A backtest tells you what a rule would have done against a clean historical dataset. Paper trading tells you what it does against the market as it actually arrives: late, incomplete, and occasionally wrong.
The two answer different questions, and passing one says nothing about the other.
What paper trading catches that a backtest cannot
| Failure | Why a backtest misses it |
|---|---|
| Data arriving late or not at all on a rebalance morning | Historical files are complete by the time you download them |
| An order rejected by the broker | No broker is involved in a simulation |
| A stock under a surveillance measure or a trade-for-trade band | The historical price series still has a price for it |
| A corporate action landing on your rebalance date | Adjustments are already applied in history |
| Your own reaction to a 15% drawdown you are watching in real time | A backtest drawdown is a line on a chart you already know recovers |
That last row is not a technical point. The documented reason retail traders lose money is behavioural persistence, not arithmetic: SEBI found 93% of individual F&O traders made net losses over FY22 to FY24, and the majority kept trading anyway (SEBI press release, September 2024). Watching a simulated position fall is the cheapest available rehearsal.
What it cannot catch
Paper fills are still assumed fills. Your simulated order never consumes the order book, so it never pays market impact, and it never queues behind anyone. For a retail-sized order in a Nifty 50 name that gap is small. For a smallcap position sized at several days of average volume, the paper result and the real result are different numbers.
Paper trading also runs forward in real time, which means a three-month run gives you three months of sample. That is enough to validate plumbing and nowhere near enough to validate an edge.
Availability on saral.money
Paper trading and live deployment are in preview, enabled per user rather than open to everyone, with broker integrations for Dhan and Zerodha. Treat the description above as what the capability does, not as something to go and switch on today.
Caveats
The main risk is misreading the evidence. A profitable paper period is a small sample from one market regime, run without impact cost, and it is easy to convert into conviction it does not support. Use it to confirm the mechanics work end to end, and keep your position sizing anchored to the backtest and to what you can afford to lose.