really hasnt looked good in a weak conference even though they dominated the regular season. Losing to a team in the final who wasnt going to get in the field otherwise not a good look. Besides Tennessee one could argue Purdue over Duke as the last 2 too. UNC still gets favoritism. 8 seed which I got right because I knew they would the committee would overseed them. 3 wins vs tourney teams...alot of Q3 wins, if they didnt beat Duke they still would have got in with very little quality wins.

Big 10 also getting alot of favoritism. 9 schools out of 14 in. I mean agree with it but its alot. One can argue about spreading the wealth but with the ACC and Pac 12 being garbage this year. The Big 10 benefits from that.

Non conference wins matter alot...especially high end wins. Alabama just 19-13 and 9-9 in SEC play had a slew of high end wins. Notre Dame's win over Kentucky was a game changer. Non conference activity is a big factor in determining how your NET is going to end up and how your conference is going to do in the NET. That really helped the SEC and Mountain West out and you see it represented in their strong numbers. And while the NET isnt all that important its going to be important if you are a power 6 school and you have both a great net and great wins. In Iowas case they had a great net without the great wins. PS yes RU got but yes they need to schedule stronger OOC

15 losses is a bridge too far. See Oklahoma. Only one school ever with that many losses. You have to be extraordinary to get in. Oklahoma wasnt...they were just 4-12 vs the field. Perhaps if they beat Texas Tech and made the Big 12 final, 19-15 would be a new test for the committee but that did not happen. Michigan at 17-14 was just 3 above 500 which means they became the 5th school to make it in a non covid year less than 4 games above 500. If they were 17-15 they were not getting in. Record matters but its a tightrope walk. Michgan was also 7-12 vs the field. Better stuff than Oklahoma. You do see that schools with bulky loss totals like Michigan and RU will be put into the field if they have a significant amount of quality wins...

A big blow to metrics guys like Norlander, Parrish, Warriner, and the younger bracketologists. Efficiency and Predictive ratings are not how the field is selected. Many suspected this and it was confirmed. Old school wins out. The Kenpom, bpi, blah blah blah appear on the team sheet as numbers, nothing more. The committee is given no instruction to even consider them. If number looks out of whack or numbers in RU's case, the committee is going to see they are outliers given the meat of the resume. Personally when I evaluate, I do not even look at any of the 5-6 numbers like sor, bpi, kenpom, totally ignore them. The NET I do pay attention to but not until I have made my full evaluation. The committee knows full well stuff like Houston at 3, San Fran at 22, Va Tech at 27 , Rutgers at 77 does not fool anyone. What your resume says rather than the number matters more. I never want to see KenPom on the this board or his partner Bart Torvik...they are literally meaningless especially stuff like wins above the bubble..what does that even mean...someone please explain. its literally useless and so is Bart.

Bracketology should not begin in December and heat up in January. Bracketology should start in February and heat up by the end of the month. Alot of worthless brackets in January. Xavier was a projected 4 seed in January. Wake Forest was being talked about as a lock for 6 weeks. Both schools were talked about as locks even 2 weeks. Wyoming was a lock for weeks. Things can change in a hurry once the resumes are actually looked at rather than looking at a number or efficiency ratings. It was funny to watch how these schools dropped day after day and the big adjustments made.

Bracketologists have big egos including me. No one wants to be wrong. We all are. I need to do a better job on seeding, some of that is I have to get more in tune with how the committee sees things and not how I see things. Much to clean up there. Bracketologists are always learning.

Joe Lunardi is such a polarizing figure but I like him. He was the first. He is generally good at what he does. He knows his stuff. He is more consistent than most and he is old school. Jerry Palm is the worst guy that CBS can have. He is awkward not photogenic. In some ways I do like having a geek up there and it seems to be embracing his role as the Village Idiot but CBS is covering the NCAA tourney they should have a guy that give it to you straight. His bracketing the final week was awful. He was putting BYU in the field, like the only one on the matrix to do so. He was keeping Michigan, Indiana and RU out and then suddenly putting the former 2 in even though no one played. Yes we all scrub the field, but Jerry knew better, he knew what he was doing and he knew he was going to move at least one if not two Big 10 schools in. Historical data matters until it does not. Jerry knows this and will go with the odds over rational discussion about Rutgers.
bumping this from last year and the example of Dayton last year is interesting.....they were 23-10 and the last team left out. Somewhat surprising because the A10 already had 2 schools in the field, if Richmond didnt win the conference tourney, Dayton would have went. Remember Daytons numbers...Q1: 4-2, Q1/2: 9-6 but Q3: 5-1 and Q4: 9-3. A whopping 3 Q4 losses plus a Q3 and these were really bad losses like LaSalle, Lipscomb, Austin Peay and UMass Lowell. They had great wins: Kansas, Va Tech, and Miami with league wins over Richmond and Davidson so 5 wins vs the field. Again they made their hay OOC but still they were going to dance with those 3 Q4 losses until Richmond stole their bid. I didnt have them in my last 8 out, I was so dismissive of those losses...so keep that in mind this year


The original post is a must read to refresh ourselves on what we forget.....note the stuff about 15 loss schools and that will be put to the test with Oklahoma State should they lose today.
 
ADVERTISEMENT
ADVERTISEMENT