Author here: I did a quick experiment with my Greyhound, Bebop, to figure out the treat he prefers best using pair-wise comparison analyzed with the Bradley-Terry model. Same tech as Elo scores in chess, and several other places! Enjoy!