It seems you're outputting other's peoples' lists where there's at least one game in common with those I like.

Couldn't you take this a step farther and aggregate all those lists, with those games that appear the most frequently showing up at the top? For bonus points, weight more highly the games coming from lists that more strongly match (and using Bayesian averages, too!)

So I typed in "Backgammon" and "Go" and it doesn't seem to do anything.