Sounds more like they're driving into a ditch than on cruise control.
2024-25 NET/Bracketology Thread
- Thread starter scottdude8
- Start date
AustinDevil
Member
Oooooooof. There are bad dad jokes, then there’s … that.
mattyoung18
Member
Whoever ends up playing Mcneese St its going to be a street fight!
throatybeard
Member
OPhK is not like us
DallasDevil
Member
FYI, here is the full committee principles and procedures document for this year: NCAA Tournament Principles 2025
DBKilby
Member
Hopefully it turns out better for them this year compared to last year against Gonzaga.
AtlBluRew
Member
Did you notice that if the Cheats win their play-in game, Lunardi has them facing Caleb Love and Arizona?
SkyBrickey
Member
I guess this shouldn't be too surprising...
So I ran the numbers for each seed between #4 and #11, with three categories each. "+2" means the seed should be worse (e.g., the team is a #6 but it should be a #8 or worse); "-2" means the seed should be better (e.g., the team is a #6 but it should be a #4 or better); "+1,0,-1" means the team is within one seed line of what it should be (based on KenPom pre-tournament ratings).
NOTE that I'm saying all overseeded teams are +2, but the numbers I've run are >=2, and I'm saying all underseeded teams are -2 but the numbers I've run are <=-2. In other words I'm lumping teams that are overseeded (or underseeded) by 2 along with teams that are overseeded (or underseeded) by 3 or 4 or 5 or more and calling it +2 (or -2). This might matter for the 2nd table, but there are too many possibilities for me to do it differently.
| Seed | W | L | Pct | |
| +2 | 4 | 21 | 25 | 0.457 |
| +1,0,-1 | 4 | 74 | 41 | 0.643 |
| -2 | 4 | 33 | 17 | 0.660 |
| +2 | 5 | 16 | 22 | 0.421 |
| +1,0,-1 | 5 | 59 | 43 | 0.578 |
| -2 | 5 | 25 | 19 | 0.568 |
| +2 | 6 | 24 | 29 | 0.453 |
| +1,0,-1 | 6 | 31 | 40 | 0.437 |
| -2 | 6 | 14 | 15 | 0.483 |
| +2 | 7 | 26 | 37 | 0.413 |
| +1,0,-1 | 7 | 37 | 31 | 0.544 |
| -2 | 7 | 20 | 15 | 0.571 |
| +2 | 8 | 24 | 31 | 0.436 |
| +1,0,-1 | 8 | 22 | 33 | 0.400 |
| -2 | 8 | 18 | 20 | 0.474 |
| +2 | 9 | 20 | 38 | 0.345 |
| +1,0,-1 | 9 | 19 | 27 | 0.413 |
| -2 | 9 | 12 | 19 | 0.387 |
| +2 | 10 | 13 | 30 | 0.302 |
| +1,0,-1 | 10 | 12 | 22 | 0.353 |
| -2 | 10 | 22 | 32 | 0.407 |
| +2 | 11 | 24 | 38 | 0.387 |
| +1,0,-1 | 11 | 21 | 25 | 0.457 |
| -2 | 11 | 19 | 21 | 0.475 |
Totals for these eight seeds in the aggregate are:
.512 winning percentage for teams within one seed line;
.508 winning pct for teams that should be seeded better;
.402 winning pct for teams that should be seeded worse.
That said, the winning pct is best (of the three categories) for the teams that should be seeded better in six of the eight seed groups, and the winning pct is worst (of the three categories) for the teams that should be seeded worse in six of the eight seed groups. And that makes sense, because teams that should be seeded better are generally better teams than teams seeded correctly and teams that should be seeded worse are generally worse teams than teams seeded correctly.
BUT, that wasn't your question.
Below, I have re-arranged the table to group the items in a way that compares, e.g., a well-seeded #6, an overseeded #4, and an underseeded #8. (And remember my note above, which means I'm lumping together teams that were over- or under-seeded by 2, 3, 4, etc. and calling them "+2" or "-2," but it can't be helped.)
| Sd | W | L | Pct | |
| +1,0,-1 | 2 | 164 | 67 | 0.710 |
| -2 | 4 | 33 | 17 | 0.660 |
| +1,0,-1 | 3 | 105 | 49 | 0.682 |
| -2 | 5 | 25 | 19 | 0.568 |
| +1,0,-1 | 4 | 74 | 41 | 0.643 |
| -2 | 6 | 14 | 15 | 0.483 |
| +1,0,-1 | 5 | 59 | 43 | 0.578 |
| -2 | 7 | 20 | 15 | 0.571 |
| +2 | 4 | 21 | 25 | 0.457 |
| +1,0,-1 | 6 | 31 | 40 | 0.437 |
| -2 | 8 | 18 | 20 | 0.474 |
| +2 | 5 | 16 | 22 | 0.421 |
| +1,0,-1 | 7 | 37 | 31 | 0.544 |
| -2 | 9 | 12 | 19 | 0.387 |
| +2 | 6 | 24 | 29 | 0.453 |
| +1,0,-1 | 8 | 22 | 33 | 0.400 |
| -2 | 10 | 22 | 32 | 0.407 |
| +2 | 7 | 26 | 37 | 0.413 |
| +1,0,-1 | 9 | 19 | 27 | 0.413 |
| -2 | 11 | 19 | 21 | 0.475 |
| +2 | 8 | 24 | 31 | 0.436 |
| +1,0,-1 | 10 | 12 | 22 | 0.353 |
| +2 | 9 | 20 | 38 | 0.345 |
| +1,0,-1 | 11 | 21 | 25 | 0.457 |
| +2 | 10 | 13 | 30 | 0.302 |
| +1,0,-1 | 12 | 16 | 17 | 0.485 |
| +2 | 11 | 24 | 38 | 0.387 |
| +1,0,-1 | 13 | 1 | 4 | 0.200 |
In this table, if you're underseeded (should be seeded better), you're better than scratch in 3 of the 8 seed groups. If you're overseeded (should be worse), you're better than scratch in 4 out of 7 (with one tie).
I think the reason for the inconsistencies are mostly due to where you're seeded. Underseeded 4s, 5s, and 6s perform much worse than accurately-seeded 2s, 3s, and 4s, presumably because their first round opponents are much tougher. Underseeded 9s also perform much worse than accurately-seeded 7s, presumably because the #1 you have to play in your second game as a #9 is usually a lot better than a #2 you play as a #7. But the underseeded #11 does a fair amount better than an accurately seeded #9, for the same reason. And yet, underseeded 7s, 8s, and 10s do about the same as accurately-seeded 5s, 6s, and 8s. Similarly, overseeded 11s and 12s that get stuck as 9s and 10s are in trouble, while overseeded 13s more or less hit the jackpot (I realize there are only five games in the accurately-seeded 13-seeds, but ALL 13-seeds win at a rate of 21% so the principle holds; and yes, I'm flabbergasted that only five #13-seeds in 21 years were within one seed line of what they should be based on their KenPom ranking. I manually double-checked it and it appears to be true.). But I can't really explain why overseeded 5s do so much worse than accurately-seeded 7s.
Anyway, I'm not sure the data tells us anything authoritatively, but you asked for it, so here it is.
BlueDevil2K
Member
Did you filter by tournament teams, or does the kenpom side use all teams? If the latter, I think you're running into the fact that there are ~15-20 conference champions each year who are not among the top 68 teams. Put differently, this is the bubble line, which is why the play-in games feature 10-12 seeds depending on the year.
If you filtered for tournament teams, beats me!
No I didn't filter for only tournament teams. I used KP rank, so that if a team is ranked 72nd they should be a 16-seed. Which I understand explains it, or most of it, anyway. But still, five accurately-seeded 13-seeds in 21 tournaments? I think that's crazy. Though I suppose it also explains why 4-seeds perform so much better in the first round than 5-seeds. From 2002 to 2023, #4s won 78.6% of their first-round games while #5s only won 60.7% of theirs, while in the same period, if both survived first-round upsets, 5-seeds won 50% of head-to-head against 4-seeds (20-20).
BrummieDevil
Member
It behooves these bracketologists with a following to move teams around daily as if there's actually constant movement going on rather than us being in the midst of a month-long period between the Top 16 release and when the selection committee reconvenes when nothing is happening with the bracket and barely anything is known.
Ohio State did itself no favors losing that game, but they still have five more Q1 wins than UNC, who itself is playing a Q2 game today against a team just behind Iowa in NET.
Damnit. UNC won. That stinks. Xavier barely lost. Indiana lost. Did ANYONE help kill UNC today?
scottdude8
Moderator, front page and podcast contributor
Actually Texas won a big one. Huge Q1 versus Texas AM. … Texas has way more Q1 than UNC
SkyBrickey
Member
Texas has two Q1 wins in two days. UNC has one all season! It shouldn't even be a debate.
