Last year, Duke was easily the biggest value to win a title entering the tournament. Fourteen percent of all #1 seeds have won a National Championship, but Duke was picked in roughly 5% of brackets, despite many statistical systems, including Ken Pomeroy’s, having them as the best team in the country entering the tournament (this was in line with my experience where I played Duke heavily and very few in those pools had them to win it all). It’s hard to know for sure why the masses thought so poorly of Duke, but I have a pretty good guess.

Even though Mike Krzyzewski had a reputation of going deep in tournaments for much of his career, the Blue Devils had been a disappointment in March for the previous five years, likely leading to selector fatigue (everyone had probably been burned by picking Duke over this span). Duke had lost twice as a #1 seed in the Sweet 16, lost in the 2nd round as a #2 seed, and had a first round loss as a #6 seed over that span. They were an incredible -5.8 wins below expected based on their seed from 2005-2009.

But as we saw, that didn’t matter in 2010, and Duke’s not alone. When I looked at all teams that played in three straight tournaments between 2006-2010, and compared the Performance Above Seed Expectation (PASE) for years N-1 and N-2 to year N, the correlation coefficient was -0.09. That means there was basically no relation between how a school had performed relative to expectations in the two previous years, and the current tournament. Yes, some teams that played poorly for two years did so again, but just as many played well.

Seventeen teams had a PASE of at least -1.0 wins or worse over a two year span; 11 of them performed above average in the third year, including two of the last three national champions. As a group, they were +7.7 wins above expectation.

Maybe you didn’t like that method of thinking about how people may perceive teams. What about just experience in the tournament, whether it was good or bad? Again, I don’t see anything in recent years that would suggest you should bump up a team because it has big game experience. The PASE for teams that were top 4 seeds but were not in the NCAA tournament the year before (2006-2010) is +1.4–they basically played to seed on average. Top 4 seeds who were repeaters as top 4 seeds from the previous season: PASE was +0.1. Again, basically in line with expectation, and no indication that merely being seeded highly the previous year helped.

So where does that leave us in 2011? Well, without going into full detail of the data in every round and for every team available on a site like TeamRankings.com (since it is a subscription service), I can tell you that the undervalued (in terms of percentage of picks to advance in online pools) teams at each of the top seed levels are Pittsburgh, San Diego State, Purdue and Texas. What do those teams have in common? None has any recent history of elite tournament success.

Pittsburgh is a perennial disappointment because they have not advanced to a Final Four, and have underperformed as a high seed. They are a poor man’s Duke from 2010. I remember when Michigan was a perennial disappointment in the tournament in the late 1980′s, until they went on a run with Glen Rice in 1989 as a #3 seed, and played another, Seton Hall, that had no history of recent success.

San Diego State is a relative unknown, and has no tournament success to their name, losing last year in the first round as an #11 seed. Purdue is another team that has not appeared in a Final Four any time recently despite having several good regular seasons. Texas has made one Final Four this decade, but has been a high seed frequently and lost earlier than people expected on a few occasions, and has no history as a national champion.

In some ways, I think people have a hard time envisioning making a pick like one of those teams when they are making instantaneous decisions on a bracket, because they haven’t seen it before. Kansas is safe, Duke is safe (again), Ohio State is safe (they were in the finals a few years ago), even Connecticut is safe because they’ve won national championships in the last decade.

Part of winning a bracket pool (other than the biggest factor, luck) is taking advantage of inefficiencies in the market, and separating yourself from the pack. I’m not saying that I think any of the above four teams are the most likely to win the national championship. But if I’m trying to place in the top 1% of a pool of people, and it doesn’t matter if I finish 5% or 95% otherwise, I’m taking calculated risks where I think I have pot odds. Texas (4th), Pittsburgh (5th), San Diego State (6th) and Purdue (8th) are all in the top 8 in what I think are a pretty good set of rankings. That’s in reasonable range to advance to a Final Four and win it.

I may think Ohio State and Kansas are rightfully the favorites, with Duke right behind. But I also know that even heavy favorites have less than a 50% chance of playing on Monday Nights in April. I’m throwing my lot in with some unloveable losers who we simply cannot win it.

[photo via Getty]