Backtest first,
then trade
Retail entries and exits often run on gut feel, with no backtestable basis and no discipline. This tool makes “validate before you trade” the default flow.
- Market data
- Clean & align
- Screening / signal rules
- Backtest (history)
- Parameter comparison
- Risk gate → live order
Backtest and live share the same signal logic to avoid “great in backtest, loses live” drift; every order passes a risk gate first.
Problem
Entries and exits on gut feel are hard to review after the fact; without a backtestable basis, you cannot tell luck from strategy.
Constraint
Strategies must be backtestable and parameters comparable; it can reach live execution, but one bug must not fire wild orders; market data quality varies and has to be handled first.
Decision
Backtest and live share one signal implementation to kill drift between them; a risk and confirmation gate sits before orders; paper first, small size next, scale last.
Result
Entries and exits have a backtestable basis and strategies can be compared; live runs the same logic, cutting the “pretty backtest, ugly live” risk.
What was hard
Survivorship bias and time alignment in historical data, the gap between backtest and live environments, and the easiest trap of all — overfitting parameters that look great on history and break live.