Should there be a trend in my simulation model? I assumed not, as market prices in principle can't be EXPECTED to trend for extended periods. The race obviously goes back and forth, so whatever trend we see in the last few days or weeks, it's foolhardy to imagine will continue for the next ~50 days.
Empirically, the correlation between today's returns and yesterday's returns over the last 90 days for the general 'Obama wins the presidency' contract is -.1. Not only is it close to 0, but it is actually negative. Meaning if the price went up one day, it is slightly more likely to go down the next.
No comments:
Post a Comment