METHODOLOGY / HOW IT WORKS
How the scanner works
Understand each signal before deciding what deserves further research.
From universe to research queue
- Use the configured core universe, filtering securities below $3 or $5 million in 20-day average daily dollar volume.
- Read daily bars and calculate volume, price changes, opening gaps, daily range versus ATR, 20-day breakouts and price patterns.
- Rank candidates, enrich the top N with available option chains and news, then rescore with fixed weights.
- Publish up to 100 ranked rows and count threshold-crossing candidates across the full scored universe.
Seven inputs, fixed weights
| Input | Weight | Basis |
|---|---|---|
| Volume | 25% | Volume / prior 20-day average; ramps above 2× |
| Price move | 20% | Absolute daily move; ramps above 5% |
| Opening gap | 8% | Open versus previous close; ramps above 3% |
| Range / ATR | 12% | Daily range / ATR; ramps above 1.8× |
| Price action | 15% | Rule-based candlestick patterns mapped to 0–1 |
| Options activity | 12% | Unusual volume / open interest and put / call ratios |
| News sentiment | 8% | Absolute mean VADER sentiment on up to 20 headlines from 3 days |
The first four inputs ramp linearly above their trigger and cap at 1 at three times the trigger. Total score = Σ(input score × weight). Missing enrichment data does not redistribute weights.
The default opportunity threshold is 0.2. The original risk_level is high at ≥ 0.55, medium at ≥ 0.3, otherwise low; the interface labels these as intensity. Direction describes the daily price move and does not predict the next session.
Confirm the signal with source evidence
Check the trading date and whether earnings or another event explains the unusual activity, then review liquidity, company disclosures and your risk budget. Narratives are generated by rules rather than company-specific manual research. VADER is not finance-tuned, so sentiment is a rough indicator.