Skip to main content

Methodology / ε

Make sensitivity visible.

EPSILON operationalizes a mathematical neighborhood: if x is the baseline configuration and ε is one nearby input change, the relevant object is not only f(x), but the observed field f(x + ε).

01 / Claim

Pre-commit the interpretation

Write a falsifiable statement before computing the result. The statement remains user-authored and is never upgraded into an observation.

02 / Baseline

Fix the reference configuration

Record strategy, universe, window, capital, fee, minimum fee, and slippage. The baseline is one controlled computation—not a robustness claim.

03 / Perturb

Apply atomic changes

Change exactly one documented input per run. EPSILON currently tests fee, slippage, start window, and universe or minimum fee.

04 / Challenge

Keep failure attached

Direction reversals, failed runs, provenance, and the pre-committed rejection rule remain visible in the portable Evidence Plate.

Direction

“Preserved” reports only whether completed returns kept the baseline sign. It does not prove the user's full rejection rule.

Normalization

Every equity trace is normalized to its own first point so shape changes remain comparable across windows.

Provenance

Data mode, source, sampling interval, execution model, assumptions, and generation time travel with the artifact.