Thoughts

[nealmick]

The recent addition of new features and spread data has provided invaluable insights. I've discovered that the spread can fluctuate significantly. what I mean by this is you can predict a game, save the spread, and then look again a hour later and the odds change. The spread has also been getting less accurate over the past decade. A strategy where the margin between the spread and prediction is maximized could produce greater expected value compared to using randomly timed lines. Hitting the peak deviation from the prediction. With a break even model at around 52.5% EV, capturing variance in spread could be profitable.

[nealmick]