Freqtrade

Freqtrade is a free and open source crypto trading bot written in Python.

This repository is a collection of strategies, configurations, dry-run and backtest results I collected overtime.

Disclaimer

I am not a financial advisor and I am not responsible for any financial loss you might incur using this repository.

Quickstart

This will install freqtrade and dependencies to work with this repository tools. Or you can just ignore this and look through the repository.

# Clone the repository
git clone https://github.com/TheoBrigitte/freqtrade.git
cd freqtrade
scripts/setup.sh

Update freqtrade

Freqtrade can later be updated with the following command

scripts/setup.sh -u

Repository structure

• backtest_results/ - Contains my backtest results, good or bad, they are all here
• config/ - Contains a bunch of configurations I found, some I used, some I didn't
• scripts/ - Contains some scripts I use to work with freqtrade and this repository
• pairlist/ - Contains a collection of pairlists
• strategies/ - Contains all strategies I found and played with. Folder names are arbitrary and each folder might contain additional configurations for the specific strategy(ies)
• sources/ - Contains sources of strategies I found. They mainly are git submodules to other repositories

How I use this repository

Freqtrade virtual environment should be loaded first

source freqtrade/.venv/bin/activate

Update pairlist and download data

scripts/fw generate-static-pairlist ./pairlist/pairlist-volume-usdt.json
mv pairlist.json pairlist/new_pairlist.json
ln -fsrv pairlist/new_pairlist.json pairlist/pairlist_futures.json
scripts/fw download-data --timeframes 1m 3m 5m 15m 30m 1h 4h 12h 1d --timerange 20241201-20250101

Run tests

I usually run lookahead-analysis, backtesting and plot profit in order to see how a strategy performs.

fw lookahead-analysis strategies/BinHV45/BinHV45.py --timerange 20240101-20250101
fw backtesting strategies/BinHV45/ --timerange 20240101-20250101
fw plot-profit backtest_results/backtest-result-2025-01-01_01-01-01.json

Based on my observations I might tweak the strategy or the configuration and run the tests again. Configuration tweaking for this strategy would go in strategies/BinHV45/config.json.

Navigate backtest results

I usually use the following command to navigate through backtest results

scripts/backtest-browser.sh

Backtest results interpretation

Many aspects comes into play when it comes to evaluating a strategy. Backtest results are the first step I use to evaluate a strategy. Once I found something that looks promising I will run a dry-run over couple of weeks and ideally months to see how it performs in near real conditions.

In a nutshell those are the most important metrics I look at:

• Expectancy: The average amount you can expect to win (or lose) per trade with this strategy. Above 0.2 is usually considered good.
• Profit factor: The ratio of gross profit to gross loss. Above 2 is usually considered good.
• Avg profit: The average profit per trade. Above 1% is usually considered good.
• Drawdown: The maximum drawdown of the strategy. This depends on your risk tolerance, but below 10% is usually considered good.

This is a very high level overview and there are many other metrics and aspects to consider.

Here is a good knowledge base to learn from: https://botacademy.ddns.net

Credits

• https://github.com/freqtrade/freqtrade