After some more SQL, some pointing and clicking, some jiggery and a fair bit of pokery, some more useless graphs have appeared. The paper committee members rated all proposals on a scale of -2 to +2, and this of course lends itself to the creation of lies statistics.

After doing some more poking around in the database and twiddling with Gnumeric, I have another pretty graph with usseless stats. This one shows the ratio of accepted versus rejected talk and tutorial proposals by date.

Offhand it seems your chances of having a proposal accepted are higher if you submit in the week just before or after the deadline.