Start with a hypothesis.
Define the instrument universe, signal rationale, timeframe and data requirements before optimising parameters. Record what would invalidate the idea.
A good trading system is more than a backtest. It is a set of assumptions, controls and operational decisions you can inspect.
Define the instrument universe, signal rationale, timeframe and data requirements before optimising parameters. Record what would invalidate the idea.
Separate development and holdout data. Include fees, spreads, slippage and funding where applicable. Test sensitivity, liquidity and parameter stability.
Check signal timing, order handling, missing data and reconciliation without risking trading capital. A successful simulation is not proof of future profitability.
Use explicit position and exposure limits, alerts, audit logs and a kill switch. Deployment requires the client’s own venue accounts, infrastructure and permissions.
Research reports should state drawdown, turnover, cost assumptions, sample period and out-of-sample behaviour. We do not publish invented performance figures or imply that a larger licence buys greater returns.
Trading can lose capital. Backtests and paper trading have limitations. Denki sells software licences and optional compute credits; purchase payments are not managed investment funds.
Explore the range