Verdict Accuracy Backtest
137 report snapshots · 13 completed checkpoints · 398 pending
73 past checkpoints pending price lookup
— run
npm run backtest:fetch locally to fill⭐ North Star — +30% within 14 days
The actual job of a 30% hunter, measured. MAE = max drawdown endured before the target hit.
1%
hit rate · n=117
-13.0%
median MAE
| Catalyst type | n | Hit +30% | Median MAE | Avg max gain |
|---|---|---|---|---|
| 📊 Earnings | 56 | 2% | -8.6% | +5.3% |
| 🚀 Product Launch | 14 | 0% | -17.7% | +1.9% |
| 🤝 Partnership / Contract | 14 | 0% | -16.8% | +5.7% |
| 🛰️ Satellite Launch | 9 | 0% | -18.6% | +7.4% |
| 📦 Other Catalyst | 6 | 0% | -2.8% | +6.2% |
| 💊 FDA Decision | 3 | 0% | -13.7% | +2.5% |
| 🚚 Delivery / Volume | 7 | 0% | -24.8% | +0.1% |
| 🏭 Foundry / Chip Validation | 2 | 0% | -16.1% | +1.7% |
| ⚖️ Regulatory Approval | 6 | 0% | -11.2% | +5.0% |
🎰 Turbo — +50% within 30 days
Moonshot picks (web/data/speculative/) are lottery tickets, not 2-week catalyst events — measured on their own thesis, kept out of the core stats/table above.
n = 84 turbo snapshots total · 41 complete · 43 accumulating
5%
hit rate · n=41
-24.7%
median MAE
25.5d
median time-to-hit
Time-to-hit distribution (days)
Day 24Day 27
| Ticker | Report Date | Status | Max Gain | MAE | Hit +50% | Time-to-hit |
|---|---|---|---|---|---|---|
| ABAT | Aug 10, 2026 | accumulating | +1.9% | -3.4% | ✗ | — |
| ACHR | Jun 26, 2026 | complete | +6.3% | -12.1% | ✗ | — |
| ACHR | Jul 2, 2026 | complete | +8.0% | -10.7% | ✗ | — |
| ACMR | Jun 24, 2026 | complete | +21.5% | -21.0% | ✗ | — |
| AREC | Jul 14, 2026 | complete | +59.3% | -13.7% | ✓ | 24d |
| AREC | Jul 22, 2026 | accumulating | +60.2% | -5.5% | ✓ | 16d |
| AREC | Aug 4, 2026 | accumulating | +31.2% | -8.1% | ✗ | — |
| AREC | Aug 13, 2026 | accumulating | — | — | — | — |
| BBAI | Jul 10, 2026 | complete | -1.5% | -22.0% | ✗ | — |
| BIVI | Jul 22, 2026 | accumulating | +58.1% | -22.5% | ✓ | 14d |
Page 1 of 9 · 84 total
📊 Calibration — Avg 30d Return by Verdict
BUY
n=80
-10.1%
Hit rate: 13%
SPECULATIVE
n=48
-20.3%
Hit rate: 0%
HOLD
No completed checkpoints yetn=9
🕵️ Near-Miss Log — Selection Counterfactual
Candidates that passed the screen but missed the top 5. If picks don't beat near-misses, the ranking isn't adding alpha.
Picks avg 30d: -14.0%Near-miss avg 30d: -7.1%Selection edge: -6.9%
| Screen | Stock | Why missed | +14d | +30d |
|---|---|---|---|---|
| Aug 13, 2026 | NUVL | Two confirmed PDUFAs (Sep 18 zidesamtinib, Sep 26 neladalkib) inside the 6-week window — the best catalyst provenance on the whole longlist. Failed on upside: price ~$123.4 vs consensus $142.05 (18 analysts) = +15%, short of even the widened 20% gate. One source quoted '$135 forecasts +30.35%', which is arithmetically impossible against a $123 price — stale figure computed off a lower base. Cut on the gate, not on the catalyst. | — | — |
| Aug 13, 2026 | IRON | Cleared upside easily (price ~$64.8 vs consensus $103.18 from 11 analysts = +59%) but failed catalyst verification per backlog #37. The aggregator's 'DISC-1459 Phase 3 Est. Readout Sep 2026' is a CT.gov estimated completion date, not company guidance — actual guidance puts bitopertin Ph3 EPP topline in Q4 2026 with an FDA decision mid-2027, both outside the 42-day window. | — | — |
| Aug 13, 2026 | MLTX | Cleared upside (price ~$18.5 vs consensus ~$27.6 = +49%) but the catalyst had already happened: VELA Ph3 readout is done, positive 52-week data reported and Week 40 presented at AAD 2026. What remains in September is a BLA submission — a filing, not a binary outcome. Its 'Ph3 Est. Readout Sep 26' is the same CT.gov completion artifact as IRON (#37). Target dispersion ($2-$46) also signals the market has not settled on a value. | — | — |
| Aug 13, 2026 | VNDA | longlist-cut: headline upside was the largest on the board (price $6.20 vs $14.13 average target across 4 analysts = +128%, B. Riley at $17) but the catalyst does not verify. The aggregator's 'Sep 30 Hetlioz PDUFA' is an AI-detected row; the only real regulatory date found is a Jan 7 2026 sNDA re-review for jet lag disorder. Rejected on provenance (#37) despite passing the gate by the widest margin — exactly the trap the new rule exists to catch. | — | — |
| Aug 13, 2026 | RVMD | longlist-cut: consensus is unresolvable at screen time — median $140.00 across 27 analysts (range $73-$170) versus a $214.71 mean, straddling the 20% gate depending on which is taken, against a price around $136. Sep 30 daraxonrasib regulatory decision is AI-detected, not confirmed. Cut for measurement ambiguity rather than a clear fail; worth re-screening when consensus tightens. | — | — |
Page 1 of 23 · 113 total
Prices fetched from Yahoo Finance · Cached in web/data/backtest-cache.json · Not financial advice