Author Topic: Likelihood Distribution of Number of SpaceX Launches in 2015 -- DISCUSSION (Read 3214 times)

Proponent · « **on:** 12/19/2014 01:58 pm »

Gee whiz, it's been only a few minutes since I opened the poll on the likelihoods of numbers SpaceX launches in 2015, and already I'm realizing multiple mistakes I've made.

For starters, let us please use the present thread for discussion, questions and comments about the poll itself, so as to confine posts in the other thread to poll entries only. That will make tallying the results easier.

Proponent · « **Reply #1 on:** 12/19/2014 02:04 pm »

Lar, thank you very much for the inaugural entry. It is exemplary in its conformance to the requested format and in the clarity of thought behind it.

Lar · « **Reply #2 on:** 12/20/2014 04:29 am »

I fixed a few others but I got tired.

deruch · « **Reply #3 on:** 12/20/2014 07:34 am »

Can you explain what you mean by "scoring each entry"?

Tass · « **Reply #4 on:** 12/20/2014 09:12 am »

People should receive a number of points depending on the logarithm of the probability they assigned to the correct option. That way the expectation value of points is maximized by giving the best possible estimation of probabilities.

sittingduck · « **Reply #5 on:** 12/20/2014 10:57 am »

Tass, I think you have independently rediscovered the concept of gambling / sports-betting.

Proponent · « **Reply #6 on:** 12/20/2014 03:04 pm »

Quote from: deruch on 12/20/2014 07:34 am

Can you explain what you mean by "scoring each entry"?

Once the number of launches in 2015 is known, each entry's score will be the likelihood assigned to that number.

Quote from: Tass on 12/20/2014 09:12 am

People should receive a number of points depending on the logarithm of the probability they assigned to the correct option. That way the expectation value of points is maximized by giving the best possible estimation of probabilities.

That's a good idea, and can easily be implemented as well. And, of course, anybody is free to score entries in any way desired, since all of the information is public.

I've noticed that the single-choice poll on the number of SpaceX launches in 2015 runs until 19 January. Should the likelihood poll be extended so that, in some sense, the two polls incorporate the same information? (Though they still won't be strictly comparable, since the other poll, I believe, does not allow one to update one's vote.)

Proponent · « **Reply #7 on:** 01/05/2015 08:51 pm »

Quote from: averagespacejoe on 12/19/2014 03:05 pm

averagespacejoe: 4 5 7 10 10 10 15 30 10

This is a bizarre format that shows me why I was never an engineer.

Yeah, sorry about that, averagespacejoe. I realize it's not a human-friendly format -- it's just something I can collate easily. I'm sure that someone who's better with computers than I could figure out a way to scrape data in human-friendly form.

Proponent · « **Reply #8 on:** 01/05/2015 08:52 pm »

As previously threatened, I've extended the closing time for the likelihood poll to match that of the single-vote number-of-launches poll (19 January).