David and I met for lunch over Thai food, with me getting an update on his company's decision process and visualization tools.
Consider a science fair. One may be hard pressed to find even a single judge who has compared a given pair of projects (A,B), given the comparison matrix increases as a 2nd power of the number of projects. Flooding the scene with judges means too little overview on the part of each judge and/or contestants overtaxed with too many interviews.
However, even with a sane number of judges, chances are very good of finding multiple cases wherein two judges have a 3rd project in common, i.e. the pairs (A,X) and (B,X) both occur for multiple Xs. The differences (A-X)-(B-X) give a sense of (A-B), the comparison we wanted to make. Collating all such instances may net significant new information, which David's latest prototype is good at extracting.