NIFTY----BANKNIFTY----RELIANCE----HDFCBANK----INFY----TCS----TATASTEEL----GOLD----SILVER----COPPER----CRUDEOIL----NATURALGAS----
Playground
Backtest a strategy
Pick an instrument and a rule, drag the parameters, and the server runs it over real daily history and hands back an equity curve. Buy-and-hold is drawn alongside every run, because most simple rules lose to it and a backtester that hides the benchmark is selling something.
Strategy
Long while the fast average is above the slow one.
Window
How it is calculated
- Long only, one position at a time. No shorting, no leverage, no position sizing.
- Signals act with a one-bar delay — a rule derived from today's close is filled at tomorrow's. Entering on the same bar that produced the signal is the most common way a backtest quietly cheats.
- Costs are charged on entry and again on exit, and the slider goes to 50bps a side. A frictionless backtest is a fantasy; drag the cost up and watch a high-turnover rule stop working.
- An open position at the end is listed but excluded from the win rate. Marking it to market as a win would flatter the number.
- Daily bars from a free delayed provider. No survivorship adjustment, no dividends, no liquidity or slippage modelling beyond the flat cost.
This is a toy for showing how the machinery fits together, not evidence that any of these rules work. Past data is the one thing a backtest is guaranteed to fit. It is not advice, and nobody should trade off it. The same engine is exposed at /api/v1/backtest.