Backtest JS: Framework

A comprehensive and user-friendly framework to fetch candle data, backtest any trading strategy and compare results.

Enhance your trading strategies with Backtest, meticulously crafted for trading developers. Leverage the power of TypeScript to backtest your strategies with unmatched precision, efficiency, and flexibility.

Key Features 🌟

Intuitive Methods: Utilizes intuitive methods for smooth and efficient operation.
Comprehensive Candle Data: Access historical candle data from Binance or effortlessly import your own CSV files.
Integrated Storage: Efficiently store your candle data, strategies, and results in the internal SQLite storage.
Documentation: Maximize Backtest’s capabilities with thorough guides and resources.

Quick Start

Installation

To install the package in your project, use the following npm command:

npm install @backtest/framework

How to use this package

You can incorporate this framework directly into your project by installing it as described above.

Alternatively, you can clone the quick-start repository, which will allow you to start writing your strategies without needing to set up a project from scratch. The project itself provides all the necessary instructions.

If not, you can use the command line interface that will handle everything for you. In this case, we recommend checking out the specific project @backtest/command-line. This way, you can easily navigate and use the command line interface without any confusion.

Documentation

If you enjoy reading the code, you can find a comprehensive (and extensive) demonstration method here that showcases most of the available methods. Additionally, you'll see examples of how to run strategies (with or without support, for instance).

In this README, you will find a comprehensive table that lists and describes all the methods available within the framework.

As on overview, some of the areas covered by these methods are:

Historical data operations

Finding historical data by name.
Deleting historical data.
Downloading historical data for specific intervals and time periods.
Exporting historical data to a CSV file.
Importing historical data from a CSV file.

Strategy operations

Scanning for available strategies.
Finding all available strategies and their names.
Running a strategy with specified parameters and historical data.
Parsing the results of running a strategy.
Saving the results of running a strategy.

Results operations

Finding result names.
Finding all results.
Deleting a result.
Saving a result.

Historical Candle Data

Easily download candle data from Binance, no coding or API key required (thanks to Binance!). Alternatively, you can import historical data from a CSV file. Additionally, you can export your data to a CSV file for further analysis.

Custom Strategies

In addition to the demonstration strategies already present, you can create your own by adding a file under src/strategies.

Use one of the existing files or the examples in this guide as a reference. Each file should contain a runStrategy method, and if it uses external or dynamic parameters, it should also include a properly filled-out properties structure.

Whenever you create a new strategy, modify the properties structure of an existing one, or delete an existing strategy, you need to run the scanStrategies method.

There’s no need to stop or restart the backtest process if it’s running, or to exit the program. The program will reload the contents of your file with each launch, as long as it’s synchronized.

Using well-defined or dynamic parameters (instead of constants within your strategy) will allow you to run multiple tests simultaneously.

Candle Data

Each candle have the following information available:

export interface Candle {
  symbol: string
  interval: string
  openTime: number
  open: number
  high: number
  low: number
  close: number
  volume: number
  closeTime: number
  assetVolume: number
  numberOfTrades: number
}

Buy and Sell

It is possible to execute a buy or sell by following this:

export interface BuySell {
  price?: number // Price of which you will trade the asset
  position?: string // Can be "short" or "long" (long is the default)
  amount?: number | string // Amount of asset to buy can be number or string (string must include a %), f.e. 300 or "10%"
  baseAmount?: number // Trade the base amount to use for percentage calculations (total worth for baseAmount equals to amount)
  stopLoss?: number // Price of which a stop loss will trigger (all will be sold on the stop loss)
  takeProfit?: number // Price of which a take profit will trigger (all will be sold on the take profit price)
  percentFee?: number
  percentSlippage?: number //
  note?: string // Add a simple note to identify this trade
}

Pay attention: follow these rules:

You CAN short and long at the same time but they need to be two seperate calls
If you try to buy or sell but you already bought or sold everything, the buy or sell will be skipped and not recorded
You cannot use “amount” and “baseAmount” params together
If in a short and a long you cannot use “amount” or “baseAmount” when selling without specifying a position.
You cannot use stopLoss if you long and short at the same time
You cannot use takeProfit if you long and short at the same time
Amount param can be a number or a string, if a string it must contain a percent sign “%”

In particular, the buy signal:

bth.buy()

/* or */
await bth.buy({
  position: 'short', // or 'long' (default)
  amount: '10%', // or baseAmount
  note: 'a simple note here',
  stopLoss: stopLoss,
  percentSlippage: percentSlippage,
  percentFee: percentFee
})

while the sell signal:

bth.sell()

/* or */

await bth.sell({
  amount: 250, // or baseAmount
  note: 'a simple note here'
})

Examples: Buy and Sell

Beginner: The simplest buy & sell

// Lets say you have $1000 and want to trade bitcoin
// Put in a long order and buy all which is $1000 worth of bitcoin
await buy()
// Lets say you bought bitcoin and are now worth $1000
// Put in a sell order and sell all which is $1000 worth of bitcoin
await sell()

Beginner: How to specify amount

// Lets say you have $1000 and want to trade bitcoin
// Put in a long order of $400 worth of bitcoin
await buy({ amount: 400 })
// Same thing can be achieved here
await buy({ amount: '40%' })
// Lets say you bought bitcoin and are now worth $1000 in bitcoin and put in a sell order of $400 worth of bitcoin
await sell({ amount: 400 })
// Same thing can be achieved here
await sell({ amount: '40%' })

Regular: How to specify stop loss and take profit

// Lets say you have $1000 and want to trade bitcoin
// Put a short order in with all which is $1000 and a stop loss at $24,000
await buy({ position: "short", stopLoss: 24000 })
// The application is smart enough to know that it's a short and only sell if a candles high goes above $24,000
// Lets say you bought bitcoin in a long and a short but only want to sell some of the shorted amount
// Put in a sell order to sell 50% of the shorted amount
await sell({ position: "short", amount "50%"})

Regular: How to specify base amount

// Lets say you have $1000 and bitcoin is currently worth $2000
// Put a long order in of .25 bitcoin which is $500 worth
await buy({ baseAmount: 0.25 })
// This can also be achieved by doing
await buy({ amount: 500 })
// You cannot use amount with baseAmount in the same buy / sell call
// Lets say you bought bitcoin and are worth $1000 and bitcoin is worth $2000
// Put a short order in of .25 bitcoin which is $500 worth
await sell({ baseAmount: 0.25 })
// This can also be achieved by doing
await sell({ amount: 500 })

Advanced: How to place an order at a specific price

// Lets say you have $1000 and bitcoins close was $2100 but you had a trigger to buy at $2000
// Put a long order in of $1000 worth but bitcoin at a price of $2000
await buy({ price: 2000 })
// Lets say you bought and bitcoin is worth $2200 but you had a trigger to sell at $2100
// Put a sell order in where bitcoin is worth $2100
await sell({ price: 2100 })

Write a Strategy

When a strategy is executed, the runStrategy method has access to the BTH object, which contains useful information. For example, it provides methods (like: getCandles) to obtain ohlc data, calculate technical indicators, and manage trading positions.

BTH and getCandles

Below is the interface:

export interface BTH {
  tradingInterval: string // Trading interval
  tradingCandle: boolean // Indicates if this candle is tradable
  currentCandle: Candle // Current candle
  params: LooseObject // Strategy parameters
  orderBook: OrderBook // Order book
  allOrders: Order[] // All current orders
  buy: Function // Function to buy (long / short)
  sell: Function // Function to sell
  getCandles: Function // Function to obtain price data (see below)
}

The getCandles function is an asynchronous function that returns an array of Candle objects.

Parameters:

type: Specifies the type of data to return (a key of Candle or 'candle' to return the entire Candle object).
start: Indicates the starting index from which to begin retrieving the candles.
end (optional): Indicates the ending index up to which to retrieve the candles. If not specified, the method uses only the start.

The getCandles method can be used as follows:

const closes = await bth.getCandles('close', 10, 0) // last ten closes
const open = await bth.getCandles('open', 0) // last open
const candles = await bth.getCandles('candle', 5, 0) // last five candles

Details on the start and end parameters:

If end is not specified, the method will return the candle at index candleIndex - start.
If end is specified, the method will return the candles from index candleIndex - end to index candleIndex - start.

Examples:

getCandles('close', 5) will return the close at index candleIndex - 5.
getCandles('open', 10, 5) will return the opens from index candleIndex - 10 to index candleIndex - 5.
getCandles('candle', 10) will return only the 10th candle counted from the last candle (candleIndex - 10).
getCandles('candle', 10, 0) will return the last 10 candles.
getCandles('candle', 10, 5) will return candles from candleIndex - 10 (inclusive) to candleIndex - 5 (exclusive).
getCandles('candle', 10, 1) will return candles from candleIndex - 10 (inclusive) to candleIndex - 1 (exclusive, i.e., excluding the last one).

How to run strategies

When you want to execute a strategy, you need to call the runStrategy method. Remember to perform a scanStrategies if, for example, you have changed parameters or created a new strategy.

import { scanStrategies, runStrategy } from '@backtest/framework'

const scan = await scanStrategies()
console.log('Scan strategies:', scan)

const runStrategyResult = await runStrategy({
  strategyName: 'demo', // ./strategies/demo.ts
  historicalData: ['BTCEUR-1d'],
  params: {
    lowSMA: 10,
    highSMA: 50
  },
  startingAmount: 1000,
  startTime: startTime,
  endTime: endTime
})
console.log('runStrategyResult:', runStrategyResult.name)

When you run your strategy, you can provide multiple parameters. Below is the general structure:

export interface RunStrategy {
  strategyName: string // name of the strategy to run
  historicalData: string[] // symbols to use for trading (e.g. ['BTCEUR-8h', 'BTCEUR-1d'])
  supportHistoricalData?: string[] // symbols to use as support (e.g. ['BTCEUR-1h', 'BTCEUR-8h', 'BTCEUR-1d'])
  startingAmount: number // how much money to start with
  startTime: number // from which date start to evaluate yor strategy
  endTime: number // to which date evaluate your strategy
  params: LooseObject // parameters to use for the strategy, you can pass multiple value for each parameter
  percentFee?: number // 0.1 means 0.1% fee
  percentSlippage?: number // 0.6 means 0.6% slippage
  rootPath?: string // sometimes is useful specify a different path (uncommon case)
  alwaysFreshLoad?: boolean // if true the file of the strategy is always reloaded by scratch, the default is false
}

Pay attention: If alwaysFreshLoad is set to true, it's important to note that you cannot use global variables in your strategy. As a result, you won't be able to take advantage of the benefits of using support historical data.

Examples: Strategies

Beginner: The simplest strategy

Below is an example of a simple 3 over 45 SMA strategy. You buy once the 3 crosses the 45 and sell otherwise. In this example, we don’t use the power of params.

import { BTH } from '@backtest/framework'
import { indicatorSMA } from '../indicators/moving-averages'

export async function runStrategy(bth: BTH) {
  const lowSMACandles = await bth.getCandles('close', 0, 3)
  const highSMACandles = await bth.getCandles('close', 0, 45)

  // Calculate low and high SMA
  const lowSMA = await indicatorSMA(lowSMACandles, 3)
  const highSMA = await indicatorSMA(highSMACandles, 45)

  // Buy if lowSMA crosses over the highSMA
  if (lowSMA > highSMA) {
    await bth.buy()
  }

  // Sell if lowSMA crosses under the highSMA
  else {
    await bth.sell()
  }
}

Pay attention: hard-coded parameters will prevent you from running multiple tests simultaneously!

Regular: the same strategy with parameters

Below is an example of a simple SMA strategy like above but it’s not hard-coded to the 3 over 45. When you run the strategy through the CLI, you will be asked to provide a low and high SMA. You can even provide multiple lows and multiple highs, and all the variations will be tested in one run.

import { BTH } from '../core/interfaces'
import { indicatorSMA } from '../indicators/moving-averages'

export const properties = {
  params: ['lowSMA', 'highSMA'],
  dynamicParams: false
}

export async function runStrategy(bth: BTH) {
  const lowSMAInput = bth.params.lowSMA
  const highSMAInput = bth.params.highSMA

  // Get last candles
  const lowSMACandles = await bth.getCandles('close', 0, lowSMAInput)
  const highSMACandles = await bth.getCandles('close', 0, highSMAInput)

  // Calculate low and high SMA
  const lowSMA = await indicatorSMA(lowSMACandles, lowSMAInput)
  const highSMA = await indicatorSMA(highSMACandles, highSMAInput)

  // Buy if lowSMA crosses over the highSMA
  if (lowSMA > highSMA) {
    await bth.buy()
  }

  // Sell if lowSMA crosses under the highSMA
  else {
    await bth.sell()
  }
}

Advanced: use of multiple historical data

Your strategy can also use other intervals as a support, that is one or more intervals of the same symbol. This way, on the trading interval, you can execute buy/sell actions, while you can use the supports to perform statistics or validate any trading signals.

export async function runStrategy(bth: BTH) {
  if (bth.tradingCandle) {
    // For example, use BTCEUR-1d data to execute your trading strategy (buy, sell, etc.)
  } else {
    // do something else when BTCEUR-8h candle is closed
  }
}

Backtesting Results

Backtest not only delivers performance insights but also returns your strategy's effectiveness through comprehensive statistics.

Import Candle Data from CSV

Although there is an option to download data from binance for crypto assets there is no automatic download available for traditional symbols such as apple or tesla stock as well as forex symbols such as usdyen.

This candle data can be downloaded from third party sites such as yahoo finance and can then be easily imported to the Backtest database to use with any strategy.

How to prepare CSV file

The CSV file must have the following fields:

Close time of the candle: closeTime or date
Open price of the candle: open
High price of the candle: high
Low price of the candle: low
Close price of the candle: close

The CSV file can have the following optional fields:

Open time of the candle: openTime, openTime
Volume in the candle: volume
Asset volume of the candle: assetVolume
Number of trades done in the candle: numberOfTrades

Pay attention: follow these rules:

Each field can be written without considering case sensitivity.
The order of the fields in the CSV file is not important.
Any additional fields will not cause an error but won't be added to the database.

API Documentation

The following table outlines the primary methods available within this framework.

Method	Description
deleteHistoricalData	Deletes historical data of a symbol and interval
deleteMultiResult	Deletes the saved result of a multi-symbol execution
deleteResult	Deletes the saved result of an execution
downloadHistoricalData	Downloads historical data of a symbol and interval from Binance
exportFileCSV	Exports historical data of a symbol and interval to a CSV file
findHistoricalData	Returns the historical data of a symbol and interval
findHistoricalDataNames	Returns the names of the saved historical data
findHistoricalDataSets	Returns all saved historical data
findMultiResultNames	Returns the names of the saved multi-symbol execution results
findMultiResults	Returns the saved multi-symbol execution results
findResultNames	Returns the names of the saved execution results
findResults	Returns the saved execution results
findStrategies	Returns the strategies saved in the database
findStrategy	Returns the strategy by its name
findStrategyNames	Returns the names of the strategies saved in the database
getCandleStartDate	Returns the date of the first candle (1m) through Binance
getCandles	Returns the candles of a symbol and interval
getIntervals	Static list of usable intervals
getMultiResult	Returns the saved result of a multi-symbol execution
getResult	Returns the saved result of an execution
importFileCSV	Imports historical data from a CSV file
isValidInterval	Checks if an interval is valid (among those from getIntervals)
parseRunResultsStats	Processes the results and returns an object with the statistics
runStrategy	Runs a single strategy, multi-symbol with or without supporting intervals
saveMultiResult	Saves the result of the previously executed strategy
saveResult	Saves the result of the previously executed strategy
scanStrategies	Rereads and updates the list of strategies and associated parameters

Prisma: Useful commands

Prisma is a modern DB toolkit to query, migrate and model your database.

In this project, Prisma is used with SQLite to avoid the need for installing other databases. If necessary, the database file can be deleted or updated.

Below are some useful commands to run from the terminal/shell:

Run npx prisma to display the command line help;
Run npx prisma validate to validate the prisma.schema;
Run npx prisma generate to generate artifacts, such as the Prisma client;
Run npx prisma db push to push the Prisma schema state to the database.

However, it's always recommended to refer to the official Prisma documentation for detailed information.

@backtest/framework