Pillar 01
Research Questions
Formulating inquiry parameters with conceptual rigor before touching syntax. Distinguishing investigative questions from mechanical indicator overlays.
Methodological Architecture
The Six-Stage Market Scan Workflow
Quantitative screening fails when technical sliders are adjusted without prior inquiry. A rigorous screener translates qualitative market questions into verifiable criteria, bounded universes, and mandatory post-scan chart inspections.
Stage 01Formulation & Hypothesis
Formulating the Research Question
Every scan must begin with an explicit structural hypothesis rather than software indicator knobs. What specific market phenomenon are you isolating? For example, whether an asset is consolidating near a multi-month range high or responding to post-announcement volume shifts. When the inquiry is ambiguous, output interpretation becomes impossible.
- Core Objective: Isolate one verifiable economic or structural condition.
- Common Error: Confusing technical indicator values with an actual market question.
- Outcome: A plain-language inquiry document defining observable targets.
Analytical Guardrail
"If your question cannot be described in natural sentences without naming an indicator oscillator, the research question is not yet framed."
Stage 02Population Bounds
Universe Delimitation
Filter equations operate on whatever data pool they receive. Applying a volume scan to the entire equity market produces drastically different distributions than evaluating large-cap index constituents or specific industrial sectors. Universe selection sets the statistical denominator for every filter.
- Boundary Conditions: Market capitalization bands, minimum daily dollar volume, index listings.
- Common Error: Scanning an unbounded composite exchange list filled with illiquid penny instruments.
- Outcome: A clean, uniform population sharing baseline liquidity mechanics.
Universe Dependency
The exact same criteria applied to a broad market index versus a cyclical semiconductor industry basket will output unrelated qualitative behaviors.
Stage 03Quantifiable Parameters
Observable Criteria Design
Market participants often describe abstract traits like "strength" or "consolidation". In this stage, qualitative notions are converted into observable, verifiable mathematical metrics such as ATR-normalized range spreads, rolling average volume multiples, or moving benchmark relationships.
- Metric Construction: Relative percentage calculations rather than raw nominal values.
- Common Error: Stacking three identical oscillators that calculate the same underlying price delta.
- Outcome: Distinct mathematical definitions devoid of redundant indicator correlations.
Orthogonal Logic
Each criterion must evaluate a distinct market dimension: price dispersion, volume activity, or time persistence. Redundant metrics create false precision.
Stage 04Syntax & Boolean Rules
Automated Filter Execution
With clear questions, delimited universes, and defined criteria, conditions are translated into screener syntax using nested Boolean operators (AND, OR, NOT). Execution timeframes and aggregation periods are set to ensure indicators do not look across disparate time scales.
- Syntax Translation: Expressing conditions in scripting environments like thinkorswim Stock Hacker.
- Common Error: Nesting OR conditions improperly, causing filter leaks or zero results.
- Outcome: A deterministic query that executes reliably across active trading sessions.
Execution Integrity
A scan filter is a mechanical search query, not a predictive engine. It returns instruments meeting specified mathematical thresholds at that exact second.
Stage 05Population Evaluation
Result Set Output Analysis
The screener output is a population of candidates, not an action directive. Analysts evaluate the result set distribution: if a scan returns 180 tickers, the filter criteria are too broad; if it returns zero, the conditions are over-constrained. Population size informs query calibration.
- Distribution Review: Sector concentration, median volatility, and candidate count.
- Common Error: Assuming a filtered list represents a ready-to-execute opportunity set.
- Outcome: A manageable cohort of 5 to 15 candidates prepared for visual inspection.
Cohort Reality
Candidate outputs illustrate market sector rotations and regime behavior across the chosen universe rather than individual security directives.
Stage 06Post-Scan Audit
Mandatory Manual Review
The final step can never be automated. Each instrument in the result set must be visually inspected against higher-timeframe charts, structural liquidity patterns, upcoming corporate events, and data anomalies such as illiquid gaps or bad print ticks.
- Qualitative Inspection: Verifying chart context that screener formulas cannot compute.
- Common Error: Bypassing manual inspection and treating raw scan outputs as final answers.
- Outcome: Identification of false positives and calibrated feedback for Stage 01.
Human Verification
Manual review eliminates data artifacts and discovers whether the filter logic actually isolated the scenario intended by the original research question.
Focused Methodology Explorations
Curated research studies illustrating filter logic and manual review stages
2026-08-16
Boolean Operators and Condition Nesting in Screener Design
Deep dive into building logical parenthetical groups, chaining AND/OR criteria, and avoiding boolean leaks during multi-variable automated execution.
2026-08-25
Constructing an Objective Visual Verification Checklist
A structured inspection protocol for evaluating post-scan candidates, flagging illiquid wick anomalies, and refining query criteria.
