Reverse-engineer trading strategies from results
polymarket-reverse-engineerworkflowsetup L3★1
kimlage/crypto-edge-search ↗What it does
Reverse-engineer reproducible trading edge from skilled-wallet fingerprint
Best for
Designing independent, first-principles trading strategies from observed edge patterns, falsifiable through rigorous gauntlet testing, without relying on wallet-copying.
Inputs
- · [object Object]
Outputs
- · [object Object]
Requires
- · $0 free data only: Gamma metadata, CLOB prices+books, $0 data API
Preconditions
Proof phase complete; wallet cohort verified as profitable; all claims about cohort behaviour grounded in 500-market corpus; survivorship bias rules written
Failure modes
- · Proposing wallet-following strategies (already proven dead in proof phase)
- · Survivorship bias not eliminated from the edge claim
- · Gauntlet tests fail on holdout data
- · Strategy requires private information, not market state alone
Trust signals
- · Gauntlet phases explicitly sequenced (net cost → baselines → deflated Sharpe → CPCV/PBO → Harvey-Liu → right-null)
- · Named failure modes (long-beta disguise, h=0 tautology, selection inflation, etc.) mapped to each hypothesis
- · Right-null hypothesis framing explicit per hypothesis