Skip to main content
The HftBacktest Python API provides a comprehensive framework for backtesting high-frequency trading strategies with accurate latency modeling and order fill simulation.

​
Installation

Install HftBacktest using pip: 
pip install hftbacktest
Requires Python 3.11 or higher.

​
Quick Start

Here’s a minimal example of a market-making strategy: 
from numba import njit
import numpy as np
from hftbacktest import (
    BacktestAsset,
    HashMapMarketDepthBacktest,
    BUY, SELL, GTX, LIMIT
)

@njit
def market_making_algo(hbt):
    asset_no = 0
    tick_size = hbt.depth(asset_no).tick_size
    lot_size = hbt.depth(asset_no).lot_size
    
    # Time in nanoseconds
    while hbt.elapse(10_000_000) == 0:
        hbt.clear_inactive_orders(asset_no)
        
        depth = hbt.depth(asset_no)
        mid_price = (depth.best_bid + depth.best_ask) / 2.0
        
        # Submit buy and sell orders
        hbt.submit_buy_order(
            asset_no, 
            order_id=1, 
            price=depth.best_bid, 
            qty=lot_size, 
            time_in_force=GTX, 
            order_type=LIMIT, 
            wait=False
        )
        hbt.submit_sell_order(
            asset_no,
            order_id=2,
            price=depth.best_ask,
            qty=lot_size,
            time_in_force=GTX,
            order_type=LIMIT,
            wait=False
        )
    
    return True

# Setup backtest
asset = (
    BacktestAsset()
        .data('data.npz')
        .linear_asset(1.0)
        .intp_order_latency('latency.npz')
)

hbt = HashMapMarketDepthBacktest([asset])
market_making_algo(hbt)

​
Core Concepts

​
Backtesting Workflow

  1. Configure Assets: Set up one or more trading assets with market data and latency models
  2. Create Backtester: Initialize a backtester instance (HashMap or ROIVector)
  3. Implement Strategy: Write your strategy as a Numba JIT-compiled function
  4. Run Backtest: Execute the strategy through the backtester
  5. Analyze Results: Use the statistics module to evaluate performance

​
Market Depth Models

HftBacktest provides two market depth implementations:
  • HashMapMarketDepth: Uses a hashmap for efficient sparse order book storage
  • ROIVectorMarketDepth: Uses a vector for dense order book storage within a range of interest

​
Numba Integration

Strategies run in Numba JIT-compiled functions for maximum performance. All core API methods are compatible with Numba’s @njit decorator.

​
API Organization

The Python API is organized into the following modules:
  • Main Classes: Core backtesting classes and market depth interfaces
  • Types: Type definitions, constants, and data structures
  • Data: Data loading, validation, and preparation utilities
  • Stats: Performance metrics and statistical analysis

​
Time Units

By default, HftBacktest uses nanoseconds for all timestamps and durations. Ensure your data uses the same time unit, or adjust accordingly.

​
Order Management

​
Order IDs

Each order must have a unique ID. There should not be any existing order with the same ID on both local and exchange sides.

​
Order Status

Orders progress through states: NEW → FILLED / CANCELED / EXPIRED

​
Time-In-Force Options

  • GTC: Good ‘till cancel
  • GTX: Post-only (maker-only)
  • FOK: Fill or kill
  • IOC: Immediate or cancel

​
Asset Types

When configuring assets, specify the contract type:
  • Linear assets: Use .linear_asset(contract_size) for standard futures/spot
  • Inverse assets: Use .inverse_asset(contract_size) for inverse perpetuals

​
Next Steps

Build docs developers (and LLMs) love

Get started for freeTalk to us