The Backtest Engine

The engine is chosen by what you backtest, not by a setting. Open a strategy on the Backtesting page (Automation → Backtesting) and it goes to the matching engine:

Coded strategies (Python, TypeScript, JavaScript, C++, SQL) only decide what position to hold. A shared simulator turns those decisions into fills, costs and an equity curve. The builder DSL and QuaScript engines simulate their own orders, because stops, targets and brackets live inside them. Either way, the statistics, analytics and Monte Carlo are computed once, by the same code.

• Fills happen at the next bar’s open by default, so a decision made on a bar’s close is never filled at that close (no lookahead). You can choose the signal bar’s close instead.
• Costs: percent of notional, per unit, per order, Indian statutory schedules (STT, stamp duty, exchange and SEBI charges, GST) or crypto maker / taker fees.
• Slippage: basis points, ticks, or square-root market impact that grows with your share of the bar’s volume.
• Delay, lot size and a volume-participation cap model latency, contract sizes and liquidity.

The same SMA crossover written in five languages produces the same trades, bar for bar. The engine’s test suite checks this on every change.

Pick the broker and symbol on the Backtesting page and the venue follows from it. Through the API, pass exchange (or broker), symbol, interval and a date range. A bar that has not closed yet is never included, and up to 500,000 bars fit in one run.

Return a target position for every bar: signed units (10 = long ten, −10 = short ten, 0 = flat), or a fraction of equity with sizing: "fraction". Signal lists are accepted too.

from quatick import ta

def strategy(bars, params):
    fast = ta.ema(bars.close, params.get("fast", 12))
    slow = ta.ema(bars.close, params.get("slow", 26))
    return [0 if ta.is_na(s) else (1 if f > s else -1) for f, s in zip(fast, slow)]

export function calculate(bars: Array<{ time: number; close: number }>, params: { n?: number }) {
  const n = params.n ?? 20, out = [];
  for (let i = n; i < bars.length; i++) {
    const mean = bars.slice(i - n, i).reduce((s, b) => s + b.close, 0) / n;
    if (bars[i - 1].close <= mean && bars[i].close > mean) out.push({ time: bars[i].time, type: 'LONG_ENTRY' });
    if (bars[i - 1].close >= mean && bars[i].close < mean) out.push({ time: bars[i].time, type: 'LONG_EXIT' });
  }
  return out;
}

Every route is POST /backtest/engine/<route> with your Quatick token, and costs one backtest credit (refunded if the job fails). check is free.

{
  "data": { "kind": "historical", "exchange": "BINANCE", "symbol": "BTCUSDT", "interval": "1h",
            "from": "2025-01-01", "to": "2025-06-30" },
  "strategy": { "language": "python", "files": [{ "path": "main.py", "content": "..." }] },
  "execution": { "capital": 100000, "costs": { "kind": "percent", "value": 0.02 },
                 "slippage": { "kind": "bps", "value": 1 } }
}