🏅 Top traders

So now we're betting if the feature stays?

@MartinRandall Correct. After the first version of fixed-payout multichoice as added for about 3 days last year I decided this was a better way to run these markets.

How would a native poll differ from "Will there be more YES holders than NO holders at the close of this market?" Do we need a CAPTCHA and IP Address logging to have greater confidence in a poll than just going by YES holders vs NO holders?

@ShitakiIntaki I'm not sure I understand the question. What you describe is a self-referential prediction market. A poll would allow people to select one of multiple answers and show how many times each option was chosen.

Do people use prediction markets as polls like that? If so, key differences include polls being free to answer, presenting no monetary incentive for anyone to 'answer' contrary to what they believe, and allowing more than two options.

@NcyRocks I guess as long as the format of the poll and the format of the market do not have a 1 to 1 correspondence. A Binary market with a single binary poll question would not add much value, same with a multiple choice market with a 1 to 1 correspondence in answers to a single multiple choice question. But if it is a whole survey, or if the market resolves to the poll percentage some of these issues are less patently abusable or otherwise incentivized to conform your poll responses to your market trades or vis a versa. Generally though I would expect a poll of manifold traders to be heavily incentivized by trading behavior in the market that the poll is presented in.

@ShitakiIntaki Are you assuming that native polls would have prediction markets attached? The current implementation of polls is a completely separate feature from the prediction markets.

@NcyRocks AH! yes I was assuming that polls would be presented as embedded within a market and the market probably asking a question about the poll results. Like "What percent of poll respondents will answer yes to this question, market resolves to poll %"

So polls would be more like the long form posts that are somewhere on the site,. (not sure where but I am pretty sure they exist) That would be AWESOME. I had some questions that I posted as free response "polls" but I don't have any legitimate way to resolve them since I am just asking a question, so I marked them each as "resolves to market". Not ideal.

I think there are comments on, one or more of the these markets how being a market really probably most definitely makes the responses suspect at best.

If that is the case a real poll function could be a more logical boost mana sink, where the pollster pays respondents and the respondents' return is not conditional upon their answer.

https://github.com/manifoldmarkets/manifold/pull/1908#pullrequestreview-1535653059

This PR was opened , it's for adding poll functionality. I like how you markets for whether something will be there on date X, instead of by date X. It's a better question because it filters out for what will stick around after being added, not rolled back.

@firstuserhere Thanks! Yeah, after the first version of fixed-payout multi choice was added for ~3 days last year I decided that this was a better way of doing it.

Native polls are complimentary with markets - it would be helpful to resolve a market based on the ultimate results of the poll