> what I’m looking at is probably the most likely avenue for me to make a helpful contribution.
Personally I'm eager to run a two-level neural network on the problem and see how well it does, along with various other "off the shelf" prediction methods, merely to 1) learn about different methods, 2) see how well they do on different variants of this problem, 3) start studying different ways to define "how well they do". I don't mind getting mediocre results! I just want to learn stuff.
It looks like maybe Dara can help me with these things.