Research

Replay and simulation lab

Replay historical opportunities through the real deterministic scorer, policy engine, and router in a sandbox. This lab does not create live signals, executions, or positions.

Research / Replay

Controlled policy and routing replay

Compare baseline vs candidate thresholds against recorded opportunities and outcomes. This is a research surface for internal validation, not a production action path.

New research run

Select a time range, keep the baseline, and test a candidate threshold profile without touching production state.

sandboxed only
Baseline thresholds

Leave blank to use the current policy defaults.

Candidate thresholds

Use candidate values to test a more selective or looser policy without touching production behavior.

no live writespaper simulation only

Recent runs

Each run replays the same deterministic logic over historical opportunities in a fully sandboxed context.

No research runs yet

Create a sandboxed replay to inspect how baseline and candidate policies would have behaved over recorded opportunities.

Run detail

Inspect how policy changes affect signals, paper-execution readiness, and blocked-vs-allowed outcomes.

Select a run

Choose a research run to inspect the summary, candidate deltas, and changed opportunities.