Implement this simple BTC-USD trend following algorithm today, using Coinbase Pro API and python
Orange: BTC-USD hourly close; blue: strategy equity /

Here is a simple algorithm you can implement immediately for BTC-USD, using the Coinbase Pro API and python. This is just a jumping-off point (but includes ideas for improvement in the future).

Clone repo here:

Here’s what to do to get started

Step 1.) git clone

Step 2.) Open and replace credentials for api_key, passphrase, and secret from your Coinbase Pro account. Make sure your API key is authorized for view and trades.

Step 3.) Run python

Here’s how it works

The algo is enabled by thecbpro library to make API calls to Coinbase Pro. cbpro is an unofficial client and is under development.

We’ve modified the function place_stop_order() from AuthenticatedClient to include stop_price, in case you want to place stop orders.

The strategy lives in and is very simple, clocking in at 100 lines.

The class History initializes PublicClient and calls get_product_historic_rates() to get the last 100 hourly bars. It returns either True or False (whether the 50 period average > 100 period average).

The class Account initializes AuthenticatedClient (for getting account information and placing trades). It can check your balances (BTC or USD) and place orders (buy or sell).

The function run() initializes Account, then checks if it is the top of the hour with

If so, we call History().signal, requesting the hourly data and returning a signal.

auth_client.is_balanceUSD() checks if there is USD available to buy. Vice versa if it is a sell.

After each hourly check, we call time.sleep((60*60) —*60 — ( to sleep until the top of the next hour (including microseconds, so we don’t sleep longer and longer each loop).

Potential improvements

This algo is a very quick way to get involved today. However it is not perfectly optimized. Here are some ideas for improvement:

cron: Using cron could make the algo cleaner and more robust to termination.

Managing data: Instead of calling historic rates once per hour, we could call 100 hours up front, then call just the last hour every time we want to check for a signal. We would pop the oldest hour, append the latest hour, and check for our signal. This would save time calling the Coinbase API for 100 hours instead of 1 hour.

Don’t check balance: Ignore checking the balance… just send the trade request. If you don’t have a balance (either of USD or BTC), the trade doesn’t go through anyway.

Interested in more ready-to-implement code, backtests, and algorithmic trading ideas? Subscribe to samchaaa++ here.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Sam Chakerian

Python, trading, data viz. Get access to samchaaa++ for ready-to-implement algorithms and quantitative studies: