They did it with the IPA category and started out with 32 beers (!). He describes it thus:
He picked 32 IPAs available from his local grocery store, and ranked them based on ratebeer.com data and held a blind tasting (names were revealed only after beers lost) to determine round-by-round winners down to a single IPA Champion!The way brackets generally work is that the field is divided in half or in fourths, and the low seeds are matched against high seeds. I'd probably do a version with 16 (32 seems more than I could manage), split the field so that there were twin sides, with #1 versus #8, #2 v. #7 and so on. In the first round, you'd have eight head-to-head matchups, then four, then two, then the final pairing.
The reason it's genius is because it allows tasters to make a binary decision. Blind flights with several beers can be overwhelming--especially when the beers are similar in style. But even very similar beers can be judged in a simple pairing. I got excited enough that I created a stout brackets using the Hops and Barley system based on beers I know are available in Portland as an example. The intention was to actually run the pairings, but I may never get that far. I include it here for fun. The wild card was for to accommodate availability--I saw two or three that are available sometimes. Also, you could get a growler from a local brewpub. (Click to enlarge.)
Someone should really run with this. Good stuff--