Back when Zach and Dennis were doing TWIP, they had a segment where Dennis did an audio version of the drag and drop list. I thought Dennis covered his thought process well. I thought at the time that this is how the top 100 should go.
Count me in for a "Head to Head" top 100 list too. I say Head to Head as I don't want to tell Robin how to code and implement it, lol.
Thinking about the implementation some more could it just be a series of machine a vs. machine b questions and an algorithm (I know it is easy to say when you don't have to develop it) to compare all the results? I wonder how many questions it would take to get a realistic sense of ranking all 500 odd games?