How can you predict the upsets? Five tips on picking your bracket

Witness 2008. Siena over Vanderbilt? Check. And look, there's Davidson in the Sweet 16. If only I'd stuck with Stephen Curry one round further.

VCU over Duke in 2007? Check. And another 10 seed, UNLV, in the Sweet 16.

Remember when George Mason met Wichita State in the 2006 Sweet 16? I got it half right. If only I'd picked JimLarranaga's crew to pull off more than just its first-round game.

But my truly golden year, my tour de force moment, was 2005. Over the course of one memorable Friday night, I watched both No. 13 Vermont (over Syracuse) and No. 14 Bucknell (over Kansas) make me look like a prophet while making new friends at an Indianapolis sports bar.

(Note: Last year I was on sabbatical during the tournament and didn't post a bracket on this site, but I do have witnesses that can confirm my Cleveland State-Wake Forest and Western Kentucky-Illinois picks.)

By now, you're probably saying to yourself, "I, too, want to become an expert upset picker. I want to dazzle my friends and co-workers. Do tell us your secrets."

Follow these five tips:

1. Identify vulnerable major-conference teams. I'm talking about teams that are either A) overseeded or B) inconsistent and unreliable.

This year's poster child for A) is Notre Dame, which has played like a No. 6 seed over the last two weeks of the season, when it went on a six-game winning streak and reached the Big East tournament semifinals. Before that, the Irish were 17-10 and in danger of missing the tournament altogether. They lost to Northwestern, Rutgers and Loyola Marymount (at home). They're an eight seed in a six seed's body.

As for criteria B), you're looking for teams that may have won some big games, thus enhancing their image, but were otherwise up-and-down. I'm thinking of a team like sixth-seeded Tennessee, which has the distinction of having beaten two No. 1 seeds (Kansas and Kentucky) while also losing to two sub-100 teams, Georgia and USC, and getting crushed at Vanderbilt and Florida. The Vols overachieved in even getting this far, which makes them vulnerable against any decent foe.

Others that fit this mold: No. 2 seed Villanova, No. 3 seed Pittsburgh, No. 4 seed Purdue*, No. 5 seed Michigan State, No. 6 seed Marquette, No. 7 seed Oklahoma State.

(* Purdue is vulnerable because they stink without injured star Robbie Hummel. They were a Final Four team before that.)

2. Look for dangerous underdog opponents for these teams -- teams that dominated their league, are experienced or pulled off at least one notable non-conference win.

Both of my "poster child" teams above drew those very type of teams. The Irish have the misfortune of meeting Old Dominion, which knocked off Georgetown on the Hoyas' home floor and won both the Colonial regular-season and tournament titles. They even play the same type of slowdown, half-court style that Notre Dame switched to during its hot streak. Match made in heaven. Ditto the Vols' opponent, San Diego State, which won the rigorous Mountain West tournament.

Meanwhile, poor Purdue, as if it doesn't have enough to deal with, draws the archetypal Bracket Buster, No. 13 Siena, a veteran team that's won first-round games each of the past two years. And while No. 11 Washington hardly qualifies as a Cinderella, it's a hot team with go-to guys that could be a bad matchup for Marquette.

3. Decide which No. 1 seed you trust least, then see if there's a potential Sweet 16 or Elite Eight sleeper in its region.

Only once in 25 years of the 64/65-team bracket have all four No. 1 seeds made the Final Four, in 2008, and it's a pretty safe bet lightning won't strike again this year. Another area where I've had pretty good fortune over the years is in correctly tabbing the customary surprise Final Four team. My successes: No. 4 seed LSU in 2006, No. 4 seed Louisville in '05, No. 3 seed Georgia Tech in '04 and No. 3 seed Syracuse in '03. Obviously, in every case, I thought those teams were sneaky good, but I also felt they'd landed in a region with a vulnerable No. 1 seed.

As for this year's top seeds, some people don't trust Kentucky's youth. Others think Duke is overrated. And even the biggest Kansas proponents can't deny what a tough draw the Jayhawks garnered. For me, however, the most vulnerable is Syracuse, which didn't play nearly as well late in the season as it did earlier and may now be without center Arinze Onuaku for at least its first two games. Meanwhile, one of my two favorite sleeper teams in the whole field, BYU, is sitting on the other side of that region.

4. Don't go overboard.

While big upsets usually make for the most memorable tourney moments, it's easy to forget that there generally aren't that many of them. Usually about six to eight double-digit seeds make it out of the first round, of which two or three at most reach the Sweet 16. And the overwhelming majority of those will come from the 10 to 12 range. No 16 seed has ever won a first-round game, while No. 15 seeds are 4-96 all-time. No. 14 seeds have won 15 percent of their first-round games, No. 13 seeds, 21 percent.

With that in mind, while I do think Pittsburgh (which lost to Indiana and suffered several lopsided league losses) and Villanova (which has struggled down the stretch) are upset-prone, I don't have the stones to pick Oakland over the Panthers or Robert Morris over the Wildcats. I'll wait a round and bank on sixth seed Xavier (over Pitt) and seventh seed Richmond (over 'Nova).

5. Be careful about who you bounce.

Most bracket pools tend to place far greater weight on the later rounds than the earlier rounds. So before picking a high seed to go down early, you better be certain (or as close to certain as possible) that it won't come back to bite you. For example, in a straight-up pool where first-round games are worth one, second-round games are worth two, etc., you may earn a sweet round of high-fives for correctly picking, say, Sam Houston State to beat Baylor, but pool-wise it's of no more benefit than picking West Virginia to beat Morgan State. However, if your hunch proves incorrect, and Baylor turns around and reaches the Final Four -- you're toast. You may have just cost yourself 30-plus points.

It's a slightly different story in a weighted pool, where you get more points for correctly picking lower-seeded teams. For instance, in a scoring system that multiplies round by seed number, picking 12th seed Cornell to win one game nets you twice as many points (12x1 = 12) as picking No. 1 seed Kentucky to win its first three (1x1 + 1x2 + 1x3 = 6). But even in these pools, there's usually a significant "bonus" for each correct Final Four pick and national champion, so you still can't afford to go bouncing No. 1 and 2 seeds like mad.

In the interest of full disclosure, I must confess that for all my professed upset mastery, only once in 24 years have I actually won a bracket pool. And the main reason is that I can't seem to adhere to my own advice -- mainly, the fifth tip listed above.

Case in point: two years ago. Looking back, I could not have asked for a much better start to my bracket. In fact, I nailed seven of the eight Elite Eight teams. The problem was the eighth.

On the night the bracket was announced, I wrote somewhat prophetically on SI.com that "this may be the most top-heavy bracket since 1999, which was the last time three of the four No. 1 seeds reached the Final Four." I came very, very close to picking all four No. 1s (which to that point had never happened) but ultimately left one out: Kansas, the eventual national champion. Even worse, the team I tabbed in its place, O.J. Mayo-led USC, lost in the first round.

Cincinnati Enquirer columnist and SI.com contributor Paul Daugherty seems to think BYU is my USC pick of 2010. He wrote that I violated his own personal commandment for winning at bracket pool: "Thou shalt not overthink it." He's probably right. It wouldn't be the first time for me.

Which brings me to my last and possibly most important piece of upset advice (even if I don't always adhere to it myself): Go with your gut. If you like Utah State to reach the Sweet 16, then go with it. The second you scratch it out -- you're losing that game. Guaranteed.

You've got to be a little bit crazy to make some of these picks -- but that's what's so fun about it.