Bracket Math: Day 1

I've got limited amounts of free time that aren't taken up by Starcraft, but here's a quick update.

 

First off, I'll point out that Shannon got 15 of the 16 games right on the first day.  I think I got 10.  But more importantly, what is the impact of the new scoring system?  Well, the system rewards upsets, and for a while, I was afraid there wasn't going to be a single upset in the first day of games.  VCU managed to hold off Wichita State for the first 12/5 upset, and then Colorado beat UNLV (11/6 upset) in the last game of the night.

 

Most Valuable Games:

VCU over Wichita St - 8pts (12/5, 1st round)

Colorado over UNLV - 6pts (11/6, 1st round)

All other 1st round games - 1 pt ea (14 games, 14 points)

 

Total points thus far: 28.

 

Early on, we can start to see the large effect that the bonus points for upsets has.  If you picked every better seed to win yesterday, you'd have gotten 14/16 games right for 14 points.  If you picked every lower seed to win, you'd also have 14 points.  Both would have been somewhat foolish predictions (100% upsets being much more foolish) and would have yielded equally.

 

My concern, however, is this: either foolish method would have been better than me.  I got 10 games right, but neither of the upsets.  Now, me losing in bracket challenges to foolish people is a long-standing tradition, so that isn't a problem in and of itself.  But consider instead someone who recognizes that there is about 1 5/12 upset each year in the tournament.  Rather than trying to predict which 1 or 2 12-seeds will pull it off this year (like I did), why not just pick them all?  If you get 1 right, you get 8 points.  2 gets you 16.  And if you are wrong on them all, you only miss out on 4 points.  As I've mentioned, I think more analysis is warranted, but for now, my gut says that the upset bonuses are too high.

 

A mitigating thought: I think that bracket "goodness" ought to mirror the excitement of the tournament itself.  There are two primary components that make the tournament interesting, even if your team isn't involved.  One is good basketball and crowning a champion and all that, but the other is the early upsets.  (You know, those Cinderellas that the TV announcers can't stop speculating about.)  Picking VCU to win probably wasn't 8 times more difficult than picking Wichita State.  (Kenpom gave VCU a 22.8% chance.)  But VCU winning might be 8 times more exciting than a world with no upsets.  Sports nerd-dom is good and all, but the excitement is what gets so many people (even Shannon) involved.

 

p.s. I think with my new scoring system, the maximum possible points in a tournament would be 863.  (Unless, of course, you can find a higher arrangement.)  I haven't checked all the possibilities, as that can't be done by brute force with any computer I have access to.  (Or any computer at all, at least not without using logic to make some very significant reductions in the number of calculations that need to be made. (Which is probably possible, but I haven't done it.  My method was one of some superficial logic and trusting my mathematical gut.))  Total points possible in last years bracket was 410, for comparison.