MetaChat REGISTER   ||   LOGIN   ||   IMAGES ARE OFF   ||   RECENT COMMENTS




artphoto by splunge
artphoto by TheophileEscargot
artphoto by Kronos_to_Earth
artphoto by ethylene

Home

About

Search

Archives

Mecha Wiki

Metachat Eye

Emcee

IRC Channels

IRC FAQ


 RSS


Comment Feed:

RSS

01 November 2006

I want to sponsor a MeFi election contest. Help me choose the rules. MeFi's past contests have been for the presidential elections, where the electoral vote provided a pretty clean scoring system. It's a bit harder to design one for the midterms. It seems that a good contest will weight the house and senate fairly equally. It's probably pointless to get people to predict the outcome of individual house races, but not for the senate. More inside...
The best I can come up with so far is letting people predict the final breakdown of the house, ranking them based on that, letting them predict the final breakdown of the senate, ranking them based on that, and then adding both rankings for the final score, lowest score wins. Ties broken by the number of specific Senate outcomes that are predicted correctly. Further ties broken based on who posted first.

Would this work? Can anyone think of a better alternative?

Winner will get two MeFi compilation CD's, the idea obviously being that at least one will be given as a gift, spreading the love.
posted by gsteff 01 November | 21:56
If you want maximum participation, I'm not sure that including every Senate race is a good idea. That might make people feel the same way that I do when somebody wants me to be in their fantasy sports league. I mean, I'm a fan, but I don't know if I'm that big of a fan.

But do you want maximum participation? Or a fun discussion (or some other kind of discussion)? Or are you trying to get out the vote? The different goals might mean different strategies.

That said, how about picking out just a few Congressional races--ones that are pretty close, and/or ones that involve well-known people? That's an alternative, anyway.

Or you could shoehorn every single election into some kind of Final-Four-style tree. Or lay odds on every race, and then let people bet imaginary money. Those ideas are much worse.
posted by box 01 November | 22:48
The goal is definitely maximum participation. I was kinda worried about that too. I'll try to simplify the senate side to just predicting the final party breakdown.
posted by gsteff 01 November | 22:55
Another tiebreaker could be pick the actual percentages for a pick 'em Senate race or two. (E.g. Tester 53%, Burns 46%)

Another, I hate to say, is to predict the percentage difference in exit polls versus "actual" votes.
posted by danostuporstar 01 November | 22:58
Is it undesirable that someone who guesses nearly every race correctly may lose to someone who gets nearly every race wrong, but predicts a correct final breakdown?
posted by box 01 November | 23:00
The vote percentage tiebreaker is a good idea, dano. I'll use Connecticut.
posted by gsteff 01 November | 23:11
Connecticut's a great Metafilter choice.
posted by box 01 November | 23:59
And it's on. Thanks to gsteff for putting it out there.
posted by hangashore 02 November | 09:59
. || I used to be puddinghead...

HOME  ||   REGISTER  ||   LOGIN