Game Theory - where do we want to be seeded
Being that Michigan is on the right side of the bubble for the moment, and it's a bit of a slow news day college-wise, I've started thinking a little about where the best viable seeding would be. Obviously we don't know who will be the 1s and 2s and in what regions yet, but I'm curious where people think our best chances of advancing would be. I'm looking primarily at advancing to the sweet 16, since first round wins are as much based on individual matchups as they are seeding.
The data* (not including last year's tournament) shows that of the last 31 NCAA tournaments, there have been 41 5 and 6 seeds advancing to the Sweet 16 - put another way, roughly 33% of the time a 5 or a 6 will make it past the first weekend. I feel like a 5/6 line is the absolute best case scenario for us, and getting there would likely require winning the BTT, which I think we can all agree is unlikely. It gets more interesting looking at the 7-12 range (which I think is the mostly likely scenario where we end up). Both the 7 and 10 seeds have had 22 teams advance, making the Sweet 16 at about an 18% rate. After that, the next best advance rate is the 12 seed at 16% (20 teams advancing) and the 11 seed at 14.5% (18 teams). From there, it's a pretty percipitous fall off to the 8-9 lines; 8 seeds make it at about a 9% rate (12 teams), 9 seeds at 4% (5 teams).
Some of these make sense, as the 8-9 winner will have to beat the 1 seed to advance, which unsurprisingly is hard. Similarly, it makes sense that the 12 would do a little better than the 11, since they have an easier second round matchup in a 4 instead of a 3 (assuming chalk). I was mainly surprised though to see how much easier it is for a 2 seed to go down compared to 1 seed, since in theory those teams should be pretty close in quality, but the math shows 2s going down at a much, much higher rate. I also found it interesting that the 7 and 10 had the same advancement rate, which seems to indicate a lot of parity at that seed, whereas there was a tremendous gap between 8 and 9, the two "interchangeable" seeds.
What's the take away for us? I don't know, I just find the data interesting because I'm a dork. Obviously trying to game the system for a certain amount of wins is a terrible idea. Win all the games and let the chips fall where they may. That said, it feels like we should be cheering for a 7 or 10 seed on selection sunday (especially since the second tier of highly ranked teams seem pretty vulnerable this year), and if we should see our spot in bracketology drop a few slots over the coming weeks, don't panic, it may well be a blessing in disguise, seeing as we're sitting on the 9 line right now.
*statistics via brackethelp.com
February 23rd, 2017 at 3:29 PM ^
February 23rd, 2017 at 3:34 PM ^
If we were to win our last three games and a couple in the BTT, we might be able to play our way up to a 6 seed.
February 23rd, 2017 at 3:30 PM ^
Seed us please!
February 23rd, 2017 at 3:33 PM ^
If we end up as a 8/9 seed, I want Gonzaga as the #1 seed in our region. I think that's your best chance of advancing to the Suite 16. Play a team that hasn't played a P5 team since December.
February 23rd, 2017 at 3:38 PM ^
Suite 16 is an interesting concept, do explain more
February 23rd, 2017 at 3:42 PM ^
It's the 3rd round of the tournament when you've had a really long day...
February 23rd, 2017 at 3:44 PM ^
It's like a fancy Motel 6
February 23rd, 2017 at 3:51 PM ^
It's the wing of the hotel Louisville brings their hookers back to on the second weekend of the tournament if they make it. You should see what they do if they make the Final Fhore.
February 23rd, 2017 at 5:13 PM ^
February 23rd, 2017 at 6:21 PM ^
Oh, beautiful. The things that games involving Rutgers cause people to do, eh?
Did you stay at the Embassy Suites or at the Embassy Sweets down the road?
February 23rd, 2017 at 11:05 PM ^
February 23rd, 2017 at 3:39 PM ^
But I do agree that we need Gonzaga as the 1 in our region if we end up as an 8 or 9
February 23rd, 2017 at 4:11 PM ^
Be a 7 and potentially see a team like Oregon or Louisville as the 2nd round challenge than Gonzaga.
Top to bottom, the Zags are the best team in CBB. They have the motivation, the talent, and they've beaten some great teams this year including Zona.
February 23rd, 2017 at 5:04 PM ^
Yea wtf do people want to play Gonzaga? They would annhilate us with that size and skill. No thanks. Becasue they're a mid-major that makes them more an attractive matchup? Faulty theory.
February 23rd, 2017 at 6:15 PM ^
February 23rd, 2017 at 5:18 PM ^
February 23rd, 2017 at 3:34 PM ^
February 23rd, 2017 at 3:40 PM ^
unless a 6 is a possibility
February 23rd, 2017 at 3:54 PM ^
February 23rd, 2017 at 5:39 PM ^
If they are the 1 seed, they are going to be in the West Regional, meaning a cross-country trip for Michigan possibly to Sacramento. I'd rather play a 2 seed that gets shipped out of their region into something closer to Ann Arbor. They are 'neutral' site games, but travel and fan support can make a difference.
February 23rd, 2017 at 4:12 PM ^
Since these teams sit at the far tail end of the distribution, the difference between 1s and 2s is the largest difference between any seed line, and it's easier for the committee to "seed" them correctly. So there is also less noise in the actual seeding as well as performance.
The other piece of this is that there is also generally a huge difference between 15s and 16s, such that a 16 has never won. They are often the outliers that got super lucky to win conf tournaments. 15s aren't typically nearly as bad and they've knocked off 2 seeds like 8 or 9 times (but have never advanced beyond that). So not only does a 7/10 seed typically have a much easier matchup with a 2 seed, they actually have a chance to play a 15 seed.
I didn't realize how low the advancement rate of an 8/9 is though, relatively. I'd be curious to know what the expected number of wins for each seed is because a lot of 8s have made the final four since if they beat the 1 seed, they get their path. UK in 2014, Butler in 2011, UNC and Wisconsin in 2000, etc.
Despite the high sweet 16 rate for 10s, no 10 seed has EVER made the final four and only two 7s have ever made it, which means more than twice as many 8s have made it than 7s and 10s combined. Small sample sizes at the point, but one could argue a 7/10 is better for sweet 16, but an 8/9 is better for a final four.
February 23rd, 2017 at 4:28 PM ^
Dunk City says hello! Many of us saw the first-ever 15 seed to advance to the Sweet 16: Florida Gulf Coast, in Dallas, vs. Florida.
Hint: Trey Burke? Kansas? Nik Stauskas in the Elite 8 vs. Florida? :-)
February 23rd, 2017 at 4:35 PM ^
I stand corrected! I clearly wasn't paying attention to anything else going on in that tournament. Forgot they were a 15 seed. Point remains that 7/10s have a chance at a much easier opponent than the 2 seed.
February 23rd, 2017 at 5:21 PM ^
I think it's more a reflection that of your one seeds, there's a good chance you have one to three truly elite teams that are simply unlikely to lose a game to anyone that's not good enough to earn a 4/5 seed or better. This year may not be the best example of that with Kansas having a couple rough loses and Villanova dropping one to Marquette. But in general, the difference in quality between the best one seeds and the two worst two seeds will often be quite significant.
February 23rd, 2017 at 11:03 PM ^
that was my first point. That on the far end of the distribution the gap between 1s and 2s tends to be much larger than any other seed line (except probably 15s and 16s, as well), because the 1s include the elite outliers.
February 23rd, 2017 at 4:14 PM ^
I would be thrilled if the final bracket were like what Joe Lunardi posted today. First round against Dayton, potential Gonzaga in 2nd round (yes, I understand they're better than us, but the pressure will all be on them and we recently saw an undefeated 1 seed fall in the 2nd round when they finally played a P5 team), Sparty on the bottom side of the bracket and an Oregon team I saw Ricky Doyle dominate as the #2 seed.
Not saying I'd predict us to the Final 4, but I wouldn't be shocked.
February 23rd, 2017 at 4:26 PM ^
Not saying I'd predict us to the Final 4, but I wouldn't be shocked.
I would be... like really, really shocked. That would require us to pull off two major upsets.
February 23rd, 2017 at 4:58 PM ^
Possibly only one upset. We might get help from other teams in our region.
February 23rd, 2017 at 11:07 PM ^
dominated them two years ago, which is an eternity in college basketball. Almost an entirely different team. I do agree that they'd be an attractive 2 seed in our bracket. The PAC12 is terrible, Oregon's metrics aren't great, but I don't love how they're perimeter oriented. Seems like our guards really struggle to keep good guards in front of us and our defense breaks down from there. And if they can defend the perimeter, not great for us.
February 24th, 2017 at 1:17 AM ^
February 23rd, 2017 at 4:21 PM ^
February 23rd, 2017 at 4:27 PM ^
February 23rd, 2017 at 4:40 PM ^
Obligatory:
February 23rd, 2017 at 4:44 PM ^
I opened this thinking the OP was discussing BTT seeding and where we would want to be seeded to improve our NCAA tournament standing. Discussing NCAA seeding now is tempting fate and Murphy a little too much.
February 23rd, 2017 at 4:45 PM ^
Preferably in a bracket that has house rules which do not allow for attempting to rebound your own shots.
February 23rd, 2017 at 4:46 PM ^
Granted, I have only watched a few games of each, but I would rather see Baylor as the one seed in my region than Gonzaga. Of the teams with a realistic shot of being a #1 seed, here is the order I would least like to play:
1. North Carolina - I can imagine the rebounding margin now..
2. Villanova - They are us, but just much, much better
3. Kansas - Their guards would be too much to handle
4. Gonzaga - Big guards would be tough for Walton, and Karnowski would have a day.
5. Oregon - Dillion Brooks is good, but have been caught in tight games with average teams.
6. Baylor - I feel like they would destroy us on the boards, but sometimes struggle scoring
7. Louisville - Have struggled a bit in the ACC
8. Arizona - I just don't think they are very good.
February 23rd, 2017 at 4:50 PM ^
My preference would be 10 or worse, instead of 8-9.
February 23rd, 2017 at 4:57 PM ^
That's the order of the remaining possibilities that seems most beneficial to us.
February 23rd, 2017 at 5:04 PM ^
February 23rd, 2017 at 5:09 PM ^
February 23rd, 2017 at 5:37 PM ^
February 23rd, 2017 at 6:51 PM ^
February 23rd, 2017 at 6:58 PM ^
♪ yar true that yar wrong but yar wrong that yar true ♪
February 23rd, 2017 at 7:04 PM ^
I'd prefer a 10 seed to an 8/9 that is for sure. Hell I'd prefer an 11 or 12 seed without the play in to an 8/9/10 seed.
February 23rd, 2017 at 7:26 PM ^
are fine as long as we aren't in the play-in portion.
Otherwise, I think 6 or 7 are sneaky great places to be depending on the 2, and 3 seeds always seem vulnerable.