I'm playing in the MIT Mystery Hunt right now and am on a little break (shower & shave before dinner). In puzzle events like this it's common for the answers to puzzles to be words or short phrases that are later used as input to other, "meta" puzzles. The grouping of "feeder" puzzles together and matching to a meta is usually left to teams and can be a fun exercise. This year's Mystery Hunt had a small change I rather liked. When we got a new puzzle we got a raft of data about it (thematically-matched data), data that wasn't used to solve the puzzle per se but that was used in solving the meta. The neat part is that this twist lets teams figure out which puzzles should be grouped together, even before solving these puzzles. It's not a simple task as "collect all the puzzles with a triangle on them" but it's not as hard as solving the entire puzzle. Nice touch. It's also a nice distraction from staring at the puzzles we've long since had a fresh perspective on.