This will be the last time I pick the NFL winners based on guesses. It might not be a solid model at first, but it can’t be consistently worse than my average. Knowing me, what should take an hour will be more like eight, but whatever. I also don’t know if people want me taking a picture of a football pool, one which I did entirely for fun, and tagging it as #madewithcode. I’m sure some people might be peeved with that.
I’m still not sure what the original algorithm is going to be, the first iteration of you will, but that should hopefully work out fairly fast. Then as the days and weeks go by, just add another layer of data and have it determine the best multipliers and exponents on its own. Of course, I’ll have to start looking at secondary variables fairly soon, but that’s all doable.
Peter's stuff.