Re:Experiment HN: Crowdsourcing Tech Stock Predictions
by aantix on Tuesday November 30, 2010 @10:40PM
(#1956705)
The biggest issue with this design is that there's no incentive to put in effort for better predictions.
Take it to the next level; design this for Mechanical Turk (https://github.com/aantix/turkee), and make this offer; for each prediction received, get a nickel. If you're on the right side with your prediction, receive an extra nickel. If you're within 0.5%, receive an additional nickel.
Not enough people utilize the "bonus" feature of Mechanical Turk as financial incentive to increase the quality of data they receive. You don't have to pay everyone the same. Reward better input.
Take it to the next level; design this for Mechanical Turk (https://github.com/aantix/turkee), and make this offer; for each prediction received, get a nickel. If you're on the right side with your prediction, receive an extra nickel. If you're within 0.5%, receive an additional nickel.
Not enough people utilize the "bonus" feature of Mechanical Turk as financial incentive to increase the quality of data they receive. You don't have to pay everyone the same. Reward better input.