It's Here! The Winning Crypto Trading Bot Is Real!

My trading background goes back to 2010 when I first started learning, trading and writing strategies for Forex. I must admit, I've never made it there so after more than a year of trying, I gave up.

10 years later, a friend of mine introduced me to crypto trading, who later became my partner. Being fully involved in another startup at that time, I hesitated. Somehow I never thought that trading crypto would be worthwhile. I didn't trust it, until one sleepless night, mid-August 2021, when finally decided to sit and write a strategy on top of Freqtrade.

You know what fired me up that night?! A backtest that showed BILLIONS in profit with only $1K as a starting balance in 1 year. I couldn't stop writing, testing, backtesting, hyperopting to increase the profits even more.

Freqtrade backtesting result

In 2 weeks time I had my production version ready for dry testing. Luckily, I have another friend who's very good at DevOps. We dragged him in as a partner and within one night we had the cluster up and running with all the deployment pipelines ready. We ran it for a day or two, made sure everything is OK so we decided to go live.

We've squeezed all our pockets, wallets and credit cards and started the live mode with about $5K. It went well, but after a few days, some of the pairs started hitting stop losses so our balance started the downward trend. I'm happy we didn't give up but rather continued improving by adding new indicators, hyperopting, testing different timeframes, stake pairs, everything. I've lost the count how many times we had to flush the database and start from scratch.

Months passed, we kept making and losing money but we continued working until we've got our first client willing to invest. He started with $7K but our strategies were far from perfect. We knew that if the market drops fast, we'll lose it. And we did!

\"It’s not how many times you fall but how many times you rise\" -- they say!

See that red "icicle" on December 4th? That's the big market crash that took a whooping $1.7K out of our client's balance. It was painful but even more painful was that the same strategy trading on our own balance lost ~ $12K on that day, (sorry, don't have a screenshot to prove it). We knew it will happen one day (and I'm glad it did) if we don't do something to protect ourselves from such events.

We started working on our "saviour" merging feature weeks before the "big crash", which would detect pairs that fall too close to the stop loss mark and merge with other trades on that pair to average out the profit. In fact we had it ready but our DevOps guy had a Padel championship in Poland so he took a 3 or 4 days off so here we are. The market crashed the day before he came back. I remember waking up that morning and checking the telegram channels. I couldn't do anything but laugh within myself. Not because it was amusing but because it was so ironic and paradoxal, I couldn't resist.

Guess what happened next ... reset, restart and start from scratch. We became experts at this at some point.

Now we had our merging feature in place. We've also done fundamental improvements to the Freqtrade core to allow more than 5 open trades per pair to bypassing the limit imposed by binance.com. We've improved our strategy to calculate and consider different indicators within the same bot instance. It was far from perfect but now we had a very solid base to build our next features on.

Since Dec 4th until today, the strategy performed very well, except for Jan 10th when we lost ~1% because of a cluster upgrade issue we had with DigitalOcean. Because the cluster was down for about 1 day, the bot couldn't do the merging properly so we had some trades hit the stop loss limit.

Since Dec 28th until Jan 31st, the profits are almost flat because of the Bitcoin price was in a down trend, which started recovering about 1 week ago. Bellow is the profit totals for the same time period.

Profit, Won and Loss In Percentage

Profit, Won and Loss in USD

18.5% in 70 days, keeping in mind the market conditions during this period isn't that bad right?!

On Dec 11 we've onboarded another client who started with $20K on the same strategy as ours, except the Supertrend indicator. We wanted to test both variations, that's why we didn't include it. As you can see, it didn't lose anything on Jan 10 because the instance came back up quicker so it managed to merge trades which got close to the stop loss limit.

Profit, Won and Loss in Percentage

Profit, Won and Loss in USD

This one had 20% in 63 days which happened during the same market downtrend conditions.

You can see some losing trades across the chart but these are not actual losing trades but rather some miss-calculations that happened because of a bug in Freqtrade which recorded the profit as negative, though on binance.com it was closed in profit. We've been able to fix that bug lately so it happens rarer now.

Some of you might ask, well, where are the BILLIONS that backtesting showed me in the beginning?! The answer is: there are many aspects to this. Overall the market was in a downtrend for the last 3, 6 and even 9 months. Additionally, we chose not to trade *UP and *DOWN pairs because they are less predictable. Third, backtesting differs a lot from live market conditions because a backtest treats every order as successful while live, the market dictates and not every order places is completed. But we're happy with the result so far, and that's all that matters.

In conclusion, for those trading or looking to trade on crypto, I think that doing this manually requires huge amounts to make it worthwhile. If you're looking to trade with a $5K to $50K balance then having a bot is a must. Personally, I would recommend everyone trade using a bot because it is way more reliable, has a higher frequency and is a better decision maker for short to mid-term investments. Not to mention the fact that you have way more free time to do the stuff you love and don't bother checking your orders every 5 minutes.

So what's next?! Fist, fine-tune both merge and split features for higher cashflow. Second, futures! The moment we add futures support to our strategy then it won't matter if we're in up or downtrend. Isn't this a dream for any trader?!

At last, for those who trade, good luck and never ever give it up. For those who don't, you must! I don't think crypto is going to disappear anytime soon. There is a lot of risk but also huge potential and if you feel the need for help then you know where to find my "Chat" button.

Stay cool!