Originally posted on 09/09/2017:

I started wagering on football 50 years ago this fall. I started computer analysis with NFL box score data shortly thereafter, using as large IBM mainframe at a university. I wanted to know what statistics could be used to predict a team's score. I used a mathematical technique called multiple linear regression in an attempt to determine a formula. Unfortunately the biggest factor in a team's score was the number of their opponent's turnovers. This was followed by their turnovers. About a decade later there was an article published in Sports Illustrated that used the same technique and drew the same conclusions.

If you think about it, predicting turnovers is the most difficult thing to do. Actually the message should have been, what you're trying to do is probably impossible. But in spite of this I've continued to try throughout the decades over sports ranging from horse racing to baseball. I can honestly say that in spite of the thousands of hours of effort, I've only found one technique that I'm willing to stake a fair amount of money on year after year. While this one does win in the long run, there have been entire seasons where it was a loser.

I've reached the twilight of my years, but I still continue to look at new approaches, gather lots of data, write programs, and generate spread sheets. Do I expect to find the "Lost Dutchmsn Mine" of sports wagering? No more. But. here is the key. I really enjoy the work. It is my hobby.

So if you pursue it as a hobby and have low expectations, you are in for a lifetime of entertainment.