QuantSim

A trade simulator that calculates market impact and transaction costs with websockets and multithreading.

Dependencies (install on your system)

• Boost (system, thread)
• CMake
• Make
• C++17 compatible compiler
• (Optional, for Windows) MSYS2/MinGW for Unix build tools

Environment variables setup

Create a .env file in the project root with the following 6 variables for your websocket server and exchange details:

Network configuration

HOST=ws://example_endpoint
PORT=8080
PATH=/api/data

Simulator configuration

EXCHANGE=OKX # currently only OKX is supported (for fee calculations), if you want your own exchange you can append logic in the FeeModel.cpp file map structure
SYMBOL=BTC-USDT-SWAP
INITIAL_CAPITAL=100000.0

WebSocket JSON Message Format

The WebSocket server you're connecting to should send messages in the following JSON format:

{
  "timestamp": "2024-01-01T12:00:00Z", // ISO8601 string
  "asks": [ ["price1", "quantity1"], ["price2", "quantity2"], ... ],
  "bids": [ ["price1", "quantity1"], ["price2", "quantity2"], ... ]
}

• asks and bids are arrays of arrays, each containing price and quantity as strings.
• timestamp should be in ISO8601 format.

Mathematical Models Used

Market Impact Model (Almgren-Chriss Model)

The market impact of a trade is taken as the sum of temporary and permanent impact:

Temporary Impact = temporaryImpactFactor × sqrt(|orderSize| / dailyVolume) × currentPrice
Permanent Impact = permanentImpactFactor × (|orderSize| / dailyVolume) × currentPrice
Market Impact = Temporary Impact + Permanent Impact

• temporaryImpactFactor and permanentImpactFactor are configurable parameters.
• orderSize is the size of the trade.
• dailyVolume is the estimated daily traded volume for the asset.
• currentPrice is the current mid price of the asset.

Slippage Model (Quantile Regression)

Slippage is estimated using historical price and volume data, with quantile regression to estimate volatility:

slippage = currentPrice × volatility × sqrt(|orderSize| / avgVolume)

• volatility is estimated as a quantile (e.g., 95th percentile) of historical returns (historical being from streamed data onwards).
• avgVolume is the average historical trade volume.
• orderSize is the size of the trade.
• currentPrice is the current mid price.

Fee Model

Fees are calculated based on the exchange, fee tier, and whether the order is a maker or taker:

fee = orderSize × price × feeRate

• feeRate is determined by the exchange and fee tier (only OKX maker/taker fees for now).
• orderSize is the size of the trade.
• price is the execution price.

Running it yourself

To compile, execute the following commands in the root folder (For Windows use MINGW/MSYS2), ensure cmake and make are installed:

./build.sh
cd build

And run the executable.

Authored by: Don Chacko donisepic30@gmail.com