Each row is timestamped at detection, scored with the analyst credibility frozen at that moment, and tracked to outcome. Dollar impact assumes a $5K unit budget. No retroactive grade-inflation, no cherry-picking. M2 swaps these seed receipts for live data.
The funnel from universe to helped. We surface this so you can audit our rigor before trusting the helped-rate at the top of the page. Live receipts begin M2.
Loss-inclusive · no cherry-picking · sample is small by design (Signal Lens is M1; the funnel widens with every week of live ingest). Min sample size before publishing a per-analyst stat: n ≥ 5 (verified at n ≥ 12).
Just landed. Regime fully supportive + high credibility → FULL. 1d +3.4%, 5d/20d windows still open.
HALF on ELEVATED VIX + NEAR_FLIP gamma. 5d running +2.8% — sized for the chop. 20d still open.
Engine called SKIP — ELEVATED VIX + NEAR_FLIP gamma weighed against sizing up. The move ran anyway. A STARTER would have captured \$710. Logged as a regime-bias false-negative.
Engine called FULL on a high-credibility BULLISH setup; earnings whiff −9% in 20d cost \$455 on a \$5K unit. Engine wrong — included as a calibration receipt.
Engine called STARTER (\$1.25K); the move ran +22% in 20d. A FULL would have captured \$840 more. Mid-conviction analyst + neutral flow → engine biased cautious.
Engine called HALF (\$2.5K) — sized for the thesis but trimmed for ELEVATED VIX volatility. Net +\$295 vs sizing-blind average.
Low-credibility analyst + EXTREME VIX → engine called SKIP; a sizing-blind FULL short would have lost \$470 on the squeeze.
Regime supportive + high analyst conviction → engine called FULL; +41% in 20d delivered \$2,060 on a \$5K unit vs sizing-blind STARTER avg.
Signal Lens Pro tracks every call from the X analysts you trust and gives you the same sizing receipt — per call, per regime, per dollar of your unit budget.