Guesses are better, far better, when given numerically and as you wisely included, with a date. The date provides 'closure' and thus a method to measure how accurate the numerical guess is. As Sikander posted...that is how prediction markets work. Only problem with 'markets' of prediction is they use money as a medium for prediction, which introduces an unnnecessary additional variable in the prediction. Prediction tournaments are better... Metaculus, GoodJudgmentOpen

On

Meta /Facebook 0.3

Twitter in USA 0.1

Twitter in Turkey 0.7

etc.

Then there is the challenge of measuring 'censored'. But these challenges are part of the landscape of forecasting.

You did say platforms, so Nostr as a protocol, is excluded.