Independent Research into Parameter Sensitivity
CriteriaShift Atlas systematically models how subtle adjustments to fundamental screening parameters reshape underlying company cohorts and analytical conclusions.
Deconstructing the Mechanics of Screening Bias
CriteriaShift Atlas operates as an independent research portal engineered to analyze algorithmic stock screen drift, parameter sensitivity, and dataset volatility. We study how minor rule variations transform underlying universes.
Assumption Stress-Testing
We subject common filtering thresholds to marginal perturbations to measure structural survivor attrition across market cycles.
Zero Vendor Affiliation
Our datasets maintain complete analytical independence, free from commercial sponsorships, brokerage incentives, or software commissions.
Empirical Reproducibility
All parameter shifts, baseline criteria, and entity delta sets are documented systematically for peer replication and verification.
The Analytical Hypothesis Sequence
Every parameter exploration follows an established scientific evaluation sequence to isolate causation from correlation.
Starting Universe
Defines the total unweighted sample space without pre-filtering, documenting exchange exclusions and data lag.
Base Condition
Applies the baseline filter criteria, establishing the reference population and capturing standard survivorship markers.
One Changed Metric
Modifies a single variable or accounting standard to isolate the precise causal factor affecting company visibility.
Variance Auditing
Systematically tracks which businesses evaporated or entered, formulating rigorous structural research questions.
Addressing Algorithmic Black Boxes
Quantitative filtering software often encourages an uncritical belief in output precision. When a user defines a liquidity screen or revenue threshold, microscopic discrepancies in fundamental formulas can discard valid candidates or produce artificial clusters.
Our mandate is to publish systematic stress-tests that visualize parameter vulnerability, providing analysts with objective clarity.
- Accounting convention discrepancies between standard databases
- Survivorship and lookahead distortions in historical universe scans
- Threshold cliff effects in momentum and fundamental filters
Explore Detailed Screening Case Studies
Inspect our granular analysis on condition sensitivity, formula edge cases, and screening tools across popular platforms.