But, honestly, I don't find relying on computer projections to be that helpful. Ten different projection systems will give you ten different -- and sometimes wildly varying -- predicted scores.
I just saw another computer projection that had the score as 38-34 Navy.
Computer algorithms lack context, and so their utility is limited.
I agree, but I am just a stathead at heart. I've built everything I use now on what the eyes don't as easily see, because humans are biased and what you see with your eyes doesn't always paint an accurate picture. You can never rely on numbers alone, because if that were the case someone much smarter than me would already have the winning formula. I just like having hard data to back my picks.
I let the computers crunch all the numbers, and then I go to each matchup individually and add and subtract points on all the other "factors" that you can't gauge with numbers. It's not exact science but it is my numbers and I feel comfortable with them.
Because my system is so aggresive it is going to be wildly wrong about 20% of the time, but on those games where the "intangibles" don't greatly effect the game, I think I can find value.