Trading backtester

A highly customizable framework designed for parallel tuning of trading algorithms by reproducing and simulating the trading history of exchanges and the behaviour of brokers.

It aims to provide a complete set of interfaces needed to simulate and optimize trading and investing algorithms at any level of precision: from the level of intra-exchange messages to the level of price chart trading.

This project is committed to achieving:

• The highest possible execution speed, which is available through AOT compilation that runs directly to native code, low runtime, and great rustc and LLVM optimization abilities.

• Low probability of making critical errors, which is achieved by Rust's strong type system and borrowing rules that prevent the vast majority of erroneous programs from compiling.

• High speed of writing custom code. This goal is achieved through the relatively simple syntax of the Rust language, which makes it no more complicated than that of C#.

Usage

Put this in your Cargo.toml:

[dependencies]
trading_backtester = { path = "???", features = ["???"] }

where path should point to the location of the trading_backtester library, and features should consist of the available ones (or may not be set).

Features

The following features are available for enabling. Each of them provides access to:

• concrete

  Concrete examples of entities that implement traits from the interface module.

• enum_def

  The macro that generates an enum that can contain each of the listed types as a unique enum variant. Simplifies the creation of statically dispatched trait objects.

• enum_dispatch

  Derive macros for statically dispatched trait objects from the interface module. Convenient to use with the enum_def.

• multithread

  Utilities for running backtesters in multiple threads.

Overview

General workflow

Kernel

Kernel message management system