Research
What we have learned
Findings earned from graded outcomes, with their evidence and their limits. Where the sample is too small to claim skill, it says so.
L-001 · 2026-08-05
activev2 conviction had no demonstrated predictive power, and the point estimates ran backwards
point for a first v3 review, never a judgment to defend or anchor to. Measured on the pre-rebuild snapshot: each conviction observation paired with the stock's 21-day forward return, minus the universe median over the same window, so the market is controlled for. | conviction band | obs | excess % | ±95% CI | effective N | reading | |---|---|---|---|---|---| | 85-100 highest | 182 | -2.86 | ±3.53 | 54 | not distinguishable from luck | | 70-84 strong | 107 | +0.47 | ±3.99 | 51 | not distinguishable from luck | | 55-69 constructive | 92 | -0.34 | ±2.59 | 54 | not distinguishable from luck | | 40-54 weak | 80 | +0.31 | ±3.48 | 51 | not distinguishable from luck | | below 40 broken | 172 | -1.18 | ±3.42 | 57 | not distinguishable from luck | **Spread, highest band minus broken band: -1.67 pp.** A working scoring system produces a positive spread here. This one produced a negative one. **Hone
Evidence: 633 paired observations, 2026-04 to 2026-08 · Confidence: medium on the null result, low on the inversion