Saturday, October 17, 2009

Seeking Bayes

There is a rumor of a mayoral debate on TV this afternoon I do believe..  The highly self-selected NS readership knows that of course.  But I am wondering what the ratio is of NS readers to those who will be impacted by the debate.   Consider:
• Pa: Probability someone knows there is an election in 2 weeks
• Pb: Probability someone knows there is a debate airing today at noon on a Saturday
• Pc: Probability they were around to watch tv at noon on a Saturday at noon*
• Pd: Probability they didn't find something else better to watch at that time
• Pe: Probability they didn’t move on after watching the first 10 minutes or less
• Pf: Probability that anything they saw changed their support among the candidates presented
• Pg: Probability they are registered to vote in the city
• Ph: Probability that viewer actually makes it to vote on November 3

* Yes, I am ignoring the TIVO factor and those who saw the clip on the Internet.

Once you assign probabilities to each of those calculate the probability that the debating airing today mattered from: Pa x Pb x Pc x Pd x Pe x Pf x Pg x Ph
I would argue each factor in itself is bracketed by 0-20%.  Some factors may not be that low, but some are certainly on the low end of even that range. I am guessing Pf itself may be near zero. and yes I know it certainly isn't the case they are each independent of each other… but still for fun:   (0.2)^8*310K people in the city rounds up to one vote.    Compare that to the dozen or so folks who must have been in the studio itself between the candidates, Delano, Rice, camera people, producers and the candidates handlers.

Too much math?… let’s skip that. Anyone know the ratings in themselves to see who was watching. Any more than the 20 political junkies (meant generically) in town.
and of course... none of that was the point.   The goal for the challengers at least was to score a soundbite or possibly a gaff that would be replayed in the news or elsewhere. No early indications of anything like that as best I can tell.

Johnnyg said...

Repeat after me: The mayoral race is a foregone conclusion. People don't pay attention to foregone conclusions.

Which sets up something fascinating. As the budget battle demonstrated, nothing can get done in the evenly split legislature. That likely means that the Supreme Court will decide redistricting after 2010. The Orie Melvin-Panella race is predicted to be very close, and will give one party numerical control of the Court. I bet the Democrats wish they had a mayoral race that would actually bring out City voters this fall.

So far, I've seen no GOTV by the Democrats on this critical race. Which is amazing, because the Republicans in the "T" are famous for their excellent GOTV efforts.

Sunday, October 18, 2009 9:39:00 PM
MH said...

Plug in those numbers for a presidential debate and I bet you'd get nearly the same result as far as the impact of the debate. You pretty much have to have a gaffe that gets in the regular news for any debate to matter.

Monday, October 19, 2009 10:11:00 PM