How to Backtest a Trading Strategy on TradingLead

This guide walks through a complete backtest in TradingLead's Backtesting panel using a Python strategy. If your strategy is a Pine Script strategy(), follow Backtest a Pine strategy instead. The setup and reading steps are the same.

Before you start

• You need access to the TradingLead charts workspace (5-day free trial, then subscription; see pricing).
• Decide what you want to learn. A backtest is most useful when you have a specific rule set, for example "buy when the 20-period average crosses above the 50, exit on a 2% stop or 4% target".
• New to backtesting? The Backtesting Lab shows the common mistakes in about 20 minutes.

Step 1: Open the chart you want to test

Go to tradinglead.in/charts. Pick the data feed, symbol (for example a Nifty 50 stock or index available on your feed) and interval. The backtest captures exactly this chart context. There is no separate symbol box in the backtest form, so set the chart first.

Step 2: Open the Backtesting tab

In the bottom panel, open Backtesting. (The neighbouring Paper Trading tab uses the same engine for manual practice.)

Step 3: Choose the strategy

Set the language to Python and choose a script you saved in Indicator Studio, or keep the bundled example. Then choose an execution mode:

• Vectorized: fastest; processes up to 20,000 candles. Your code sees the full captured history, so make sure it only uses past values when generating each signal.
• Sequential: your code runs after every completed candle with only the revealed prefix and the actual simulated account (balance, position, pending orders). It is limited to 250 candles, and it suits account-aware sizing.

Step 4: Set realistic assumptions

Open the setup and Advanced settings:

Under Instrument & cost assumptions you can also choose lots, a contract multiplier, tick size and a user-assumed margin model. These are your assumptions, not verified broker rules. Taxes are not modelled.

Finally, choose the history range: leave dates empty to use the latest closed candles up to the cap, or enter explicit dates. An optional session schedule lets you pin trading hours, holidays and special sessions for intraday data.

Step 5: Run and wait for completion

Click Run backtest, review the preflight summary, then click Start backtest. The run first shows as preparing while history is fetched and the strategy executes. Wait for completed. If it fails, read the message: TradingLead stops on missing history, invalid candles or session mismatches instead of quietly testing fewer candles.

Step 6: Read the results honestly

Start with the summary, then open Analytics:

• Net P&L after fees, and net P&L including open mark-to-market.
• Max drawdown: the worst peak-to-trough fall in candle-close equity. Ask yourself whether you could sit through it.
• Win rate and profit factor: read them together, since a high win rate with small wins and large losses can still lose money.
• Expectancy: average net P&L per closed position.
• Number of trades: a handful of trades is not enough evidence.
• Buy-and-hold benchmark for the same period: did the strategy actually beat simply holding?
• Costs: total fees paid, and how much of gross profit they consumed.

Sharpe ratio and annualized volatility appear only after you enter explicit sampling assumptions (samples per year, interval, risk-free rate).

Step 7: Inspect trades on the chart

Entry and exit arrows appear on the chart. Use Chart display / range to show labels for all trades or only the selected one, plus connecting lines, stop/target segments and open-position levels. Click any row in the trade grid to jump to it on the chart. Check a few trades by eye: do entries happen where your rules say they should?

Step 8: Save, compare or go further

• Save a stopped run to the saved-run library, and compare runs tested on the same candles.
• Stress-test inputs with the Strategy Optimizer, preferably with a holdout or walk-forward split.
• Rehearse the rules manually with paper trading.

Common mistakes to avoid

• Testing only a period where the idea obviously worked.
• Leaving costs at zero. Frequent intraday strategies are especially sensitive.
• Using future values in vectorized code (lookahead bias). TradingLead cannot certify that your code is free of it.
• Tuning many settings on the full history and trusting the best one.
• Treating a backtest as a forecast. It is a simulation on past data, not investment advice.