Validation
Public accountability: what the scanner published, what actually happened, and the exact rules each scan ran on. Research and education only — not investment advice.
Published Picks — Deterministic Backtest
By pattern
| Pattern | Picks | Resolved | Win rate | Avg gain |
|---|---|---|---|---|
| Cup & Handle | 1 | 1 | 100.0% | +5.17% |
| Darvas Box | 2 | 2 | 0.0% | -3.57% |
| MA Pullback | 1 | 1 | 100.0% | +4.77% |
| Pocket Pivot | 2 | 2 | 100.0% | +7.73% |
| Three Weeks Tight | 1 | 1 | 100.0% | +1.72% |
| Unclassified | 37 | 37 | 83.78% | +3.12% |
| Wyckoff SOS | 2 | 2 | 50.0% | +2.47% |
By legend methodology
| Legend | Picks | Resolved | Win rate | Avg gain |
|---|---|---|---|---|
| darvas | 2 | 2 | 0.0% | -3.57% |
| minervini | 1 | 1 | 100.0% | +1.72% |
| oneil | 3 | 3 | 100.0% | +6.88% |
| unclassified | 37 | 37 | 83.78% | +3.12% |
| weinstein | 1 | 1 | 100.0% | +4.77% |
| wyckoff | 2 | 2 | 50.0% | +2.47% |
Every number comes from BreakoutAttempt rows the scanner already published — the same table behind the public track record. Outcomes are the stored resolutions (pivot to resolution close) using only as-of dates; open picks never enter win rate or average gain. Re-running this on the same rows always returns the same numbers — no lookahead, no cherry-picking.
Game Plans — Shadow Track Record
Every published game plan is snapshotted (entry, stop, target) and scored against the bars that actually printed afterwards: target hit, stop hit, or expired untouched. Paper only — the plans were never trades.
No game plans tracked yet — snapshots begin with the next published plan.
Scan Tearsheets vs SPY
One tearsheet per scan definition, built only from point-in-time data — resolved pick outcomes compounded in chronological order, benchmarked against SPY buy-and-hold over the same span.
William O'Neil's CANSLIM cup-with-handle base: a rounded 7-65 week consolidation, depth 12-33% (up to 40% in a weak market), then a handle breakout on volume. The institutional-edge setup.
1
100.0%
—
0.0%
—
—
Nicolas Darvas's box theory: a stock trading in a defined box breaks above the box top on expanding volume. The self-made-trader momentum setup.
2
0.0%
-100.0%
-7.1%
-4.64
—
Mark Minervini's volatility-contraction setup: a base where each pullback gets smaller on drying volume, then a pivot breakout. The scanner's precision-sniper setup.
0
—
—
0.0%
—
—
Scan Mandates — the exact rules, versioned
Each scan definition pins its filter thresholds in a versioned mandate file. What you see above was measured under exactly these rules — change a threshold and the version bumps, so history stays reproducible.
| Mandate | Version | Legend | Pattern | Trend ≥ | RS ≥ | Confidence ≥ |
|---|---|---|---|---|---|---|
cup_oneil |
v1 | oneil | Cup & Handle | 6/8 | 65.0 | 0.6 |
darvas_box |
v1 | darvas | Darvas Box | 6/8 | 65.0 | 0.6 |
vcp_minervini |
v1 | minervini | VCP | 6/8 | 65.0 | 0.6 |
Legend AI is a research and educational tool, not a registered investment adviser. Nothing here is a recommendation to buy or sell any security.