A 90-day replay designed to show when an idea would have fired, what a stated fill model would have done, and how the result can be reproduced.
THESIS TRADING / PROVISOA chronological list of rule triggers, not just a single return figure.
Fill modelling around the mark with explicit slippage. Actual execution can differ.
An outcome tied to its input data, model, and fee assumptions.
The peak-to-trough decline in the simulated series, not only the ending balance.
The test identity uses SHA-256 of the rules, data snapshot identifiers, and tester version. A test can then be rerun against the same inputs instead of relying on an unrepeatable screenshot.
Historical coverage includes Hyperliquid candles, funding and open interest, Robinhood Chain pool history, and Fear & Greed. X-post replay is a later extension dependent on available data and cost.
Confirm market availability, sampling, missing-data treatment, slippage, funding, fees, and how partial fills are modelled. These details can change the apparent result materially.
A replay cannot promise execution in a fast move, guarantee liquidity, or remove liquidation risk. A rule that worked in one period may fail in another.
Keep the rules, data snapshot identifiers, and tester version with the report. An outcome without its input history cannot be independently reproduced.
Choose a saved thesis and define the replay assumptions before submitting to a connected tester.