Biff
Member
Using markets like Betfair or PredictIt for actual probabilities of an outcome is incredibly flawed.
There are many factors built into the price of a prediction share beyond the actual likelihood of an event occurring. The biggest one to consider is the probability of getting paid.
These prediction markets have not yet lived through an event as major as the U.S. Presidential Election. There is a very good chance these websites will become illiquid and unable to pay out winnings.
This is what happened to Intrade, which failed shortly after the 2012 elections. I know friends who made $1000s in correct predictions, only to have the site shutdown without actually paying out any of the winnings.
Basically, my point is you can't look at a prediction site that has Trump winning at 65% and say "oh the market is predicting 65%." The market is probably predicting something like 80-90%, then building in a substantial risk premium on actually getting paid.
There are many factors built into the price of a prediction share beyond the actual likelihood of an event occurring. The biggest one to consider is the probability of getting paid.
These prediction markets have not yet lived through an event as major as the U.S. Presidential Election. There is a very good chance these websites will become illiquid and unable to pay out winnings.
This is what happened to Intrade, which failed shortly after the 2012 elections. I know friends who made $1000s in correct predictions, only to have the site shutdown without actually paying out any of the winnings.
Basically, my point is you can't look at a prediction site that has Trump winning at 65% and say "oh the market is predicting 65%." The market is probably predicting something like 80-90%, then building in a substantial risk premium on actually getting paid.