Commentary

Breaking down bracket predictions

Updated: March 15, 2011, 6:09 PM ET
By Christopher Harris | ESPN.com

Is it an insult to be called Mr. Chalk?

I don't mean in terms of one's complexion, because look at that author's thumbnail photo. There's no one chalkier than me. No, I mean: Is it lame to fill out your NCAA bracket mostly picking favorites?

That's sure my impression. There tends to be honor among thieves and admiration for the hubris it takes to boldly assume you can see the future, and thus envision the many upsets that will no doubt rattle the windows of several sporting arenas across this great land. The truth, of course, is that there will be upsets during the next three weeks. Unfortunately, history teaches us that it's mighty difficult to actually predict which games will feature those upsets. So are you smarter taking more chalk than is "cool," under the assumption that as the rounds wear on, the favorites will tend to rise to the top?

I'll try to help answer that question in this column, while at the same time using my own bracket and the brackets of ESPN experts AJ Mass and Brian Gramling to help isolate those games that seem ripe for upsets. In fact, to better assist you in filling out your own ESPN Tournament Challenge bracket, here's a look at the three of ours (note they won't be viewable until they lock at the first game time on Thursday, but feel free to join our Fantasy Editorial Group to play along with us):

Christopher Harris' bracket | AJ Mass' bracket | Brian Gramling's bracket

What follows are words from the horses' mouths: Queries into the wisdom of picking upsets, and a discussion of why the experts took the teams they did. Enjoy!

Second Round

How many second-round upsets is too many? Let's look at the numbers. During the past 10 years, we've seen as many as 13 better seeds get knocked off in their first game, and as few as four. In that span, the average number of better seeds to lose their first game is 8.2, though in the past three seasons, we've seen 10, 10 and eight go down. Among our experts, I have 10 better seeds losing at the outset, Brian has 11, and AJ has a more restrained seven.

In theory, the optimal second-round picking strategy would be to select as few upsets as possible, especially among top-four seeds. That's because those "protected" (i.e., top-four) seeds are historically likeliest to make deep tournament runs, and if you mistakenly bump one of them off and then see them roll into the Final Four, well, you're in trouble. How many protected seeds do each of us have losing in the round of 64? I have only one such seed losing (Wisconsin), AJ has two (Texas and Wisconsin) and Brian has zero. In the past 10 tournaments, an average of 1.4 protected seeds have lost in their first game, including two last year, and one the year before that. So as you can see, while our overall "upset" count in the second round seems high, the impact for wrong picks will hopefully be relatively minimal.

[+] EnlargeJustin Harper
Rich Barnes/Icon SMIJustin Harper shoots 54.3 percent from the field overall, including an impressive 46.5 percent from 3-point land for perimeter-oriented Richmond.

The three of us have three second-round upsets in common. We have ninth-seeded Old Dominion beating eighth-seeded Butler (hardly a shocker, given how close the two teams are), and we each think that 12-seeds Richmond and Utah State will upset Vanderbilt and Kansas State, respectively.

Says Mass: "Richmond is very solid defensively and makes few mistakes with the ball, while Vandy doesn't force any turnovers. The Spiders will dictate tempo to win." Gramling concurs: "The Spiders' top three scorers are all seniors and they're all shooting at least 43 percent from behind the arc." As for Utah State, Gramling is unequivocal: "The Aggies are grossly underseeded, with a 30-3 record and holding 19 of their opponents under 60 points, including three under 40 points! Tai Wesley is a star and will outperform Jacob Pullen and his Abe Lincoln beard." It's true that when the beard is on the floor, the Wildcats outnumber their opponents 6-to-5, but nevertheless I'm with these guys: Utah State is a great defensive team and they can impose their slow-down tempo on anyone. There's a good chance Kansas State gets frustrated in this game.

The biggest second-round surprise anyone picked was Mass taking 14th-seeded Patriot League champ Bucknell to beat UConn. He says, "Bucknell isn't nearly as talented as Connecticut, but I think five games in five days during the Big East tournament took a physical toll on Kemba Walker and the rest of UConn." Mass also has two 13-seeds winning: Oakland over Texas and Belmont over Wisconsin.

"Oakland has the talent to follow the Kansas blueprint," he says, "and Texas is likely upset that they were dropped all the way to a No. 4 seed. And in the Southeast, I can honestly see an argument for any team winning each of the games from the 3/14 matchup down, so I've blown up the bracket entirely, which includes Wisconsin losing to Belmont."

For my part, I've seen Texas play in person five times this season and watched the rest of their games since early January on TV, and I completely get what Mass is saying, because per usual, Rick Barnes' crew is stumbling at the wrong time. However, the Longhorns are one team who can neutralize Oakland center Keith Benson -- who's legit -- by running wave after wave of big men at him. I'm not predicting a deep run for Texas, but I think they avoid the upset. However, I agree with Mass on Belmont over Wisconsin. What a terrible matchup for the Badgers. The Bruins push the ball, create turnovers and their athleticism belies their small-conference pedigree. It's folly to pick against a team simply because they've let you down in the past (as Wisconsin has done, repeatedly). Rather, I'm picking against them because I think they'll lose.

Gramling and I agree on several "lesser" upsets. For instance, we both have three 11-seeds winning: Marquette against Xavier, Missouri versus Cincinnati and Gonzaga over St. John's.

"That Zags/Red Storm game was one of the hardest for me to pick," Gramling says. "I actually put it in the same category with Washington versus Georgia. All four teams are loaded with great athleticism and speed, creating fun and difficult matchups. I sided with the schools playing better of late, coming off conference championships: Gonzaga (with nine straight wins) and Washington, who I think avoids the upset."

Gramling actually does me one better in his love for No. 11s, though: he also picks the winner of the USC/VCU Wednesday first-round game to beat Georgetown. However, Gramling makes it clear he actually is taking favored Wisconsin over Belmont: "The only upset of the top 16 seeds I considered was Belmont, but I think of the Badgers as the team that beat Ohio State and Purdue, and not the school that scored 33 points in its last game. Wisconsin turns the ball over only 7.4 times per game and Belmont has zero wins versus 2011 NCAA tournament participants."

[+] EnlargeChris Singleton
Paul Abell/US PresswireThe return of Chris Singleton from a broken foot is a must for Florida State to defeat Texas A&M.

Touche, my friend.

As for me, I'm on an island taking Florida State to beat Texas A&M, a 10-seed over a 7-seed. I think people have the wrong idea about the Aggies: sure, they don't give up a lot of points, but that's mostly because they play so slowly. According to Ken Pomeroy's pace-adjusted defensive efficiency ranks, A&M is a good-not-great 39th in that category (while FSU is second). I feel fairly certain the Seminoles are going to add injured Chris Singleton, their best offensive player, back into the mix for this game, which should prick up their scoring a bit, and I trust that big guys like Xavier Gibson, Terrance Shannon and Jon Kreft can hold down that impressive Aggies front line.

Before we advance to the third round, here were the second-round favorites that all three of us endorsed: All of the No. 1 and No. 2 seeds, Syracuse, Purdue, BYU, Kentucky, Louisville, West Virginia, Arizona, Washington and George Mason.

Third round

I mentioned earlier that protected seeds tend to advance deep into the tournament, and it's not hard to figure out why: they tend to be the best teams (naturally), and also get paired with lesser squads. However, that most certainly doesn't mean that protected seeds always, or even usually, get to the Sweet 16. Last year, a full eight of the final 16 teams were 5-seeds or higher, the second time in the past decade that half the Sweet 16 was represented by "upset" squads. In the past 10 tournaments, on average, exactly six of the 16 protected seeds failed to make it to the Sweet 16.

Does that mean you should select six protected teams to get iced at this point (or earlier) in your bracket? Not so fast. If you're vainly trying to pick a perfect bracket, well, yes, you'd probably want somewhere around six of the top 16 seeds out by now. But considering how difficult it is to pick the right upsets, and how deep a No. 1 or No. 2 seed might go if they don't lose in this round, I actually think it makes more sense to start dialing back the upsets a little bit at this point. You'll probably miss a few pretty cool ones. But you're also less likely to find yourself eliminated from your bracket by the first Sunday's end.

Among our experts, Mass went the wildest at this point, selecting nine non-protected seeds to advance to the Sweet 16. Let's be clear: By my count that hasn't happened since the 1990 tournament, a span of 21 years. But that may be a paltry distinction, since eight non-protected seeds have advanced that far several times. Regardless, it's clear Mass means it when he says he expects some of the brackets to blow up; he'll either be among the Tournament Challenge leaders by early next week, or he'll be in the doldrums. Gramling and I took more conservative routes, as befits a safer strategy: Gramling has four non-protected seeds in the Sweet 16 (Washington, Tennessee, Missouri and Michigan State) and I have three (Washington, Utah State and Gonzaga). The three of us failed to agree on a single non-protected advancing team.

Certainly the two boldest calls of this round have two No. 1 seeds losing. Gramling thinks Tennessee will beat Duke. "I picked the Vols to beat Duke as my signature upset," he says, "because Tennessee has played a much more difficult schedule this season, beating Pittsburgh and Villanova on a neutral court and sweeping both Vanderbilt and Belmont this year. The Vols are also physical enough to win the battle on the boards, and Bruce Pearl has a history of overachieving in NCAA tournament play." I remember thinking the same thing about the Blue Devils last year in a regional final matchup against Baylor, and it almost came true, so this isn't a hard call to like. Mass AJ picks Old Dominion to shock Pittsburgh: "The Panthers always seems to disappoint, and there's no question that ODU can match Pitt's physicality."

[+] EnlargeIsaiah Thomas
Chris Williams/Icon SMIWill Isaiah Thomas and Washington carry their momentum from a thrilling Pac-10 tournament championship into the Big Dance?

Among the 2-seeds, Gramling and I established a mini-consensus about Washington stunning North Carolina. Says Gramling, "I think the ACC is as bad as it has ever been, and I'm just not sold on either Duke's or UNC's best wins, besides beating each other. I have Washington topping Carolina, especially with pesky on-the-ball defender Venoy Overton reinstated to the team."

This is my upset special of the third round, too; a look at Pomeroy's adjusted tempo and overall ratings indicates just how similar these teams play: They push it upcourt, and while Carolina probably has a better defensive team than the Huskies, Washington's offense has been hard to stop lately, and comes replete with junior and senior contributors that the Tar Heels just can't match. It should be immensely entertaining, but I'll take Isaiah Thomas and Washington.

Mass' upsets, of course, come fast and furious (though he does take Carolina in this round). He's got Georgetown beating Purdue ("With Chris Wright healthy, I think the Hoyas will be able to beat all comers"), Xavier beating Syracuse ("I hate picking the Musketeers over my alma mater, but I can't shake the feeling the 3s will start raining early and often, and the Orange have a tendency to start games slowly"), St. John's over BYU ("The Red Storm have a senior-heavy team and BYU just isn't the same since 'The Expulsion'"), Bucknell over Cincinnati ("Cincy hasn't impressed me all season, so I'll roll Bucknell through"), Oakland over Arizona and more. Give Mass this: Nobody's gonna call him chalky. And Gramling matches with at least one gutsy pick: Michigan State over Florida: "Tom Izzo has enough experience on his team of bruisers to make a run in what is one of the weakest regions in recent memory. Look for Michigan State to shut down the suspect-shooting Florida backcourt of Erving Walker (41.5 percent) and Kenny Boynton (38.2 percent) and move on."

And yes, I'm chalky as heck.

Where do the three of us agree for the 16 winners in this round? We all have Ohio State, Kansas, Notre Dame and Louisville advancing. That's it. Last year when we did this exercise, by this round we had three experts with eight teams in common. This year, it's only four.

Sweet 16

The first weekend is fun, but here's where you make your hay. The points accrue faster, and if you nail upsets to get to the Elite Eight, it's likely you'll leapfrog over many of your competitors. Of course, if you pick an upset and a favorite pulls through, you're severely handcuffed.

That said, this is the last round where it's really safe to be crazy. In the past 10 seasons, an average of 3.1 teams made the Elite Eight who were not a 1-seed or a 2-seed. Heck, just last year, we saw four such squads get to regional finals. But realize only once in that span have gotten more than four 3-seeds or higher this far, and again, the prudent, conservative (and yes, Mr. Chalk) route advises against going overboard in your upsets of 1s and 2s.

But Gramling is having none of this. He's got a full five of his Elite Eight coming by way of upset: he has a 10-seed (Michigan State), a 7-seed (Washington), two 4-seeds (Texas and Wisconsin) and a 3-seed (Purdue) in his regional finals. Mass has four surprise participants headed to his Elite Eight: Louisville, West Virginia, Old Dominion and Temple. And I have but three: Louisville, Purdue and Syracuse. Plus I was the only expert not to put a team seeded lower than fourth in my Elite Eight. That's partly strategy (keeping favorites alive tends to be good business in the long run, though if rare early carnage comes, I'm toast) and partly an assessment that college basketball's much-ballyhooed "parity" is overstated, and the best teams really are mostly the best teams.

[+] EnlargeTyshawn Taylor
Mark D Smith/US PresswireKansas' tournament fortunes could rest in the hands of Tyshawn Taylor, who has been rather inconsistent.

Gramling and Mass each lost their first No. 1 seeds in the last round; both AJ and I are ditching Kansas at the hands of Louisville here. "Big East teams are simply better battle-tested than other conferences," Mass says, "and Louisville is far better than advertised." I'm not sure I agree with the former statement, but I definitely agree with the latter. The Cardinals play the kind of fast-paced smothering defense that occasionally has gotten the Jayhawks out of sorts, and in particular they defend the 3 extremely well. The untold story of Kansas' season might be how dependent they are on Tyshawn Taylor from the outside. Yes, of course, their low-post game is nonpareil, but if Taylor is off, teams hang. I think Louisville hangs here.

Also, Mass chooses this moment to lose the tournament's overall favorite, Ohio State, in an upset at the hands of West Virginia. "For me, it's a matter of the Mountaineers having experience going deep into the Dance," he says.

The real shockers in the experts' brackets at this point are Old Dominion (Mass), Temple (Mass), Washington (Gramling) and Michigan State (Gramling) staying alive. And, um, I have no such bold picks. Mr. Chalk, il est moi. But let's let these guys defend their surprises.

"I love Jimmer Fredette as much as anybody," Gramling says, "but the Spartans have enough in-your-face defenders to keep Fredette in check and ultimately pull off the upset over BYU. And Isaiah Thomas has the quickness to break down the Syracuse 2-3 zone and Matthew Bryan-Amaning will give Rick Jackson fits on both ends of the court. I'm also not sold on the Orange's inconsistent offensive backcourt." Meanwhile, Mass says, "Really, Pittsburgh should be the team to face Florida, but ODU is just the mid-major version of Pitt, so I'll take them over Utah State." And while he didn't explicitly say so, one assumes Mass took Temple simply because Bucknell's magic carpet ride has to end sometime.

The harsh facts are that there are no consensus teams left among our three experts' Sweet 16 winners. I suppose that means there are vastly differing opinions about that "parity" I mentioned earlier. Personally, I expect more chalk at this point than the others appear to. But then, you knew that.

Elite Eight

How many 1-seeds should you have winning in this round? In the past 10 years, only 1.7 top seeds have advanced to the Final Four per tournament; last year it was only one, and the year before that, it was only two. Of course, back in 2008, we saw all four 1-seeds advance to the final dance, so the ultimate chalk isn't impossible. But really, this is the place where you make your stand. The points now dwarf the early rounds, so while it doesn't pay to advance anyone lower than a 5-seed, find one or two (or three!) non-1-seeds and cross your fingers.

Here again, though, I feel a bit chagrined to report that I'm the chalkiest of the bunch. I have just one "upset" in my Elite Eight: I have 3-seed Purdue headed to Houston along with three top seeds. Gramling was more adventurous: he picked San Diego State and, the most daring pick of the bracket so far, Michigan State to get to the Final Four. And Mass -- whose overall bracket we've pretty well established as the boldest of the bunch -- has Florida, Notre Dame and North Carolina, all 2-seeds, but still not pre-tournament favorites.

"Unless the Fighting Irish go completely cold from the floor, they're the strongest team in their region," says Mass, "and really the strongest on their entire side of the bracket." I take issue with Notre Dame's defense, and don't think it will hold up this long, but there's no arguing that its offense is something to behold. "I love San Diego State's senior-laden roster," says Gramling, "as well as the fact that they've allowed only five opponents to score more than 70 points this year. And I'll take Michigan State in a rubber match over Wisconsin in the Southeast final, but I admit I'd most likely be bouncing both schools in the Sweet 16 if they were in any other region."

Final Four

Let's allow each picker explain his own selections:

Mass: "I have Duke as my only 1-seed to advance, along with three 2-seeds. Because Florida's path to the Final Four is the easiest of the quartet, they'll be less prepared for the severe test they'll face in Notre Dame. Duke against UNC in the semis? Sure, the Tar Heels could win, but the Blue Devils just seem to have that extra gear. And after winning that one, any opponent will seem so much easier, and allow them to repeat as national champions."

Gramling: "After watching five months of college basketball, I think the two most impressive teams in the land, in terms of talent at every position, are Ohio State and Kansas, and I have them beating San Diego State and Michigan State, respectively. What a heavyweight battle this would be with two 37-2 schools squaring off! With both teams sporting the best big men in the nation, this game will be decided by whoever makes more shots and controls the glass. Kansas leads the nation in field goal percentage (.514), mainly because the Jayhawks are second in assists (17.9 per game). Plus they're great on the offensive glass, especially the Morris brothers. Ohio State is third in the country in shooting at 49.4 percent, which is a full two percent lower than KU. The biggest difference between the schools is rebounding. The Buckeyes rank 30th in the nation with a differential of plus-4.8 per game, while the Jayhawks place 8th in the land at plus-7.8. Kansas is my champion."

As for me, I've put upstart Purdue in Houston, along with three 1-seeds: Duke, Ohio State and Pittsburgh. I understand the negatives about Pitt: Big East officials let them get away with more physical play than the NCAA tournament refs might, and certainly Jamie Dixon's team has disappointed year after year. But I like this squad (and don't like the Southeast region). They're bruising up front, they've got terrific shooters and playmakers at guard, and their most important players are almost all juniors and seniors. "They burned me last year" isn't a good enough reason not to ride with them (though the worry over officiating is). I think the Panthers get to the final game against Ohio State, but will then bow out gracefully. The Buckeyes are the nation's most complete team, and the fact that they play only seven guys hasn't hurt them yet. And what a shock: Mr. Chalk takes the No. 1 overall seed for his national champion.

Christopher Harris is a fantasy analyst for ESPN.com. He is a six-time Fantasy Sports Writing Association award winner. You can ask him questions at www.facebook.com/writerboy and follow him at www.twitter.com/writerboyESPN.

ALSO SEE