Stockbee 20% Study Methodology

Purpose

The 20% study is a daily market observation routine. Its goal is to build a durable model book of unusual price moves, not to convert every mover into a trade.

The workflow answers three research questions:

Which stocks moved at least +20% or -20% over the selected window?
What catalyst, chart context, liquidity context, and market regime were present?
What happened 1, 3, 5, 10, and 20 trading days later?

Operating Principles

Record both winners and failures. Negative examples are necessary for setup fluency.
Separate observation, hypothesis, and execution. A 20% mover is an event, not a signal.
Classify before optimizing. The first pass should label catalyst and setup context; strategy design comes later.
Preserve data-quality flags. Missing bars, current-universe-only backfills, possible splits, and low liquidity change interpretation.
Prefer broad, stable cohorts over clever one-off rules.

Daily Review Checklist

For each high-priority mover, review:

Catalyst: earnings, guidance, M&A, FDA/clinical, contract, analyst action, short squeeze, theme sympathy, low-float speculation, capital structure, or no clear news
Direction: +20% upside event or -20% downside event
Window: 1-day, 5-day, or custom lookback
Close quality: strong close, midpoint close, weak close, reversal close
Volume shock: current volume versus 20-day average volume
Liquidity: current dollar volume and 20-day average dollar volume
Position in trend: near 52-week high, extended, breaking down, or attempting reversal
Setup type: base breakout, gap-and-go, episodic-pivot-like, climax extension, failed breakout, breakdown, or reversal candidate
Theme clustering: whether multiple related symbols moved together
Follow-up plan: which future horizons should be updated

Daily Cadence

Run scan after the daily bar is complete.
Run enrich when structured catalyst data is available.
Run update-outcomes to mature prior records.
Run summarize after enough records have matured.
Promote only evidence-backed findings into downstream research.

Historical Backfill Cadence

Backfills are useful for initial model-book seeding, but they require stronger data-quality notes. A current listed-symbol universe excludes delisted failures and can materially overstate continuation quality.

The CLI marks backfill records with CURRENT_UNIVERSE_BACKFILL_SURVIVORSHIP_BIAS by default. Use --survivorship-complete only when the supplied OHLCV includes delisted symbols and historical universe coverage.

Minimum backfill notes:

Universe source
Whether delisted symbols are included
Split and corporate-action adjustment source
Liquidity filter
Date range
Lookback window
Forward outcome horizons

Interpretation Boundaries

The output is valid as:

Study material
Event history
Cohort evidence
Edge-hint input
Manual chart-review queue

The output is not valid as:

A buy/sell signal
A broker order template
A guarantee of continuation or reversal
A substitute for position sizing, stop design, or regime gating

/stockbee-20pct-study

referencesmethodology.md