Last December, my brother finally released his top-100 chart of video game music. It took me a few days to understand his process: although he names one track per game he lists, we shouldn't consider it is a video game tracks chart. He's really trying to compare full OSTs to one another.

That was a lot of things to check at once, especially since there are games in this chart I had never played, and was aware of their music only through overclocked remixes. So I started a twitter thread, commenting one track a day, bottom up. And there are many points where we apparently disagree. (of course, the fact that we start with the soundtrack for Fury of the Furries out of the top 100 didn't help).

As I progressed towards #50, one thing turned obvious: my brother and I do not approach making charts the same way. Not to say that his way is bad. And neither of us do that through statistics. I figured out I see charts as the result of a sorting process. Being a "computer scientist", I know that sorting implies we have a comparison function that is capable of telling which of two items is the greatest. If items have multiple dimensions, it relies on simpler functions applied to their constitutents (e.g. comparing their catchiness and the quality of their instruments separatedly), then combining those things together - either through priorities or through weights - to get a final comparison result.
I also know that if I have a pre-sorted list and a new challenger, I should ...