Backtesting Software
helio / quant-backtesterDescribe the strategy logic, the instrument, and the parameters. Your AI teammate implements and backtests it, then scans for signals during market hours so you're not watching a screen all day.
Implement and backtest a moving-average crossover on SPY over 2020–2025, with 0.03% commission and 0.1% slippage, then report annualized return, max drawdown, and Sharpe. Once it's validated, scan for entry signals during market hours and alert me, and never place a trade. Ask me what you need before you start.
Swap in your strategy logic, instrument, and backtest window.
What is Backtesting Software?
Describe the strategy logic, instrument, and parameters, and it implements the backtest, runs it against a historical window, and reports metrics with the data range and assumptions labeled. It scans for signals during market hours and alerts you, it does not give investment advice, and it never executes a trade.
What it does
Any chatbot backtests a strategy once. Helio runs it as a standing job — from your context, kept current, delivered to you.
| Owns the job, not just the run | It writes the backtest, logs the assumptions, and keeps scanning for signals during market hours. |
| Runs again without starting over | Saved once, the same strategy re-runs on new data and scans every trading day, not from a blank prompt. |
| Works from your context | It uses your market data API and local code directory, versioning the backtest alongside your research. |
| Keeps itself current | Change a parameter or a signal rule and it updates the code instead of waiting to be re-taught. |
| Lands where you work | Backtest results and live signal alerts arrive in Slack, Lark, or email during market hours. |
Three things it gets right
A backtest that reproduces
Window 2020–2025, 0.03% commission, 0.1% slippage. Every parameter is written into the report, so the run repeats.
Overfit flagged, not hidden
Stable for a 3–7 day window, degrading past 10 days. Parameter sensitivity is surfaced as a possible overfit.
Alerts, never executes
A signal at 10:47 produces an alert with position size and stop-loss checked, never an automatic trade.
What Backtesting Software connects to.
Questions
- Does this give investment advice?
- No. It implements and backtests strategies you describe and reports what happened — it never recommends a trade or claims a result will repeat. Every real decision is yours.
- Can it execute trades automatically?
- No. Signal detection produces an alert only. Trade execution is a deliberate, separate step you control.
- How do I know the backtest isn't overfit?
- Every result is labeled with its data range and assumptions. Treat a strong backtest as a hypothesis to validate further, not a guarantee — the report says so explicitly.
- What data sources does it support?
- It connects to whatever market data API or historical dataset you provide access to — it doesn't come bundled with a specific paid data source.
Stop watching the market by hand.
Describe your strategy — Helio backtests it and scans for signals during market hours. You still make the call.
Ask about this automation