How to backtest a crypto strategy without Python

Who this is for

You trade or follow crypto and already think in rules: RSI below 30, SMA cross, Bollinger fade, trend filter. You want to see whether those rules held up on BTC or ETH before you put money behind them. You are not trying to build a quant research stack this afternoon.

Torquant is built for that job. You describe the strategy in plain English. It runs the simulation on built-in pair data and shows you the equity curve, drawdown, and buy-and-hold comparison. Long, short, and leverage are supported when you specify them. Price history uses spot market candles; funding and liquidation from perpetuals are not modeled.

Step by step

1. 1

   Open Torquant

   Go to app.torquant.app and sign in. You land in the strategy chat. There is nothing to install and no API key to wire up for price data on supported pairs.

2. 2

   Describe your strategy

   Type what you want to test in everyday language. Include the pair, timeframe, entry rule, exit rule, and any indicators with their settings. If you are starting from zero, click Guide me and Torquant will walk you through market, entries, exits, and position size one question at a time.

   Be specific enough that someone else could follow the rules. Vague prompts produce vague backtests.

3. 3

   Answer clarifying questions

   Torquant may ask for details you skipped: exact pair (ETHUSDT vs BTCUSDT), candle interval (1h, 4h, daily), date range, position size, or whether you want a stop-loss. Answer in plain language. One missing detail is one question, not a form to fill out.

4. 4

   Run the backtest

   When the strategy is complete, click Run Backtest. Torquant simulates your rules on historical candles for the pair and window you chose. The first run usually takes a short wait, not a setup project.

5. 5

   Read the results

   Start with the big picture, not individual trades. Ask four questions:

   • Did the strategy beat buy and hold on this pair and window?
   • What was the maximum drawdown? Could you have sat through it?
   • Does the equity curve look smooth or like one lucky streak?
   • Does the trade count match how active you expected the rules to be?

   A backtest does not prove future performance. It filters bad ideas before you risk capital.

6. 6

   Refine and run again

   Change one thing at a time: tighten the RSI threshold, add a trend filter, switch from daily to 4h, test ETH instead of BTC. Re-run and compare. Small edits are cheap in Torquant, which is the point of skipping the Python loop for early exploration.

Worked example: BTC daily RSI mean reversion

Here is a complete prompt you could paste into Torquant to test a simple long-only RSI idea on Bitcoin daily candles.

Torquant may confirm the date range or position size if anything is ambiguous. Once ready, run the backtest and check whether the equity curve and drawdown look acceptable compared to holding BTC over the same period.

Want a short idea instead? Try an ETH 4h SMA cross: “ETHUSDT 4h. Long when SMA 8 crosses above SMA 20. Exit on cross below. 50% equity per trade.” Same flow, different hypothesis.

What makes a good prompt

• Name the pair and timeframe. “BTC daily” is better than “Bitcoin strategy.”
• State entry and exit. A rule with no exit is not a complete strategy.
• Include indicator settings. RSI(14) beats “when RSI is low.”
• Say long, short, or both if direction matters to your idea.
• Mention position size when you care about risk, e.g. 10% of equity per trade.
• One idea per run. Test the core hypothesis first, then add filters in the next iteration.

Common mistakes

• Chasing perfect backtests. If you tweak until every curve looks amazing, you are probably overfitting.
• Ignoring drawdown. A strategy that returned well but dropped 60% may not be tradable for you.
• Testing one lucky window. Try a longer date range or a different pair to see if the edge holds up.
• Expecting perp realism on spot data. Funding and liquidation need a different tool chain.

Common questions

Related guides