I'd be very interested in the data and the code behind the generator function. I'd love to integrate it with my bracket-generator module (which is very basic and only does lower seed, higher seed, and random currently).
In his post he said he had trouble finding venues that would work with him on this. I know last time through Arizona he played Comerica Theatre (capacity 5500) and on this tour he is playing the Celebrity Theatre (capacity 2650). I would guess that this specific venue change was due to Comerica Theatre not wanting to work with his exclusive ticketing.
I'll bet you that if you look closely enough you'll find the Comerica (and most other equally large venues) is owned by Livenation/Ticket Master and as such will only deal with acts who work though Ticketmaster. That's the hardest problem to solve with ticketing, the fact that most venues of a decent size won't touch you unless you also use their ticketing service.
Where did you get your probabilities from? This is a cool idea and although it's getting too late to be useful now, I think it would be nice to implement in my Twitter based bracket challenge, http://tweetyourbracket.com. Sorry for the shameless plug, I'm just trying to get the word out about it since I finished it about 8 hours ago.
And I'm with you in the horribly written source code boat. I had my idea the Sunday before the bracket came out and wrote code until last night. Gonna take another look at the code once the tournament is over, we'll see how that goes.
I like it! I was thinking about adding something like that to http://tweetyourbracket.com to automatically fill out a bracket with random picks, chalk or anything else.
Not trying (completely) for a shameless plug here. I just love things that combine code and brackets.
The source is on GitHub for anyone interested. I've also opened a few issues for things that I want to do for next year (and maybe this year depending on popularity), so if you find anything feel free to post an issue.
This came out of an idea to see how/if brackets could be stored as strings and then manipulated, displayed, scored, validated, etc.
All the code is on GitHub. https://github.com/tweetyourbracket