Upcoming in Theory and Decision (in the same issue as I), David Johnstone asks a fundamental prediction market question: should we believe traders who make the most money, relative to traders with good statistical forecasting scores?
Paul, there are several possible reasons to declare predictions, but this paper doesn't care about the reason, it is looking at the consequences of such declarations.
Hmm... if there's an active prediction market, why would (competent) predictors ever want to (honestly) declare their predictions? Or is the idea here that they'd work something like the people who publish stock picking advice now -- that there'd be a layer of the prediction market that finds it more profitable to sell their opinions than to bet on them? (Does anyone know of any work comparing the records of the best of those public analysts to the best traders who don't declare their picks?)
This might be too technical a subject for me to put my mind around it... How could a zero ask-bid spread yield a profit score equivalent to a statistical (or probability) score? Traders won't trade when his/her profit forecast is negative…
Paul, there are several possible reasons to declare predictions, but this paper doesn't care about the reason, it is looking at the consequences of such declarations.
Hmm... if there's an active prediction market, why would (competent) predictors ever want to (honestly) declare their predictions? Or is the idea here that they'd work something like the people who publish stock picking advice now -- that there'd be a layer of the prediction market that finds it more profitable to sell their opinions than to bet on them? (Does anyone know of any work comparing the records of the best of those public analysts to the best traders who don't declare their picks?)
Paul, predictors can just declare their predictions, which are then later collected and compared with the true answers, producing an accuracy score.
How are predictors revealing their predictions other than through their bets?
Yan, it is a technical question, so try following the link first and see if that explains it well enough.
This might be too technical a subject for me to put my mind around it... How could a zero ask-bid spread yield a profit score equivalent to a statistical (or probability) score? Traders won't trade when his/her profit forecast is negative…