Crashing the ASC continued

• It never occurred to me that Bayesians needed a model selection criterion -- I guess I just thought it was beside the point for them -- but of course there are situations where it's necessary. So there's a DIC, and it's biased and people are trying to think of ways to unbias it. It's a culture shock to be in a situation where Bayesians are all over the place (still a minority though).
• This cross-entropy method seems too useful for me not to have heard of it before. Maybe I was asleep when someone last told me about it.
• I'm trying to ask friendly but useful questions at these sessions, mostly just to give the speakers the impression that someone has at least somewhat engaged with their presentation. But I never know what to do when I hear a completely wrongheaded (or just wrong) talk, especially because I can never tell if the audience passivity is the usual kind or the "please shoot this guy" kind. I'm not the kind of guy who can call a spade, and I rarely care enough to discuss it with them afterwards. Hence I often end up wrapping an irrelevant or unanswerable question in the kind of half-witty one-liners that audiences in that situation lap up.
• But beware, I've witnessed a couple of speakers have David Brent moments this week.
• My 25 year-old friend K is being guided through an overdue adolescence by his mate R. K thinks R is teaching him how to pick up girls, but actually R is just teaching him to be a dodgy mofo. I mean strip clubs are all well and good but they're NOT THE PLACE TO MEET WOMEN, unless you're willing to spend a lot of money (although see yesterday's post). So to those in my posse who think I'm corrupting them: you don't know what corruption is.