Its all over in San Antonio. Memphis had the game in hand, Kansas was reduced to fouling, and Memphis failed to hit their free throws. Kansas chipped away at the lead, and with 2 seconds left hit a huge 3 pointer to send it to overtime where they simply dominated Memphis, who seemed like they just ran out of steam. So there you go, another basketball season over. The good news is, the 2008 Utes are undefeated! :)

Well, the final four is set. The time for bracket watching is over, and we can just sit back and enjoy the games. I thought I would take one last look at my poor pathetic bracket. I ended up 8 for 16 in the second round, 5 for 8 in the third round, and picked 3 of four of the final four participants. That puts me down for a total point score of 81 points in our yahoo group, good for 10th right now, but if if all goes as planned (North Carolina and Memphis win and then Memphis wins it all), I possibly jump to 3rd. I can't win it, since there are people ahead of me that picked Memphis too, but 3rd isn't so bad considering. Jump for the final bracket.

For the first time since the field expanded to 64 in 1987, the total sum of the seeds in the Final Four is... four. Yep, Kansas just survived a brave effort by Davidson to join North Carolina, UCLA, and Memphis in the final four, the first time that all four top seeds have made it to the final two games. And believe it or not, much as everybody likes upsets, I was actually cheering for Kansas just now - I think its going to be way more entertaining to see four obviously superior teams beat each other up rather then see one North Carolina steamroll Davidson.

Okay, so my script sucks. I made it out of the first round with a horrific 21 out of 32.

It correctly predicted no upsets in the South. In the West, it missed on UConn, Drake, Baylor and BYU. Even worse, it picked BYU to make it through to the Sweet 16. In the Midwest (the most upset of all of the brackets), it picked the Villanova upset correctly, but whiffed on Vanderbilt, USC, Gonzaga and Kent State. Finally, in the East, it incorrectly predicted upsets for St. Joes and George Mason and missed on the Indiana/Arkansas coinflip. Updated bracket after the break. The best I can do in the 2nd round is 11 out of 16, and even worse, I already have two of my elite 8 teams out! Sigh - next year, back to the drawing board.

The first round of the NCAA tourney is in the books so lets look back at some analysis. There were 8 upsets in the first round. According to our new favorite statistics, 7 upsets is the mean for the first round, so this was a slightly more upsetting year then most. As vaguely predicted on these pages, a 13 seed beat a 4. What wasn't expected was that two of them would win (and that one of them would put a whooping on the 4 seed. I guess Vanderblit wasn't all that after all). This marks the 7th year since the field expanded that two 13 seeds won. And everybody's favorite statistic, the 12 seed over the 5 seed returned in force this year with two upsets. The 12 seed now boasts a better winning record in the first round then the 11 seed, which only had one upset this year. The remaining three upsets wern't so impressive. Two 9 seeds beat 8 seeds which mostly matches the slight edge that the 9 seed holds (but, for all intents and purposes, its still a coin flip), and there was a lone 10 over 7 seed win, which is not unexpected (last year was the first in 13 years that a 10 seed didn't win). The South bracket completely held serve, and the East bracket only had one upset (and don't call it an upset - it was 9th seed Arkansas over Indiana). Both the Midwest and West brackets are featuring 12 vs 13 seed matchups; and the Midwest got especially decimated with 4 of the 8 total upsets. People were already saying that Kansas and UCLA had easy paths, and they got easier, with both facing either a 12 seed or a 13th seed in the Sweet 16, where the 13 seed has never won, and a 12 seed has only won once (but the statistics don't tell us if that one victory was over a 1 seed or not).

So, predictions for the 2nd round? I think that Arkansas, Texas A&M, UNLV and Mississippi are done (8 and 9 seeds are a combined 12 and 80 in the second round). Hold out a little hope for a 7 seed, which has about a 40% chance of winning - I like Butler. Look for a 5 seed to fall - they have a worse then expected record in the second round. I think the rest of the top seeds will sail through. I have no guesses in the Siena/Villanova or Western Kentucky/San Diego games - The stats just don't show us how many 12 vs 13 seeds there have been, but I'm guessing there just isn't enough information to predict anything but a total coin flip.

As always, its time for our usual Sigep March Madness bracket competition. This has been a busy year, with the arrival of the little girl and work and stuff, so I haven't had a chance to watch or follow any college basketball, except for peering in occasionally on our woeful Utes (Where have you gone, Rick Majerus?). So it goes without saying that I'm not going to be very educated making my bracket picks this year. So I figured, why not do something silly? Last year I posted about A mostly random script. This year, I thought I would try my hand at something similar. I just posted about these wonderful stats from sportsline.com. It occurred to me that I can use the round-by-round statistics to construct a bracket. For each contest, both teams got a number of chances corresponding to their winning percentage in that round. So for example, in the 8 vs 9 contest, the 8 seed is 42 and 50 and the 9 seed is 50 and 42. In my script, the 8 seed gets 0 through 31 the 9 seed gets 42 through 91 and the random number generator picks a number in that range. Look after the jump for my bracket:

In honor of the field of 64, which just got announced, here's a fun stat sheet courtesy of CBS Sportsline, showing some key historical stats of the tournament. One thing that was interesting to me was the 5 seed vs 12 seed matchup that everybody always harps about. Marc was arguing the other day that really, its is not more likely to occur then an 11 beating a 6. Which happens to be exactly the truth - both the 11 seed and the 12 seed have the exact same winning percentage in the first round. However, you will notice that the winning percentage of the seed steadily rises as the seed rises, except between the 11 and 12 seeds, which have the same exact winning percentage. Furthermore, the delta between the 13 seed and the 12 seed is 11, whereas the delta between the 14 seed and 13 seed is 3, and the delta between the 11 seed and the 10 seed is 6 wins. By all accounts, the 12 seed should probably be about 4 or 5 wins worse then they are. So, what does this mean? Well, it means that the legends are trun, and that the 12 seed is the lowest seed that has been averaging more then 1 win per year for the last 23 years, so its the biggest upset that you can reliably choose for your bracket (of course, the trick is picking the right one). It also means that you shouldn't be as generous with your 11 seeds either - you should really only have one 11 seed advancing to the second round.


Its also interesting to look at the 8 / 9 matchup - the 9 seed is the favorite with what should basically be a 50/50 split. However, don't expect the 9 seed to go much further, they are a terrible 3 and 47 in the 2nd round (which happen to always be against the 1 seed, but still). 8 is a much better bet for a 1 in 3 chance of knocking off the big dog, but still not something I would take without good cause.


Finally, I'm intrigued by the records by seed - in 23 years, we have never gone 2 years in a row without a 13 seed winning (and last year no 13 seed won). Even more interesting, the 15 seed should win about every 7 years or so, and its been 6 years without a 15 seed win.


I love looking at the historical statistics of the game, because its so cold and calculating, not at all like the tournament is when it gets going. These numbers seem so insignificant when you are cheering for a 13 to beat a 3, but I'll bet when we go back and plug in the numbers from 2008, then they'll all make sense again, and we'll consider ourselves fools for not seeing that upset when the statistics told us we should. Thats why math rocks.

Which was more disappointing? the Cavs or The Sopranos? Watch out for the spoiler on the second link.

I didn't see either, but judging from the reaction on the radio this morning, I'm guessing most men are leaning toward the latter.

The first round of the NCAA tourney is in the books. The geek script did pretty well - going 25 of 32, though it picked a few long shots that didn't end up paying off. In real life, I did 24 of 32 , so the geek script ended up beating me. How did it do on round two? Read on...

These are the teams that the script selected to win round 2, based on the round 1 events. Losers (or teams that never made it) are stricken, winners in bold.

It has begun. Too late, I discovered this fun little script that does brackets, geek style. Because it is random, I don't feel bad about making a bracket (even after the games have started), and posting it here, and scoring it according to Yahoo rules to see how accurate it is. I'll try to draw this all in a pretty PDF, but until then, jump to see the current
round: