Monday, March 15, 2010

Loose Women, Hard Lemonade and Awkward Bank Shots

It's so close, you can smell it - that smell that will linger on you (and JC's basement) for a few days afterward, one that you don't even notice anymore because now, well, that's just how the world smells. And it smells good. Kind of.

We're three short days away from the opening weekend of the NCAA Tournament, where 64 teams will bludgeon each other over and over until one winner appears, ready to cut down the nets about 7 weeks after the tourney starts. Since it is an orgiastic day of basketball, gambling, and drinking, we've spent years (YEARS JERRY) perfecting how to best mix these three things together while still retaining jobs, families, and without driving anything anywhere at all. We've got it down to a science. A fat, hairy, awesome science. Here's from 07, and from 08, and from 09, although . . . I mean . . . I lived it, sister. I just don't remember any of it.

The gamboooling is clearly the most important part, however - and so I'll run down a little about my system for determining "fair" lines, which I'll use to coast to an even-money finish, losing again to Brent, who will brush his teeth before bed every night even though he passed out for hours in the afternoon.

The 'key' to the whole deal is what is known as the "log5" method of determining comparative win probability. That sounds complex, but it's not - basically, if one team is a .600 team and the other is a .550 team, we want to know how often the .600 team would win if they played a thousand times (hint: It's not .575). The formula is:

(A-A*B) / (A+B+2*A*B)

So in our example, the .600 team would be expected to win 55.1% of the time. We can then convert this to a money line/point spread - the 'real' line would be between 1.5 and 2.5 points (depending on your conversion method). Easy - and you can even do it in Excel after drinking all morning. Trust me, this is vital to any projection system.

The play-in game gives us a chance to test it out. Winthrop is a 4-point favorite over Arkansas-Pine Bluff, and when we look at the weighted "true" values for each team, it looks like our "true" spread should be anywhere from 2.8 to 3.5. Massey ( puts his guess at 3.7 in the same direction, so we know we're in the ballpark as the big boys here.

Looking at some other factors (courtesy of - one of the greatest sites in the f-ing universe), Pine Bluff plays a moderately-paced game based around a solid defense and pretty frigging awful offense, while Winthrop . . . does exactly the same thing. In fact, their numbers are nearly identical - you would sleep with one and not realize it wasn't the other, at least until it rolled over and gave off the MangleFace Vibe. I don't know where I'm going with this, but the teams are remarkably similar. As a result, we have no need to shift our profile at all, so we can say we lean juuuuust a little toward taking the points and riding "The Harvard of Pine Bluff" as hard as possible (NOTE: in this analogy, Shooter's Bronco-Bustin' NASCAR Bar is the Yale of Pine Bluff, and the local Valero is Brown). Also note: don't bet this game, it's really too close to take any edge.

No comments: