// Why Pro costs less than planned
A price list is the first place a failed study should show up.
This ladder used to carry a fifth tier — Signal, $2,000 to $10,000 a month — selling a volatility forecast built on prediction-market repricing as a feed to funds. Pro’s headline was the same forecast.
We replicated the paper it rested on, on our own data. It did not hold: t = −0.16 against the paper’s 3.63, and out of sample the signal made the forecast worse. So the node is not being built, the Signal tier is deleted, and Pro is $99 rather than the $109 it was going to be — because it is now a views tier that shows what markets imply, not a signals tier that tells you what happens next.
Charging the original price for the same page with the best thing removed would have been the easy move, and nobody outside would have known.
t = −0.16 · 313 days · p = 0.88 · out-of-sample MSFE ratio 1.039 · 66 regressions, all reported