Basically the same thing as last week. The newest Bracketology can be found here. I only did one set of simulations this week, with my slight adjustments (this time shown in points, rather than a rating), but I ran 20,000 of them.
Some quick notes about this particular bracket:
UNC has the misfortune of having to face West Virginia in the second round; it's somewhat unusual for a #1 seed to have to face the eight best team in the country (in this simulation) in the second round.
All the teams in the South bracket, most notably Michigan St., Marquette, Villanova, and Purdue, are lucky to have Oklahoma as the #1 seed in their region. The Sooners, 18th in Pomeroy, are an incredible weak top seed. Oklahoma also catches a bad break, with Arizona St., rated four spots ahead of them, as a potential second round opponent.
Missouri, with their big comeback win tonight, should end up higher than a 5 seed. For now though they're stuck with a ridiculous draw, facing BYU in round one, and then potentially an underrated UCLA team on the first weekend. If they had, say, Purdue's draw (Siena, Villanova/Northern Iowa), they'd easily be over 1.5%.
The most interesting team here, by far, is Memphis. The Tigers, second in the nation in defensive efficiency, held Gonzaga to just 37% from the field on Saturday night, blowing them out in Spokane. #5 in the country might seem kind of high, but Memphis has actually been better than their Pomeroy rating since moving freshman Tyreke Evans to the point.
Using their current KPom rating to back-project their first nine games, Memphis was projected to win each game by an average of 16.9 points, but only won by an average of 11.2. Since the Syracuse loss though, and the accompanying position shift, they've been expected to win by 15 ppg, but they've won those 14 games by an average of 18.5.
That's a 9.2 point swing, from underperforming by 5.7 ppg to beating the projections by 3.5 ppg. They don't have any more out of conference tests, but vists to UTEP, UAB, and Houston (71st, 33rd, and 50th in Pomeroy, respectively) should give us more of an idea of just how good they are. For now, though, they look like a legitimate Final Four threat.
16 comments:
It will be nice to see Memphis in the Final Four. However I don't think the Tigers got it in them.
Who does UConn lose to most often in those simulations, if there is an easy way to tell? How can Duke still be 1 in Pom ranking after some of these losses and close calls?
How can you compare the probability of winning (W%)to the odds being given to each team? For example, how can you tell if Pittsburgh, with a 13.20 W%, is over-valued or under-valued by the odds makers at 12 to 1?
"Who does UConn lose to most often in those simulations, if there is an easy way to tell?"
Not sure how I'd do that, but I can tell you how often they're eliminated in each round:
1R: 4.6%
2R: 18.3%
S16: 31.2%
E8: 17.2%
F4: 13.4%
NC: 6.8%
Nat. Champs: 8.4%
Being in the same half of the bracket as Mizzou, BYU, and UCLA, they have a pretty tough draw. Memphis is also the second best 2 seed.
Duke is still in first because they were so far ahead going into the Clemson game.
"How can you compare the probability of winning (W%)to the odds being given to each team?"
For Pitt, just do 1/(1+12)=7.7%. So according to this Pitt is undervalued at 12:1. They do have a pretty easy draw though.
UNC is a 2 point favorite at Duke. Methinks the Pomeroy opinion on Duke is not popular in Vegas.
It certainly is a strange line, but it's fairly easily to explain away within the context of public opinion. Duke has had three out of four "poor" results with a loss at Wake, getting obliterated by Clemson, and taking Miami to OT at home... whereas UNC has been on a reign of terror since that @WFU loss with the lone exception of a close game at FSU (but it's the ACC, right?).
Anyway, what I'm getting at is that there has to be at least a smidgen of value in Duke +1.5 tomorrow, as KenPom points us to. There's no way in fuck that UNC would be 8-10 point favorites at home.
I don't know how difficult it is to run these sims, but would it be possible to run one for this point last year?
I'd be curious to see how that would have turned out.
It'd just be easy to run, I just don't have the Pomeroy ratings from back then.
It's frustrating that he doesn't archive the results, but I guess a day by day archive of all that data would get pretty ridiculous pretty quickly. I dunno how complex his algorithms are for computation but the computation could probably be altered to be done server-side... in whatever similar things I've done the entire process is probably completed in under 3 seconds.
It's not like I spend a lot of time thinking about this or anything.
I'm a sucker, you just sold me on BYU at 500-1.
A few months ago some comments broke out on aaodds regarding pace and tourney success. Well, Hambone's intuition appears to have been deluded by anecdotal evidence, per this article.
Which book has the best odds for Memphis right now? I see 35-1 at VIP but I'm not sure they accept U.S. players. Otherwise, the best I see elsewhere is 30-1.
I'm seeing the same thing as you on Memphis, James. A bunch of sites don't have anything up right now, maybe they will tomorrow, but I doubt they'll put Memphis at much higher than 30-1.
So how's our team to watch Memphis doing in the new bracket sim now that they are up to 2nd in Pom and a #2 seed in Lunardi's bracket?
Patience.
Post a Comment