Xark!: NFL Picks 2006 wrap-up

My first year of picking football games in public is over, and here's my record: 172-82, with a winning percentage of .677. I dropped below .700 with a 6-10 mark in Week 16, but recovered with a relatively strong 10-6 record in the always unpredictable final week of the regular season.

What else did I learn? For the truly geeky, meet me after the jump link for my "Lessons Learned"...

When I'm confident, it's a good sign
During the final 10 weeks of the 2006 season, I thought it might be interesting to divide up my picks by confidence level (Confident, Not Betting the Farm and Just Guessing). I figured that the "confident" section would consist of fewer games than either of the other categories because it would be limited to "no-brainer" picks.

As it turns out, I gave more Confident picks (48) than NBTF choices (46). I don't know why, but it's a surprise. Anyway, I finished 37-11 (.770) on Confident picks, 30-16 (.652) on NBTF picks and 34-25 (.576) on games where I was Just Guessing. If you combine the Confident and NBTF categories, my winning percentage was .713 on 94 games, compared to .576 on the 59 games I "just guessed."

In other words, there really is a correlation between degrees of confidence and winning percentage. It's actually a pretty strong trend. One could argue that the most important skill in picking football games is the ability to spot which games are toss-ups and which games aren't.

Also, it will be interesting to see what happens next year when I do confidence levels from Week One on. My best guess is that the Just Guessing percentage will wind up closer to .500 when I add more games to the sample.

Picking against the spread isn't impossible
I always heard that Vegas has a technical term for anyone who bets on pro football: Loser. But when it came to my short experiment on picking against the Vegas line, that wasn't quite the case.

I picked against Vegas for the final four weeks of the season, finishing 40-22 (.645). Next year I'll pick all 17 weeks and see what happens, but if you can stay near .650 picking covers, that suggests that consistent betting could wind up being profitable. Not wildly so -- you have to bet $110 to make $100, and there are taxes and legal issues to consider -- but if I would have bet $110 on each of my ATS picks in the past month, I would have come out more than $1,500 ahead before taxes.

My original thought was that if I applied my "confidence approach" to gambling, I'd beat my overall average. Makes sense, right? If you only bet selected games, shouldn't you finish better than you would by including the games where you really don't have strong feelings? It works for picking winners (.770 "confident" versus .576 "just guessing") so why shouldn't it work when you're picking point-spread winners?

Maybe when I do it next year I'll do better, but the reality is that I barely finished above .500 on my "Best Bets" (5-4, .555) in December (I didn't start this until very late in the season). If I'd have bet $110 per game on that set of picks I'd have made about $60 on about $1,000 worth of risk. Anyway, it's clear that I really don't have an eye for spotting an inaccurate Vegas line... yet when I spread the risk out over 16 games, I actually perform better. This will be interesting to watch next year to see if it changes.