FYERS Adds Strategy Backtesting and New Indicators to Automate Platform

The Indian broker is pitching it as a way to validate automated strategies before deployment, and the free pricing plus the metric set it returns make it worth a serious look for anyone running systematic NIFTY or bank-nifty strategies.

What the backtester actually returns

The module runs your coded strategy against historical bars and prints the four numbers I look at first: returns, win rate, drawdown, profit factor. Every simulated fill is written to a complete trade book you can export. That last part is the real value — a single equity curve screenshot is a marketing asset; a raw trade log is data. If you can pull the fills out, you can rebuild the return distribution, stress-test the tails, and check whether the wins cluster in a single market regime or hold up across trending, choppy, and high-vol windows. FYERS uses a NIFTY breakout as its example use case, which is fine: it lets you finally see whether the edge survives something like a trending 2023 versus a sideways 2024 before you commit real capital.

The indicators — useful, not magic

Squeeze Momentum is a TTM Squeeze descendant — Bollinger Bands contracting inside Keltner Channels, then a momentum burst on the exit. ORB is the textbook opening-range breakout: first fifteen or thirty minutes defines the range, and a break on either side triggers the entry. Neither is novel if you've read a technical-analysis book in the last decade. What matters is the implementation: close-based versus intrabar breakout triggers, the slippage assumption baked into the signal, and whether the squeeze uses LazyBear's standard parameters or something modified to look prettier on screenshots. Default settings on any bundled indicator are almost always the ones that curve-fit best on the marketing material. Pull the source code or the settings sheet before you trust the signal.

Caveats that should temper the excitement

Beta means edge cases aren't fully handled yet, and access apparently requires commenting with your Client ID. Survivorship bias is still live on any NIFTY historical set that wasn't built point-in-time. Free backtesting also pulls in the worst kind of user: someone who runs five hundred parameter combinations until one looks profitable — that's not an edge, that's curve fitting with extra steps. Watch drawdown duration and the distribution of losing streaks before you get excited about profit factor. If those cluster tighter than your sample size can statistically support, the backtest is confirming your priors, not the market's behavior.