2009 Point Spread Performance

Hello everyone, I am back from a loooong vacation from NFL point spread predictions. Work and family have kept me busy, so I haven't "trained" this off-season. I wanted to let you know that this blog is still alive and that we will be doing the same thing we did last year, hopefully with the same success.

For the past 3 years, I have been publicly putting my statistically derived picks and my reputation on(the)line. I don't pick every game, only those with high probability of success. NFL point spread picks is all we do here. I have crunched more than 15 years of data, tried more than 100 different statistical methods, and had above 60% success rate in the past 3 years.

Last year I started providing a betting strategy. This strategy is based on Kelly's formula which determines the percentage of your bankroll to use each week. I modified this formula and used our success probability to weigh each pick of the week. This helped us last year to reduce risk and improve ROI to the highest I've ever gotten so far, 93.9%.

This year's picks/bets start slow the first 4 weeks of the regular season to capture enough data, and then we switch gears for some real profit making. I agree with Brian Burke in that predicting (hence betting) pre-season games is a worthless effort.

Recap of the 2009 Season and Playoffs
  • ATS record of 38-23, 62.3%
  • Return on Investment (ROI) = 93.9%, almost doubled starting bankroll
  • ATS record by week:
    • Week 1-4: 66.7%
    • Week 5-8: 73.3%
    • Week 9-12: 50%
    • Week 13-16: 58.3%
    • Playoffs: 62.5%
  • Best Week: Week 12, 3-0, $290
  • Worst Week: Week 9, 0-4, -$340
  • 6 Weeks in a row with gains, and only 2 weeks in a row with losing record
  • 12 winning weeks, 3 losing weeks, and 2 weeks breaking even (does not count Super Bowl)
Conclusion: We had a fun roller coaster ride. In the end it paid well, almost doubled the starting bankroll. Some weeks were rough and some were great and we knew this going in. Which is why I and you should never put all your eggs in one basket and do not rely on one week or game to make miracles. Minimize risk and look to make money in the long run.

Looking forward to the 2010 NFL Season,

Jaime
admin@nflpickles.com

4 comments:

Dave said...

I have a few questions, if you don't mind answering:

Did you end up changing your model from 2008 to 2009?

If so, did you back test the new model you used for 2009 on the 2008 season?

And if so, how did it do?

I ask because your archive indicates you had some trouble in 2008, but I also know that 2008 was a tough season for other modelers and am curious as to whether you figured out a better method that worked for the 2008 data, or whether you've accepted that it was just a particularly tough year to predict, and that your system will work well for the long haul. Thanks.

Jaime said...

Great question Dave! I am using the same model I used since 2007.

I think in 07 though I had more than 10 models running at the same time. In 08 I discovered the Prediction Tracker so I stopped publishing reports for all models and focused on "my baby".

Before that, when I was in grad school learning about this, I back-tested more than 100 different methods of making picks.

Yes, 2008 was a tough year, I was just above break-even point, 53%. Which is why I am more pumped to see its success for two years in a row.

Julian said...

Glad to see you back in action!

DDW said...

Hey Jaime,

Good to see you back in the swing of things again. I can't wait for this season to get started. I'm going to try and do a better job of posting my picks vs. yours this year. I kind of slacked off mid-way through last season.

Lets make it a good one this year!