Immutable settings
Run parameters lock at first start so later edits cannot silently rewrite the meaning of an existing result.
Metanyx turns market events and strategy decisions into reproducible evidence, connecting an experiment to its data, configuration and exact source.
A backtest is useful only when its assumptions and implementation can be recovered. Metanyx binds settings and strategy source to each run.
Run parameters lock at first start so later edits cannot silently rewrite the meaning of an existing result.
Strategy source is hashed and snapshotted, connecting performance and diagnostics to a specific implementation.
Run bundles make settings, identifiers and results portable for review and comparison.
Order-book strategies depend on event sequence, queue conditions and acknowledgements. Replay preserves those mechanics more faithfully than candle-only evaluation.
Books, trades, candles and BBO observations are persisted as research inputs without blocking trading I/O.
Simulated acknowledgements, cancels and confirmed fills exercise the same state transitions that strategies face live.
Comparable inputs make parameter changes and strategy drafts easier to evaluate without conflating them with a different market sample.
An integrated assistant can examine strategy code and results, run comparable tests and prepare drafts. Humans retain the review boundary before installation or launch.
The assistant works from strategy source, declared parameters and measured results rather than an isolated prompt.
Candidate settings or logic can be exercised against the same replay windows and diagnostics.
Generated changes remain drafts until a person reviews and explicitly promotes them into the strategy lifecycle.
Talk with Metanyx about quantitative research, adaptive strategies and guarded execution.
GET IN TOUCH ↗