Really Good Teams and NCAA Tournament Success

Great to see you back, my friend.

IMHO, where the H got lost a long time ago, upsets in #1 vs. #16 should be much rarer than #2-#15 upsets. Distributions have thin tails; therefore, the quality difference should be greater between #1 and #2 than between #2 and #3, etc.

Anyway, thanks for presenting your analysis.
 
Kedsy said:
In our past 25 tournaments:

Duke is 10 of 19 when having a #1 or #2 seed (52.6%);

Duke is 10 of 18 when finishing in KenPom's top 8 (55.6%);

Duke is 8 of 14 when ending the season with 5 or fewer losses (57.1%);
Click to expand...
I just realized I left one out. In our past 25 tournaments, Duke is 5 of 6 when finishing with a KP EM of 30+ (83.3%).

Our KP EM is currently 36.73 (#1 in the country). Let's hope it stays above that 30 mark.
 
sagegrouse said:
One other variant to try, Kedsy, is the preseason rankings. I have heard that preseason polls are better predictors of NCAAT success than late-season rankings.
Click to expand...
OK, it didn't take me a couple days.

I measured how many teams in the pre-season Top 8 made the Elite Eight, to make it consistent with my other measurements.

In the past 25 seasons, 92 pre-season top 8 teams made the Elite Eight, out of 200 possibilities, which is 46.0%.

In the past 10 seasons, 33 pre-season top 8 teams made the Elite Eight, out of 80 possibilities, which is 41.3%.

A little bit worse than the other numbers we've looked at.

Part of the reason for the low numbers is because in the past four seasons, only 8 pre-season top 8 teams made the Elite Eight, out of 32 which is 25%, and likely a small sample anomaly. If you don't count 2021 through 2024, the previous numbers become 84 out of 168 (50%) and 25 out of 48 (52.1%). Which are consistent with the other numbers we've looked at, but not necessarily better.

It's also likely you saw analysis of more than the top 8. If you extend it to the pre-season top 10, the 25-year numbers become 106 out of 200 (53.0%) and the 10-year numbers become 38 out of 80 (47.5%). I imagine the numbers for the pre-season top 25 are even better, but that wouldn't be an apples-to-apples comparison, so I didn't bother with that.

There were several instances where a pre-season top 8 team ended up a 6-seed, or an 8-seed, or even an 11-seed and a 12-seed in a couple instances, and still made the Elite Eight, which admittedly is pretty cool.

FWIW, Duke was #7 in the AP Pre-season Top 25.
 
Kedsy said:
OK, it didn't take me a couple days.

I measured how many teams in the pre-season Top 8 made the Elite Eight, to make it consistent with my other measurements.

In the past 25 seasons, 92 pre-season top 8 teams made the Elite Eight, out of 200 possibilities, which is 46.0%.

In the past 10 seasons, 33 pre-season top 8 teams made the Elite Eight, out of 80 possibilities, which is 41.3%.

A little bit worse than the other numbers we've looked at.

Part of the reason for the low numbers is because in the past four seasons, only 8 pre-season top 8 teams made the Elite Eight, out of 32 which is 25%, and likely a small sample anomaly. If you don't count 2021 through 2024, the previous numbers become 84 out of 168 (50%) and 25 out of 48 (52.1%). Which are consistent with the other numbers we've looked at, but not necessarily better.

It's also likely you saw analysis of more than the top 8. If you extend it to the pre-season top 10, the 25-year numbers become 106 out of 200 (53.0%) and the 10-year numbers become 38 out of 80 (47.5%). I imagine the numbers for the pre-season top 25 are even better, but that wouldn't be an apples-to-apples comparison, so I didn't bother with that.

There were several instances where a pre-season top 8 team ended up a 6-seed, or an 8-seed, or even an 11-seed and a 12-seed in a couple instances, and still made the Elite Eight, which admittedly is pretty cool.

FWIW, Duke was #7 in the AP Pre-season Top 25.
Click to expand...
I for one appreciate your hard work on the numbers/analytics you give us. CDu as well. You guys keep it up.
And Happy New Year (8 days late).

GoDuke!
 
Very cool, Kedsy!

I've been trying to keep the "Kedsy Ranking" alive and well. (Assign points based on RSCI ranking and subtract a half point for each year at Duke) but it's been a struggle. I still use it every summer/fall to predict the starters.

Good to have you back!
 
Love the topic. Fun analysis! Kedsy did a lot of work. Respect!

I've actually thought about this some over the years for Duke (most other teams don't have as high a bar, so I've had discussions with other team's fans about this). I think it's very hard to actually quantitatively define a "good" / "really good" / "great" / "exceptional" / "best" season etc. And there's definitely a perspective factor. I personally don't think it is quantitative, it's more qualitative. And to me, it's about standings in the regular season, conf tourney performance and NCAAs. And beating UNC (for Duke football for me, it's very simple. Beat UNC, regardless of every other game, it's a great season. End of discussion, let's talk about basketball).

As examples:

Let's say you finish middle of the pack in your conference, lose a couple of games you should have won but barely squeak into the NCAAs without winning your conf tourney, then get to the FF. To me that's a "great" season.

Let's say you go undefeated in the regular season, then lose in your conf tourney, then lose before the FF in the NCAAs. Hard to label this season, but I'd say that's a "good" season that was ultimately disappointing.

Let's say you are a #1 seed in the NCAAs, but lose to a #16 seed. I think it's hard to call this a "good season". One bad apples spoils the whole bunch, girl.

Anyway to me for Duke, any of these would be a "good" season:

Good

- Make the S16 without winning the regular season or conf tourney. In any given game the #16 ranked team (not the 16 seed, talking about polls) could beat the #1 ranked team
- Win the conf tourney, get past the 1st round in the NCAAs. We know this was Scheyer's first season. Cutting down post-season nets is always a thrill.

Really Good

To get better (perhaps "really good", though I never defined "really good" in my head):

- I tend to agree with Kedsy - finish in the top half of the regular season, make the E8, regardless of conf tourney performance. I'd also say don't get swept by UNC.

Great

For "great", either of these:

- FF, unless you are undefeated and lose in the FF. Then you drop back down to "really good". Again, it's that bad apple. There's another example for me where "great" drops down to a lesser rank, but I won't talk about that.
- beat UNC every time you play them in a season. Even if Duke had a losing record, but beat UNC each time, that would be great!

Exceptional

For "exceptional":

- win the NCAAs. Doesn't matter what else happened. '91, '00, '10, '15 all are exceptional.

Best

For "best":

- After winning multiple NCAAs, or comparing multiple NCAAs, decide that one is the "best". Totally subjective. I think '10 was Duke's best season. Obviously the natty, but sweeping UNC that year and the beatdown on them on Senior Night (82 and 50 are some of my favorite numbers), along with UNC sucking so bad that year, it was the best. '92 would surpass that one for me, but Duke lost in Chapel Hill. Of course, others may disagree. It's great to have multiple championships to have the argument.


For this year, Duke could have its "best" season for me. Go undefeated in conference (thus beating UNC twice and winning the regular season), beat the last remaining non-conf game vs. Illinois, win the ACCs (and beat Carolina in it) and win the natty. They finish with only 2 losses. Will that all happen? Not likely. There's a lot of season left (it's only January 8th), so very hard to be confident about it. But I'll settle for another "exceptional" season ;)

9F
 
House P said:
Welcome back, Kedsy.

Remind me again how the 2022-3 preseason #1 team did in the tourney. :)
Click to expand...
But we all knew that at the time, didn't we?

Actually, it's impressive that the pre-season rankings predict as well as they do, considering they pretty much always over-rank teams that did well in the previous NCAAT.
 
Kedsy said:
But we all knew that at the time, didn't we?

Actually, it's impressive that the pre-season rankings predict as well as they do, considering they pretty much always over-rank teams that did well in the previous NCAAT.
Click to expand...
And I suspect they'll do less well moving forward in the era of free agency. It was one thing when teams largely stayed the same from year to year. But now teams turn over soooooo much that it's likely to get harder and harder to get it right in the preseason. And it was already probably getting harder with the increase in early entry.
 
You must log in or register to reply here.
Back
Top