The weeks are passing and the contests are getting more difficult to pick. It's no longer Powerhouse U versus Sister Mary's School of the Blind, so getting more than 80% right is becoming an increasingly difficult task. For the most part teams are starting to settle into their positions for this year and the rankings are starting to reflect that. Depending on how quickly we age out old data, somewhere between 45% and 55% of the current rankings are based solely on the four or five games teams have played this year. This slow phasing out of old data prevents the systems from reacting too quickly to a single game (anyone who thinks UCLA beats Texas by 50 consistently in a best of 7, please raise your hand .... right, thought not), but rather adapt to the season as it unfolds.
That said, here's the good, the bad, and the ugly from week 5.
Justin
Accuracy: 36-13 (73.5%)
The Good: Not much. I suppose I should take consolation in the fact that if I had not done some slight tweaks to my system over the summer, I'd be in even worse shape than I am now.
The Bad: In 4 of the 5 games where I disagreed with Eddie, I was wrong. For three of those games my computer was less than 50.3% certain in the predicted outcome, and the fourth was a whopping 53.8% confident. In effect, I lost four coin tosses in a row. Two of those were even one-point games.
The Ugly: Boise State at New Mexico State. Not because I got it wrong, but just because this is -- by our GUGS metric -- the fifth least interesting game of the entire season, getting a whopping 11.6 on our 100-point scale. This isn't surprising considering that New Mexico State is ranked dead last by TFG; the Aggies are involved in four of the five worst matchups this season, including week 4's snorefest against Kansas that ranked dead last in GUGS (10.7).
Eddie
Accuracy: 39-10 (79.6%)
The Good: Not only did I go 4-1 head-to-head against Justin, but I was pretty confident about it. I had Baylor at 56.2%, Notre Dame at 62.5%, and Northwestern at 62.4%. Louisiana-Lafayette was effectively a coin flip at 51.6%.
The Bad: Clemson over Miami. I lost 12 points in the TotalFark pick'em thanks to this one. Clemson did everything in their power to give this game to Miami, including three turnovers inside the Miami 30. There's no rational reason for this game to have been a 68.3% confidence.
The Ugly: Louisiana Tech over Hawaii. I was the only person in the entire TotalFark pool to pick the Bulldogs, and it looks absolutely silly in retrospect. This game was over at halftime.
Overall Season Totals:
Eddie: 190-45 (80.9%)
Justin: 186-49 (79.1%)
