"As far as how to optimally distribute money to charity, that is very much an unsolved problem, but I think it's one that we can mostly worry about when we get that far."

I like the rest of your proposal, but I seriously think we need to look more carefully at this part. Once a billion dollars is already on the line, it's worthwhile for large charities that won't do very much good to spend $100M on marketing for a 12% chance at getting it, which does no one any good at all (except the marketing companies). If we make the decision beforehand- even if it is completely arbitrary (eg., we take all the charities recommended by GiveWell and put them on a giant roulette wheel)- then charities won't spend large amounts of money competing amongst themselves for the money, which would defeat the original purpose.