Yeah, it's primarily software projects where I developed my risk heuristic of "will backer updates and development logs be amusing enough, whether or not anything is delivered" and for videogames the subheuristic of "is the value of the soundtrack worth whatever state the game is in at release". A lot of KS games I have backed for their soundtracks and anything about playing the game itself has been a welcome surprise, usually. Most games on KS you have a good idea of who the composer will be and the likelihood of delivery of a strong product from that composer, no matter whatever else with the game.

I've actually think I had more projects not deliver that weren't software, though, than software projects. At some I should review the statistics. Under the "did the backer updates entertain" heuristic though, I don't count any real losses.