Strongest result across all runs. Treatment earns +7.83% while BTC falls -34.7%. Alpha concentrated in 2 crash detections
| Arm | Return |
|---|---|
| Treatment (Briefings) | +7.83% |
| Control (Price Only) | -8.14% |
| Placebo (Stale Briefings) | -7.70% |
| Control-WS (Web Search) | -1.55% |
Extended to 202 ticks (through Mar 21). Refined composite regime classifier. Treatment SHORT accuracy: 65% vs Control 50%.
The longest window, Sep 2025 through Mar 21, 2026. BTC fell -34.7% overall, but the final 15 days (Mar 7 to Mar 21) added a new dynamic: a tentative recovery from $71K to $74K, followed by another pullback. The extended tail tested whether the treatment could handle choppy, range-bound conditions after the crash.
This is the strongest result across all 7 runs. The treatment arm returned +7.83% while BTC fell -34.7%, a spread of over 42 percentage points versus buy-and-hold. The treatment-vs-control delta of +15.97pp is the highest recorded, and the treatment-vs-control-WS delta of +9.38pp confirms the edge is specifically in structured analysis.
The treatment arm's alpha was concentrated in two crash detections. First, it went short on Nov 8, 2025, two days before the November selloff accelerated, based on cross.regime flagging deteriorating breadth. Second, it went short on Feb 3, 2026, two days before the February crash, when btc.energy flagged hash ribbon capitulation and cross.breadth showed equity correlations spiking.
The control arm returned -8.14% over 202 ticks. Its trade log shows a clear pattern: buying every dip without distinguishing between a pullback in an uptrend and a leg down in a bear market. It went long 14 times during the Oct to Feb decline, each time getting stopped out at a lower price. By March, it had accumulated significant fee drag from the repeated round trips.
(3 more observations in the full report)
Previous: Run 6 | All Runs | View Tick Data
Read the full research findings | How we test