In November of last year I published a post demonstrating that using VIX + normal distribution, or +/- 1 std 68% rule is far far different from real distribution of short-term SPX returns. To validate this, and demonstrate concrete results to the readers I started publishing weekly market forecasts where I predict 50% confidence intervals for the next trading week. If my methodology was well calibrated we would see SPX finish the week inside the range 50% of the time, and 50% outside of the range - this what 50% confidence range means.
Last week I wrote that after 2 months ( 8 forecasts ) we can start looking at the data and drawing conclusions. The results are in: 4 out of 8, or 50%, which is obviously the best result I could hope for.
No comments:
Post a Comment