Giant Killers: The List

POTENTIAL GIANT KILLERS

The percentage listed next to each team represents its chance of becoming a Giant Killer. Teams that have been eliminated are crossed out.

RIP, GK's

Wednesday, March 25, 4:44 p.m. ET

Posted by Jordan Brenner and Peter Keating

Well, it's almost 24 hours until tipoff of the Sweet 16, and you know where things stand: All four No. 1 seeds are in for the fifth straight year ... Gonzaga, Memphis and Xavier, the remaining non-BCS teams are Giants ... Arizona, the only team not seeded in the top five of its bracket, isn't exactly an unknown quantity.

The Giant Killers are ... no more.

There were three Giant Killings this year; we thought two might happen (Cleveland State over Wake Forest and Western Kentucky over Illinois) and missed big on the third (Dayton over West Virginia). But from the beginning, our model was chalk-heavy. It essentially favored every Giant in every matchup, often overwhelmingly. There were no natural upset-pullers facing vulnerable high seeds, (such as Davidson vs. Georgetown last season). That's partly because there's a growing gap between the very best teams in the country and everyone else, which the NCAA has responded to by drastically reducing the number of at-large bids it rewards to non-power conference teams. And it's partly because of unlucky seedings (super-vulnerable Florida State, for instance, drew Wisconsin instead of, say, VCU). Some might take this whole trend as a cue to fold up the spreadsheets and go home. We're braver men than that: Go home we will, but we shall return. With more data than ever.

Sunday GK Specials

Saturday, March 21, 7:57 p.m. ET

Posted by Peter Keating

Time to preview Sunday's games, and only two of the eight matchups have Giant Killer implications. Here's how they stack up:

No. 3 Kansas (40.4 percent) vs. No. 11 Dayton (24.7 percent)

The numbers say Kansas should actually be an easier takedown for Dayton than West Virginia, whom the Flyers upset on Friday, humiliating our spreadsheets along the way. The Jayhawks held off North Dakota State for two main reasons -- they turned the ball over just five times, and Cole Aldrich got inside with impunity. Neither feat will be easy to repeat against Dayton. On the season, Kansas has committed considerably more turnovers than it has generated (21.1 percent of possessions vs. 19.0 percent). And the Flyers' grinding, physical defense, which limited West Virginia to 49 shots, slows the game down while limiting opponents to effective field goal shooting of just 45.6 percent. Dayton's Chris Wright torched West Virginia for 27 points (and made a strong bid for the All-Giant Killer team), but that was more than double his season average. And if Dayton's defense isn't solely a product of tempo, neither is its offense, whose adjusted efficiency (103.9) ranks just 127th in the country. The Flyers should at least keep this close, and this is the likeliest chance for a Sunday Giant Killing.

No. 1 Louisville (26.8 percent) vs. No. 9 Siena (28.1 percent)

The Saints make their first appearance in this tourney as a Giant Killer after edging Ohio State (in a non-GK game) in double overtime Friday night. It probably will be their only shot at a Giant, too. Siena does a nice job of protecting the ball and generating turnovers (18.2 percent of possessions vs. 22.7 percent, respectively). But the Saints beat the Buckeyes by compiling a staggering 23-9 edge in offensive rebounds, which allowed them to take 12 more shots than Ohio State. Little suggests they'll be able to repeat that feat against Louisville. The Cardinals rank sixth in the country in steals, leading to turnovers on 23.4 percent of opponent's possessions. And things won't get any easier if Siena hangs onto the ball: Louisville lets opponents shoot just 43.1 percent on 2-pointers (partly because they rank seventh in the NCAA in blocks) and only 30.6 percent on threes. This is where we wish the MAAC had sent Niagara, a team with true Giant Killer potential.

POTENTIAL SLAIN GIANTS

The percentage listed next to each team represents its chance of becoming a Slain Giant. History: Over the past five tournaments, eight of 11 teams that were given at least a 50 percent likelihood of losing to a GK did, in fact, go down. Teams that have been eliminated are crossed out.

*Note: The model generates negative results for some teams, which have been assigned odds of 0 percent.

Cleveland State!

Saturday, March 21, 1:00 a.m. ET

Posted by Jordan Brenner

Just when we'd reached rock bottom, Cleveland State lifted our spirits. We've been believers in the Vikings since the Horizon League tournament, when they beat Butler in what was, essentially, a Giant Killer tuneup. And despite Wake Forest's relatively safe profile, we gave the Vikes a pretty strong endorsement anyway. Cedric Jackson and Co. delievered, and now we don't feel quite so, well, stupid. The only problem? Cleveland State now faces Arizona, and since the Wildcats are a 12-seed, that's not a Giant Killer matchup. But since our model was considering frying its own circuits out of frustration, this was a good development for Team GK.

Licking Our Wounds

Friday, March 20, 7:33 p.m. ET

Posted by Jordan Brenner and Peter Keating

The day started so beautifully at GK Central. We were congratulating our model for wonderful advice in Thursday's games. One of our best bets, North Dakota State over Kansas, was looking good as Ben Woodside drained 3s and got to the tin with ease. Utah State was looking even stronger. Both, eventually, went down, but we were ready to deal with that. After all, we've been endorsing sticking with favorites all week.

Then West Virginia happened.

We could try to tell you that you didn't need our Giant Killer analysis to consider West Virginia a lock over Dayton. Vegas made the Mountaineers nine-point favorites; Ken Pomeroy put their odds of beating the Flyers at 87 percent. Or we could carry on about how the beauty of March Madness is in the unpredictability of its upsets. A statistical model capable of calling Villanova over Georgetown is a spreadsheet that would drain the world of all joy.
But you don't want to hear that. Because we're the morons who claimed our model showed there was no way West Virginia could lose.

As commenter alexvr18 put it (so eloquently): "U idiots I thought west virginia had a 0% chance!!!!" [sic]

Zero percent? Nothing is zero percent. Landing on the moon wasn't zero percent. Mickey Rourke's comeback wasn't zero percent. And David over Goliath certainly wasn't zero percent. But we just had to go there, because our model said West Virginia was precisely the kind of team that wouldn't lose to a Killer, even a great-looking Killer. And Dayton clocked in at 24.7 percent -- solid, but far from spectacular. (And if you're not ready to throw something at us yet, try to digest the fact that the model now thinks the Flyers are a darn good bet against Kansas next round.)

To make matters worse, we also told you to ride Temple against Arizona State. You know how that turned out. So all we can do is paraphrase the legendarily abject Artie Fufkin and say that we did it, it's our fault, and we invite you to kick our butts should we ever meet in public.

Until that time, remember that stuff happens, especially in March. Objectively speaking, there's no way that West Virginia should have lost this early. Just like there was no way Duke should have lost to VCU in 2007, based on our statistical model. Against Dayton, WVU forced only nine turnovers, which is a big reason why the Mountaineers took only 49 shots ... ah, forget it. We want to crunch more numbers right now about as badly as you want to read more of them.

Instead, it's time for a walk around the block, a beer and some hot Xavier-Portland State action. The way today's going, you know how that prediction is bound to turn out ...

Saturday Preview

Friday, March 20, 12:57 p.m. ET

Posted by Peter Keating

Gonzaga (13.1 percent) vs. Western Kentucky (25.4 percent)
We had planned to post a blown-out preview of Saturday's Giant Killer games. Turns out, this is the only one. And while the Hilltoppers pulled Thursday's only major upset (making our model appear smart in the process) the numbers say this is where the fun stops. Western Kentucky was able to build a 17-point lead over Illinois, but Gonzaga is much less prone to digging itself into holes: On offense, the Bulldogs shoot almost 40 percent on 3s and have an adjusted offensive efficiency of 119.1 -- a whopping 13.4 more points per 100 possessions than Illinois. And on D, Gonzaga does a better job than any team in the country at stopping two-point buckets (38.5 shooting percentage).

Western Kentucky nearly threw away the Illinois game with 16 turnovers, and that kind of sloppiness isn't uncommon for the Hilltoppers, who commit slightly more turnovers than they generate (20 per 100 possessions versus 19.8). That won't cut the mustard against Gonzaga, which hardly ever gives the ball up (its 16.2 turnover percentage ranks fifth in the country) and is above-average at forcing miscues (21.5 per 100 possessions). Enjoy Orlando Mendez-Valdez's outlet passes and long-range shooting for one more game -- and remember him when it comes time to vote for the All-Giant Killer team.

Day 1 Recap: The Model Looks Good

Friday, March 20, 12:30 a.m. ET

Posted by Jordan Brenner

Western Kentucky tried its best to blow it against Illinois, and VCU tried its best to steal the game from UCLA. But in the end, Thursday's games produced exactly one Giant Killer: the Hilltoppers. What do we make of that?

For starters, our model started strong. If you've been following us for the past week, you know that we're not surprised by the chalky action we witnessed Thursday. We've explained -- often -- that not only did this year's field produce a poor crop of Killers but the matchups didn't help, either. The good news is that all three of our "Best Bet" games take place Friday. So there's still room for a little GK excitement. And of the next wave of games, Western Kentucky over Illinois was -- by far -- our model's favorite pick. It came through.

We also spent a lot of time talking about UCLA's zero percent vulnerability rating mattering more than VCU's 40.2 percent GK mark. We advised you to avoid that trendy pick. And, yeah, we were sweating out Eric Maynor's final shot, but our model turned out to be on target. And that means WKU vs. Gonzaga will be Saturday's only GK game, which is shockingly little mid-major action for a day featuring eight second-round matchups. (And, because people keep struggling to understand this in the comments sections of various GK stories, teams from BCS conferences can't be Giant Killers). Let's see what happens in Friday's games with GK potential. In case you've forgotten, they are:

East Region

(1) Pittsburgh (16.7 percent) vs. (16) ETSU (17.8 percent)

(4) Xavier (62.7 percent) vs. (13) Portland State (17.6 percent)

South Region

(3) Syracuse (34.9 percent) vs. (14) Stephen F. Austin (21.6 percent)

(6) Arizona State (27.9 percent) vs. (11) Temple (32.6 percent)

Midwest Region

(1) Louisville (26.7 percent) vs. (16) Morehead St. (2.6 percent)

(2) Michigan State (41.7 percent) vs. (15) Robert Morris (8.8 percent)

(3) Kansas (40.4 percent) vs. (14) North Dakota State (16.1 percent)

(4) Wake Forest (17.8 percent) vs. (13) Cleveland State (34.9 percent)

(6) West Virginia (0.0 percent) vs. (11) Dayton (24.7 percent)

West Region

(3) Missouri (0.0 percent) vs. (14) Cornell (12.3 percent)

(6) Marquette (10.5 percent) vs. (11) Utah State (20.0 percent)

Late-Night Musings

Thursday, March 19, 11:59 p.m. ET

Posted by Jordan Brenner

The first day of games is winding down, and if UCLA and Western Kentucky hold on, we'll have witnessed exactly one Giant Killer upset. Eric Maynor is trying his best to make that two, though. More to come shortly.

Last-Minute Reminders

Wednesday, March 18, 6:43 p.m. ET

Posted by Jordan Brenner

Peter did a great job today of explaining the problems with this year's field -- ostensibly we have a weak crop of Giant Killers, and in the first round our best bets didn't draw any of the most vulnerable Giants. So, it's certainly fair to avoid most mid-majors in searching for first-round upsets.

But it's worth remembering that the battle doesn't end after Thursday and Friday, and that if you're filling out a bracket, you need to think ahead. So, let's say you trust your gut over our model and feel like picking VCU over UCLA. Fair enough. Just don't stop there. If Villanova advances, you'd suddenly end up with VCU with a 40.2 percent upset rating (that number will change after the first round when we re-run the stats) against Nova, which checks in with a 39.1 percent vulnerability mark. Now, that's the kind of upset we can promote. The point is, if you're going to go against the grain in the first round, you'd might as well look for opportunities to double down after that.

We'll check in again tomorrow, of course, and see how our model performed. And we'll be with you all Tourney long. Who knows, we might still be talking about a GK game in a regional final, just like with Davidson-Kansas last year.

Commenting on Comments

Tuesday, March 17, 4:00 p.m. ET

Posted by Jordan Brenner

We in Giant Killer Central are nothing if not men of the people. As such, not only are we reading your comments -- we're responding. Many have been good. Some have been confusing. A couple ... uh ... wow. Anyway, let's get on with it.

erlewein2013: I guess I don't get the "Slain Giant" concept. Duke lost last year to a team 5 seeds below it in West Virginia, after beating 15 seed Belmont by 1, and in 2007 lost to VCU, an 11 seed (and 5 spots lower). That obviously is within the 5 year window, and, yet, somehow Duke has a 0.0% chance of being a slain giant? How? I am sure I am missing something.

GK Response: You're confusing historical Giants with an individual team's history. Duke's 2009 team is being judged against the 25 Giants that have lost in the Tourney over the last five years, and purely on a statistical basis. Duke's results last year have nothing to do with the analysis (plus, West Virginia isn't a Giant Killer, as a Big East team). This year's Duke team, which is all that is relevant, ends up with a 0.0 percent likelihood of losing to a GK based on how its statistical profile matches up with Giants who actually lost over the last five years.

cyork987: Can we get the percentages for all the teams in the field? Does a teams Giant Killer % work in the inverse, say if VCU were to play ETSU, does it make them less likely to be beaten by a team 5 spots away?

GK Response: The percentages for Giants and Killers are listed in the two charts on the right side of this page. Some teams won't count as a Giant or a Killer. For instance, Wisconsin can't be a Killer, because it's from the Big 10. But it can't be a Giant, because it will never play someone seeded at least five spots lower. And if two GKs were to meet (as in your VCU/ETSU example), it wouldn't count as a GK game. After all, neither would be slaying a Giant.

alexvr18: y cant a team fron a major 6 conference be a giant killer?
in the games like usc/mich st (2nd round), wash/miss st (1st round), and like u even mentioned, wisconsin/florida st (1st round) it would be good to know the percentages

GK Response: Because those are the rules! Giant Killers don't come from major conferences. The world wouldn't stop if Wisconsin were to beat FSU. If Akron wins, it's a different story. Because of our rules, we have no historical basis for power-conference teams beating other power-conference teams, so we can't give you percentages for the likes of Wisconsin, USC, etc. Sorry -- that's not how this model works. But if you really want the odds, I recommend my good friend, Bracket Predictor.

MCBuser "History: Over the past five tournaments, eight of 11 teams that were given at least a 50 percent likelihood of losing to a GK did, in fact, go down." Does this mean in the first round or could this also apply to a 2nd round upset?

GK Response: Good question! Indeed, it does apply to the second round. In fact, Wisconsin didn't meet its GK fate until the Sweet 16 last year against Davidson.

AdrianKhan: can someone explain the unrivaled hatred from espn towards the spartans?

GK Response: Dude, you're getting mad at a spreadsheet. This is all math, sorry. Besides, Bracket Optimizer loves the Spartans. (P.S. -- that's math too.)

rkonks29: Basically all you did was contradict both sides of the matchups you picked! Did you just watch Dicky V on selection Sunday to get this or what? You made no predictions and admittedly put up some predictions that are the "trendy picks". Great job ESPN. An 8 year old could have come up with this stuff. TERRIBLE!!!

GK Response: Two words: Reading. Comprehension. Peter was previewing every Giant Killer game from the first round and ranked them in order of likely upsets. Not sure what was so difficult to understand about that.

grundel1023: Um, Maryland will upset Cal. That's your best bet.

GK Response: That's lovely and all, but a) It's a 10/7 game and b) they're both power teams. So, it has nothing to do with Giant Killers. Other than that, sounds good!

tim.mccahill: Way to jump on the bangwagon with the rest of the country. I can listen to basketball commentary and reiterrate what they say as well. KU had a tough loss to Baylor but I would rather have them stumble now than the first round of the tourny. They will beat Michigan State to play in the Elite 8 vs. Louisville.

GK Response: Again, you're arguing with a spreadsheet. There's no subjective analysis involved in this. Kansas just happens to have a lot in common with teams that have lost to Giant Killers over the last five years.

Thanks for the all the feedback, suggestions and confusion. Keep the comments coming!

Killed by the committee?

Monday, March 16, 5:21 p.m. ET

Posted by Peter Keating

At first glance, Sunday's seeding selections didn't look too promising for Giant-Killing -- and after second and third glances, they still don't. For one thing, small conferences, the breeding grounds for potential Killers, got shafted. Non-power conferences, as you probably have heard, got only four at-large bids; in 2004, they got 12. Several of our most promising Killers (Davidson, San Diego State, Creighton, St. Mary's) found their bubbles burst on Sunday. Bummer.

Some of our other top prospective Killers drew lousy matchups. Akron ended up with a safe Giant in Gonzaga. VCU, as we've mentioned, has super-safe UCLA. And several vulnerable Giants either drew weak Killers (Michigan State vs. Robert Morris) or were rendered ineligible because they got matched against a medicore BCS team (we're looking at you guys, Florida State and Wisconsin).

But there still will be upsets, as we're previewing. There always are. At least three Giants have gone down every year since 2004. And in 2006, which lacked a single Killer with a rating of higher than 38 percent (sounds a lot like this season), six GKs emerged, including a certain George Mason team. The trick was the fact that a number of Giants were ripe for the picking. So maybe that's the template for this year -- pick on the weak Giants, even if a GK's profile isn't that strong.

Clearing up confusion

Monday, March 16, 9:55 a.m. ET

Posted by Jordan Brenner

In the first-round matchup post below, the percentages listed next to each team aren't game-specific. They are indicators -- based on historical stats -- of a team's likelihood of fulfilling its destiny to be a Giant Killer or a Slain Giant. So, to answer tothman22's question (yes, we're reading the comments), UCLA's profile indicates that it has a 0% chance of being upset by any Giant Killer, irrespective of its opponent. Similarly, VCU's likelihood of being a Giant Killer is 40.2 percent, regardless of which team the Rams play. That's why we started computing these numbers prior to the Tourney -- we're purely comparing a team's profile to those who pulled of the feat in the past.

So, which team should you trust more? Well, no Giant with a 0 percent rating has lost to a Killer in our five-year sample. In fact, only two Giants below 23 percent have fallen during that time. Obviously, VCU would be a stronger pick against someone like Florida State, but we'll help you reconcile the difference between weak Giants and strong Killers (and vice versa) shortly. For now, if you're still confused, read our methodology. And if you're struggling after that, post another comment. We'll get back to you!

First Round: Killer Matchups

Sunday, March 15, 8:20 p.m. ET

Posted by Jordan Brenner

In case you forgot our rules, here's a refresher: A Giant Killer is a team that beats another squad seeded at least five spots better. A Giant Killer can't come from any of the six BCS conferences. Gonzaga, Memphis, Butler and Xavier can't be Killers either, and we've added Utah to that group, since the Utes earned a 5-seed. With that in mind, here are the first-round Giant Killer matchups. Each Giant and Giant Killer is listed with the percentage our model computed as a likelihood of falling or pulling off an upset, respectively. Stay tuned Monday for a detailed breakdown of the East and South regions, followed by a closer look at the Midwest and West on Tuesday.

East Region

(1) Pittsburgh (16.7 percent) vs. (16) ETSU (17.8 percent)

(2) Duke (0.0 percent) vs. (15) Binghamton (21.2 percent)

(3) Villanova (39.1 percent) vs. (14) American (12.9 percent)

(4) Xavier (62.7 percent) vs. (13) Portland State (17.6 percent)

(6) UCLA (0.0 percent) vs. (11) VCU (40.2 percent)

South Region

(1) North Carolina (6.6 percent) vs. (16) Radford (0.0 percent)

(2) Oklahoma (44.5 percent) vs. (15) Morgan State (30.9 percent)

(3) Syracuse (34.9 percent) vs. (14) Stephen F. Austin (21.6 percent)

(4) Gonzaga (13.1 percent) vs. (13) Akron (38.6 percent)

(5) Illinois (37.4 percent) vs. (12) Western Kentucky (25.4 percent)

(6) Arizona State (27.9 percent) vs. (11) Temple (32.6 percent)

Midwest Region

(1) Louisville (26.7 percent) vs. (16) Alabama St. (13.2 percent)/Morehead St. (2.6 percent)

(2) Michigan State (41.7 percent) vs. (15) Robert Morris (8.8 percent)

(3) Kansas (40.4 percent) vs. (14) North Dakota State (16.1 percent)

(4) Wake Forest (17.8 percent) vs. (13) Cleveland State (34.9 percent)

(6) West Virginia (0.0 percent) vs. (11) Dayton (24.7 percent)

West Region

(1) Connecticut (33.4 percent) vs. (16) Chattanooga (0.0 percent)

(2) Memphis (3.8 percent) vs. (15) Cal State Northridge (20.9 percent)

(3) Missouri (0.0 percent) vs. (14) Cornell (12.3 percent)

(5) Purdue (5.2 percent) vs. (12) Northern Iowa (6.8 percent)

(6) Marquette (10.5 percent) vs. (11) Utah State (20.0 percent)

Now it's for real

Sunday, March 15, 7:35 p.m. ET

Posted by Jordan Brenner

The brackets are up! We'll be back shortly with a list of all the first-round games featuring Giant Killer ramifications. Stay tuned ...

Zip-a-Dee-Doo-Dah

Sunday, March 15, 12:20 p.m. ET

Posted by Jordan Brenner

It's judgment day, and once again, we've trimmed our list of Killers, while still generously including several bubble teams without much of a hope of landing in the bracket tonight. One team we didn't have to remove, though, was Akron. The Zips featured a strong Giant Killer profile last season, but failed to make the tourney. That's not an issue this year, since Akron captured the MAC championship Saturday. Power teams should be concerned by that result: Akron currently features our fourth-highest GK rating (38.6 percent), and we're not exactly confident that Davidson will make the field. We'll have the answer to that and every other question soon enough, though, so check back during the selection show as we live-blog all the GK matchups as they appear.

New Stats, New Thoughts

Saturday, March 14, 1:42 p.m. ET

Posted by Jordan Brenner

We've finished crunching today's numbers, and some teams are on the move. No one is as interesting as San Diego State. The Aztecs have emerged as our second likeliest Giant Killer at 46.2 percent. And while they have a solid shot at an at-large bid, they can make that issue academic by winning the Mountain West title Saturday night at 7 ET against Utah. Among the Giants, Syracuse's run in the Big East tournament continues to manifest itself in a declining vulnerability. The Orange is now just 36 percent likely to lose to a Giant Killer -- the second straight day that number has dropped. So take some time to scan the stats, then sit back and watch the remaining Killers and Giants in action today.

CLEAN-UP DAY

Saturday, March 14, 12:17 p.m. ET

Posted by Jordan Brenner

The bubble keeps growing, and a bunch of teams that once had at least minor hopes of grabbing an at-large bid have bounced off. So, even though we've approached the process of eliminating teams conservatively, we've pruned our list. In addition to removing teams that had to win their conference tourneys to get in -- Long Beach State, Bowling Green, Coppin State, South Carolina State, Holy Cross, Nicholls State, Texas A&M-CC, Prairie View, Arkansas-Pine Bluff and New Mexico State, we had to make a couple of other tough calls. Not long ago, Rhode Island seemed to have a reasonable shot at a bid, but the upsets around the country have basically eliminated the Rams from contention. And, sadly, we had to say farewell to our previously top-rated Giant Killer, Niagara. There's just no way the MAAC will get two teams. Stay tuned for updated numbers.

PATRIOT GAME

Friday, March 13, 6:09 p.m. ET

Posted by Peter Keating

The Patriot League final going on right now -- it's American University 45, Holy Cross 29 as I write this -- has got to be the least exciting conference championship in the country from a Giant Killer perspective. Both of these teams have had nice seasons, and American beat Army by a single point last Sunday just to get here. But they both commit more turnovers than they generate. And neither packs much scoring punch, even after adjusting for the fact that watching the Eagles play is like watching paint dry -- their tempo ranks 329th in the country. Our model estimates American has just a 10.9% chance of pulling an upset if they make the Big Dance, and Holy Cross no chance at all. Go Eagles ... I guess.

FRIDAY: THE TEAMS, THEY ARE A-CHANGING

Friday, March 13, 12:44 p.m. ET

Posted by Jordan Brenner

After updating all the stats through Thursday's games (yes, even Syracuse-UConn), we started pruning our list a bit. A few choices were easy -- removing teams such as North Carolina A&T and Sam Houston State that could have made the Big Dance only by winning their conference tournaments. We've been pretty liberal with our bubble teams, just in case there's a Sunday surprise. Still, we decided enough was enough for Illinois State -- they just trail too many other teams even to think about an at-large berth. They're done, so they're off our list.

And while the same teams remain on the Giants list, the numbers have changed quite a bit. Apparently, 70 minutes of basketball can alter a team's statistical profile dramatically: Syracuse's likelihood of falling to a GK dropped from 42.6 percent to 37.2 percent after knocking off UConn. Oddly, the Huskies' numbers barely changed. More to follow, in advance of a Saturday filled with conference championship games.

THURSDAY UPDATE

Thursday, March 12, 2:58 p.m. ET

Posted by Jordan Brenner

We've updated the numbers through yesterday's games and are taking a moment to say farewell to Mount St. Mary's, the latest team to drop off the Potential GK list. Mountaineers, we hardly knew ye.

A FAST START

Wednesday, March 11, 5:40 p.m. ET

Posted by Peter Keating

Butler and Cleveland State took our Giant Killer system for a test drive Tuesday night when they faced off for the Horizon League championship. And baby, we're humming!

One quirk in our analysis is that we treat Butler, Gonzaga, Memphis and Xavier as Giants, not potential Killers, despite the conferences they come from, because of their past success in the NCAA tournament. That means even before the Big Dance, those teams have a chance to run into Killers in their conference tourneys.

So on one side there were the Butler Bulldogs, ranked 16th in the country, winners over Cleveland State twice this season already on their way to a 26-4 record, favored by six points, supported by most of the raucous crowd of 5,107 at the Hinkle Field House in Indianapolis. On the other side: the Cleveland State Vikings, who hadn't made the NCAA tournament in 23 years.

Butler's raw stats are impressive. For example, the Bulldogs give up just 57.4 points per game, one of the lowest totals in the country. But before this game, our analysis showed Butler as the second-likeliest of all Giants to topple, with a 70.1 percent chance of falling. Cleveland State, on the other hand, was the sixth-likeliest team to pull a big upset, with estimated odds of 33.4 percent.

We recalibrated the numbers too late to run them, but I e-mailed a small group of friends who are sports obsessives, just to predict the upset. At 8:37 p.m., I got an e-mail from an editor at a well-known magazine: "Call me in the next 15 minutes & I can get your money on Cleveland St."

I called, and when my friend picked up, I heard typing in the background. He asked, "Is it OK if the bet is in euros?" The game, however, had already tipped off, so you will never know how I would have settled the whole ethical conundrum.

Butler stopped Cleveland State from getting inside all night long, took a 39-31 lead and stayed ahead at halftime, which seemed significant because the Bulldogs were undefeated this season when leading at the half. But Cleveland State went on a couple of runs to come back, relying on precisely the strengths that make our model like the Vikings as a Giant Killer. They shot 53 percent on 3-pointers (10-19) and held Butler to an equally amazing 21 percent (4-19) on 3s, maximizing the value of their own possessions and limiting the value of their opponents'. They committed just seven turnovers. And PG Cedric Jackson, a highway robber during the regular season, had three steals.

So congrats to Cleveland State coach Gary Waters -- bet Rutgers wishes now that it had kept him. And to Jackson, who is in the running to make the All-Giant Killer team, if we -- or you -- can figure out how to pick it. (Use the comments section, we're paying attention.)

NUMBER CRUNCHING

Wednesday, March 11, 3:11 p.m. ET

Posted by Jordan Brenner

Welcome to our first look at the stats for Potential Giant Killers and Potential Giants. Over the next few days, as small-conference teams are knocked out of their tournaments, the group of Giant Killers will shrink dramatically.