Apr 14

I’m trying to understand what proportion of Manifold users meditate regularly and if that will change over time if I poll again next year.

This market will resolve in 30 days based on this poll:


I will not participate in either the poll or this market.

Get Ṁ200 play money
Sort by:

How does this resolve ? Is it based on your impressions after seeing the poll results, or is there an explicitely defined formula ?

@Mich Impressionistic only, unless you can propose a better analytical approach.

@suzumebatchi Well that’s your market so your call but by explicitely defined formula I was thinking for example of resolving to probability = percentage of yes answers in the poll + λ*percentage of “intermittently” where lambda is in [0,1] and is for you to decide depending on how much you think someone who meditates intermittently counts as meditating. Or if you don’t want to resolve to probability, something like YES if the formula I suggested yields a result above a certain threshold, NO otherwise


Thanks for this input, Mich. I may check back in with you when I resolve the market.


Thanks so much for your thoughts. Before signing in to resolve this market, I decided to set lambda to 1/7, rather arbitrarily, but loosely on the frequency of one day per week out of seven. Tell me if I have this right: Yes = 38%, 1/7 = 14%, therefore, if I add the 38% and the 14%, we get 52%, resolving the poll to yes. Does that sound right?

@suzumebatchi If I understand well, you’d like to use my formula with lambda = 1/7 AND use a threshold of 50%, resolving to YES if the result is above and NO if the result is below ?

I tried to compute everything using the answers on the poll and I can’t get the same result you get. I think there are at least two reasons : first, it seems the poll is still open so we may not have used the same information as input. I suggest either closing the poll or if you’d like to use the results from earlier, sending a screenshot. Second : to make people who meditate intermittently count only for 1/7 meditator each, you have to multiply the proportion of intermittent answers by 1/7. If there were 28% of « intermittently », then my formula yields 38% + (14/7)% = 40%, resolving to NO if using a threshold at 50%. (I’m assuming the 14% you mentioned were the proportion of votes on “intermittently”, because if you had already divided by 7 to get the 14% the proportion would have to be 98% of intermittent meditators)