eMidas
Bounded optimisation

Challenge a baseline without losing the trail.

Run deterministic parameter candidates, optionally add a GPT-assisted proposal, and rank every result against a transparent multi-objective score.

Start researching Explore all features
eMidas / Bounded optimisation
Bounded changesFull ledgerHuman promotion
What is included

Professional depth,
without hidden logic.

AI proposes research candidates; it does not deploy or guarantee improvements.

  • Selectable optimisation targets and parameter bounds
  • Deterministic and GPT-assisted candidates
  • Live stage, progress and candidate updates
  • Return, drawdown, Sharpe and trade-aware scoring
  • Permanent experiment ledger and save-as-new flow
How it works

One connected
research movement.

  1. 01

    Set the allowed research boundary and objective.

  2. 02

    Backtest each isolated candidate against the same baseline.

  3. 03

    Review the rationale and validate promising candidates out of sample.

Research boundary

Evidence supports judgement. It never replaces it.

AI proposes research candidates; it does not deploy or guarantee improvements.

Turn your next idea
into evidence.

Create your workspace