NCAA tournament: Last-minute bracket advice

Mar 17, 2011

Note to Readers: As you probably have noticed, many familiar discussion-related features on have moved as part of the recent redesign. We apologize for any inconvenience and confusion. Here are a few links to help you find your way around:

- The weekly chat schedule
- The Discussions and Debates page, which will feature recent and upcoming chats and a full daily schedule.

Thanks for bearing with us during this transition. If you have questions or concerns, please feel free to send us an email.
The discussion: Need last-minute help with your bracket? Just want to talk about the NCAA men's basketball tournament? Join Brad Carlin, Kenneth Massey and Jon Wertheim for some traditional and non-traditional analysis of the 2011 tournament.

Jon Wertheim here. Hope everyone's well - and hates the 68 draw as much as I do. For what it's worth here are some tips re: how to use investing to aid your bracketeering

go hoosiers...oh, wait....

Hi everyone -- this is Brad Carlin at the U of Minnesota.  I've made my hoop picks and just got out of  a boring funding-related meeting here, so I'm sure the next 30-60 minutes are bound to be more interesting!  ;)

if your picks are due and haven't given the matter enough thought, pick your champion and then go to the poologic calculator,


It's a little Flash program that will do all the picks for you, subject to the rules of your pool (the usual 1-2-4-8-16-32 scoring is in there by default) and subject to your picking a champ.


I advise changing a few picks here and there so you don't just pick all favorites, which is what the computer will advise!

What would likely be final score between Ohio State and Kansas in final game?

Here are the Sagarin ratings.

Sagarin's "predictor" ratings (last column), which are the best for guessing future outcomes, have OSU about 1.5 points better than KU.  So this means his best guess is that OSU would be KU by 1 or 2 points, were they to meet on a neutral court

Big assumption both top seeds make it through. But I'd tip OSU, say 78-72....

According to the Massey model, 75-72 Ohio St.

I think patterns of Injuries, Suspensions, and Illness during the season tend to bias the ratings. How can one get a handle on this? Recent injuries are in the news, but teams that are healthy relative to their season performance are harder to identify.

Great question. This is a prime example of data you can use that the selection committee hasn't necessarily accounted for. You have more creativity and more time than the suits on the committee. Use it to your advantage....

Yes I quite agree; the poologic site wants me to take BYU to the regional semi... but didn't they just suspend a key player who was most of their inside game?  The computer ratings all assume that the same team that played during the regular season is going to show up in the tourney.  Injuries and suspensions obviously change this reality.  If you have knowledge of these things, using it will give you a small edge on your competition  :)

Is there a smart bet in the 5-12 seed range to make the Final Four? Or at least a region where it's safer to not pick the 1 or 2?

Again Richmond over Vandy is the hot pick....Utah State over Kansas State intrigues, too....These are the games that really help your pool score

Yes I agree re: Richmond and USU; both are just 2-point dogs despite being 12 seeds.  Clemson is just a 1 point dog and they just won big in the play-in game.  The only 5 that looks secure to me is Arizona, which is favored by 6 but many ratings I've seen indicate the spread is maybe closer to 9 there.

Statistically, Ohio State looks to be a cut above the rest of the field. But "How to win your pool" articles always say you don't win if you pick tournament favorites to win it all. So what do I do with the Buckeyes?

You can win your pool picking the favorite - especially of your pool has no "upset bonus" ...but a little variance is essential/ if you're hellbent on OSU, sprinkle your field with upsets, i.e. Richmond over Vandy. I suspect more people  are picking against the favorite  here than any 5/12 game in tourney history!

It depends on your goal.  If you want to win your pool, and there are a lot of participants, then you probably need to take some chances.  If you just want to finish in the 90th percentile, then go with the favorites.

The Southeast bracket is driving me crazy. I just can't trust Pitt after years of struggle to get past the 16. Florida's rank is way too high. BYU is a one man team. A lot of people seem to be going to Michigan St just for the Izzo factor, but 14 losses? HELP.

Definitely the soft quadrant. I don;t trust BYU, not with one player so responsible for the overall output. I don't trust Pitt either. Inasmuch as there's a George Masonesque  total Cinderella reaching the FF, it's coming from here....

I like Kansas State as a sleeper.  Nobody is giving them much chance, and don't forget, they have tons of talent and were preseason top 10.  They started the season poorly, but have been improving.



This is Brad:  can you really pick KSU when they're favored by just 2 in Round 1?  I'd hate to be relying on a coin toss type outcome in the very first game!


Totally agree that region is a wreck; I have Fla coming out of there but not much confidence in that pick

The Gophers beat UNC early this year, so I have UNC going out in round 2 (because I don't have the guts to have them lose in round 1). Thoughts on UNC losing early?

UNC has a 54% chance of making it to the sweet 16.


for all such probabilities.

ken I'm playing 2 sheets this year, and I'm a contrarian so I'm playing two 4 seeds:  Texas and Kentucky.  Both are in the top 10 for Sagarin predictor so seemed like "bargains" from a contrarian view.  I hesitate to add that my pool has an "insurance" feature, and I have insured both of these teams, so if they lose prior to the FF I get the team that beat them instead (which might well be Duke and OSU/UNC, of course).


Anyway, looks like your ratings are not as enamored of UT or Ken?

What do you think of SDSU?

They can't beat Duke, but they may get to the elite 8.  I think Temple is a possible upset pick vs SDSU

I don't see them beating Duke. Can they live up to their exalted seeding? Perhaps....Call it East Coast bias but, boy, do a lot of people have them losing early.

For Jon W: Your book promises to explain "The Hidden Influences Behind How ... Games Are Won." What's the most underrated, hidden or otherwise unknown influence in NCAA tournament games? Or do I have to buy the book?

Buy the book! No, seriously, thanks for the interest.  I posted a few Scorecasting-y tips here:


Don't exaggerate momentum. Also a prior poster asked about accounting for injuries v/v RPI. This is precisely the kind of data a creative fan can access that eludes the selection committee.....

I've heard that long distance travel West to East or East to West and the resulting jet lag affects pro outcomes. Does that extend to NCAA and predicting upsets?

Very little data to suggest that jet lag affects road teams. (The home winning percentage doesn't budge when teams play nearby - Ravens/Redskins - versus when they fly across country. .... I've always wanted to study arena optics: that is, when a team accustomed to a small gym has to shoot in, say, the massive old RCA Dome, does that impact percentages - controlling for defense? Something to consider perhaps...

Wertheim here. Question for the vox populi: Can anyone outside Indianapolis HQ make a credible case for 68 teams? To me, those first four games rob the tournament of symmetry, take away from a clear start, and do nothing to increase interest/excitement....

If anyone's interested, my 1st round picks include

Georgia, Florida St, Gonzaga

Can they really beat North Carolina? They're kind of a trendy upset pick, but I'm having a hard time believing.

Watch out - they may not get past UGA

Huskies (Washington, not my UConn guys) are favored by just 5.5 over Georgia, so about a 2/3 chance to win (ie 1/3 chance to lose)

Are there any injuries in the mid-range seeds (say 9 to 13) that could hurt their upset chances?

Not an injury, but Purdue suspended a player (Kelsey Barlow) yesterday. Not only a loss of personnel but potential chemistry disruption. A good example of why you may as well wait as long as possible before submitting your picks. Folks who filed Monday were unable to take advantage of this additional info...

I think there is room in the statistical modeling of these games to account for things like autocorrelation -- e.g., a team like KSU that started poorly but is playing better lately (UConn might be another example).  A lot of the current methods probably assume the games are all independent, but that's probably an oversimplification.  Time series methods could be used to account for the 'streakiness' in some teams' performance and thus do a better job of predicting who's going to persist in the tourney.

Ken, have you tried to model autocorrelation in this way?  I think there are a few JASA articles on this subject in NFL football...

Chances the Tigers pull off the upset?

Have a hard time seeing it. But your question provides an excuse to plug this piece by Sean Gregory:,8599,2059230,00.html

and this companion piece by Zach Kwartler:


Great halftime reads...

oops put this answer in the wrong spot.  Kentucky is favored by 13 so :

> spr(13)
 P(favorite wins) = 0.871 ;  underdog has 1 chance in 7.74

Ken and Jon did not have a good word for Pitt. But isn't Pitt a good contrarian pick for champ? A one seed that is only getting 5 or 6% of the champ picks in the big online pools that report such data.

Yes, according to the computer models, they are a good pick.  However, from personally watching them, I do not feel comfortable with Pitt.

I concur.  I often advise taking "the weakest #1" as a fairly safe yet contrarian pick for the top of your sheet. But with Pitt's struggles down the stretch it's hard to feel great about them.


Keep in mind that how "contrarian" you want to be depends on the size of your pool.  If you're only playing against 10-15 other guys, you can play it safer and just take OSU or KU or Duke.  But if you're playing against 200 other guys (as I am), now you need to get away from the pack a little more and reach down into the 2-3-4 seeds for your champ, I would say.

The various rating systems are all over the place on Belmont. What gives?

Belmont has played only three top-100 teams, losing them all.  But they've dominated everybody else.  So there's not much good info to evaluate this team. 

Mo' Money from the TV networks in the contract. The symmetry was ruined when they didn't give up the at-large bid when then MWC/WAC split occurred and created another automatic bid. Once they went to 65, they ruined the best American sporting event.

How about a 256 team monster tournament, where the first weekend is a 4-team tournament at home sites of the top 64 seeds?

I did like having a game or two to watch the last two nights. Kind of a nice warm-up act. Maybe the teams that played in the games wouldn't want to hear it put that way, but that's how I felt.

Fair point.

so is the NIT totally dead?  I thought NIT games were what we used to watch on Tues and Wed nights before the Big Dance?

I'm vaguely aware that CU was in the NIT (as they were snubbed by the Big Dance) and that my undergrad alma mater Huskers were promptly shucked, but that's about it.  ;)

good luck today everyone!

Besides the quality of the games -- did the NCAA structure these games correctly? Should they have reseeded the teams?

I frankly don't understand the logic in having play-in games with non-16 seeds.  Why is it fair to require Clemson to beat a good opponent (UAB) just to get into the tourney proper, where as other similar teams (also seeded 12) just got in without a play-in?


I think the women's tourney made the right call by NOT expanding beyond 64.  If you're going to expand again, you can't use these ludicrous piecemeal solutions.


Plus, even President Obama didn't make finalize his picks until after the play-in games were over  ;)

In This Chat
Brad Carlin
Brad Carlin is a biostatistician and department head at the University of Minnesota who has developed a side expertise in statistical analysis of NCAA basketball tournament bracket strategy. His early work was implemented in a the Poologic Calculator, and he has been featured in Slate and in the New York Times and the Wall Street Journal, among other publications. Carlin's professional specialty is the analysis of clinical trials and of spatially associated data such as cancer rates by county.
Jon Wertheim
Jon Wertheim is a senior writer at Sports Illustrated and co-author of the recent best-seller Scorecasting.
Kenneth Massey
Kenneth Massey produces the Massey Ratings, which provide objective team evaluation for professional, college, and high school sports. His college football ratings have been a component of the Bowl Championship Series since 1999. Massey is an assistant professor of mathematics at Carson-Newman College in Jefferson City, Tenn.
Recent Chats
  • Next: