So to sum up, I liked the use of prediction markets to estimate the conditional general election probability for a candidate at a point in time, and I think it’s worth getting some estimates of how particular events moved this probability. I think at this stage the conclusions are a bit underdeveloped and oversold, considering how many factors are at play and how unclear it is what information each primary introduced. But I look forward to future revisions.
-
The paper: PDF file
-
