Switching From Manual to Automated Crypto-Trading in 3 Steps

Before we start, rest assured I’m not advocating throwing year's worth of accumulated knowledge and experience down the drain to go out and find the latest hyper-hyped AI bot promising to take over your duties while you sleep.

That’s a fantasy.

That’s not the way it works.

No worries if you can’t code. There are solutions for non-coders and I’ll address that rather important matter head-on later in this piece. And, no, those solutions do not require learning to code.

Automating your trading has nothing to do with having to trust a black-boxed algorithm either. That would be crazy!

Now, if you'd argue you wouldn't trust your own bot, then I'll be quick to challenge you on that. Your bot will do whatever you tell her to do. And she will do it exactly as you tell her, with machine efficiency. So, why exactly wouldn't you trust her?

After all, we trust algorithms and all sorts of computer systems all the time, with things a lot more valuable than cash— like our lives.

Need an example? Does air-travel ring a bell?

Jetliner pilots have rather little understanding or control over how modern aircraft do everything they do to take to the sky and go from A to B—let alone us, passengers.

It's algorithms running the sky-borne show of thousands of aircraft dancing through the atmosphere every day. And we trust them.

It's not algorithms that make mistakes... it's programmers.

Industrial-grade systems and algorithms may be trusted because they go through lengthy testing, quality assurance, and certification processes before they go live to do their mission-critical jobs.

Something similar happens with broadly-used open-source software. There are so many eyeballs looking into the code that most significant bugs are weeded out on the first few iterations.

Now that we have established what trading bots are not, clarified a few of the popular misconceptions, and debunked some of the myths around trading algorithms, let's make a quick list of what your trading bot may do for you and the reasons why you would want to automate your trading...

Trading bots may work 24/7, make a thorough evaluation of the market every other second, identify programmed situations and conditions across different markets and exchanges, evaluating strategy rules and taking immediate action on every opportunity that arises.

When trading is automated, the decisions on when to enter and exit positions are made ahead of the moment in which the actions take place. This removes the anxiety, fear and greed factors that may cloud your judgment when making trading decisions on the spot, in realtime.

Trading manually entails making swift decisions on the fly, while being fully aware of the possibility that you may be making a mistake, with the associated monetary cost. If you don't move fast enough, you may miss the opportunity. But if you make rushed decisions, you may miss some crucial piece of information and make the wrong choice. This dilemma makes manual trading a stressful activity.

There's only so much you can do when trading manually. Your physical limitations—best-case scenario being a pair of hands, a pair of eyeballs and one brain—set a rigid constraint to your capacity to monitor different markets or work with multiple strategies. On the other hand, automated trading allows you to conceive, develop, test and put to work one strategy after the other, allowing to explore different markets, different tactics and uncorrelated strategies—a key to true diversification.

Physical health hazards associated with the sedentary lifestyle of traders who spend all day staring at a screen are well documented. When trading is automated, the time spent seating in front of a computer is drastically reduced. Also, automated trading saves you from the emotional roller-coaster produced by the ups and downs of alternating success and failure events. When trading is automated, you periodically evaluate the bot's work and focus mainly on the big picture and overall results. This makes you a more balanced person than you would be if having to deal with the emotions triggered by each trade.

Now that we have set the right expectations, let's go on and get you started on the path to automating your trading.

First thing is to do a little reprogramming of your brain so that you can grasp the key concepts behind the automation of your job as a trader.

You need to stop thinking of trading as the collection of decisions you make and actions you perform to enter and exit positions. Instead, you will think of trading as a two-step process: strategy design and strategy execution.

You will use your knowledge and experience to design your strategies, and you will leave the execution — the actual act of trading — to your algorithm.

This means that your new job is not going to be about staring into the screen, scanning the market and placing orders. Your new job is going to be about squeezing your brain for all the juice it can give. It'll be about extracting your trading wisdom and using it to create, tune and polish the strategies that you will feed your algorithm.

On the other hand, the algorithm’s job is mostly about acting on a certain set of rules.

What rules?

The ones you came up with during your learning process and trading career. You tweaked and polished those rules over and over through your experience trading manually, even if you never got to write them down.

Of course, you may still be polishing your skills and there’s always new stuff to learn and ideas to explore. As soon as you have switched to trading in an automated fashion, you will have plenty of time to upgrade and enhance your education, and come up with new rules to explore new opportunities.

Your rules are your body of knowledge; your toolbox.

You will think of your algorithm as your trading assistant. She will do whatever you tell her to do. And she will do it exactly as you tell her.

This means that you will need to think very carefully what you will instruct her to do!

Now, bear in mind that your trading assistant does not speak English. She speaks math and computer code. The kind of instructions she understands well is of the type resulting from conditional statements; for example:

If a certain situation arises, then you must take this action.

Think of a situation as a specific state of the market that can be described in mathematical terms through a set of conditions that need to be true. Put in other words, when all the specified conditions are true, then your assistant will assume that the market has entered into a specific situation in which she needs to take a certain action.

No need to go into further details at this point.

All you need to do now is to acknowledge that you have been following this same kind of logic throughout your trading career. You have been monitoring the market until certain situations arise, and have been taking certain actions in response to those situations.

When I say that you will need to squeeze your brain to extract trading knowledge, I mean that you will need to sit down and describe — as precisely as possible — how it is that you make your trading decisions.

It is those rules that you will feed your assistant.

To structure the set of rules that you will feed your algorithm, you will need to follow some sort of framework.

This is fundamental in all types of automation.

It is quite straightforward to automate a very specific process with a clear structure. However, it can be quite complex to automate an unknown or undefined process.

In my article Developing Your Own Trading System: A Step by Step Logical Guide, I propose a specific framework that you may use to structure all your trading strategies in the same manner.

This framework will put you a stone’s throw away from automating your trading!

I have to admit this is where things get a bit tricky.

Choosing a trading automation platform is a very personal thing, as the software needs to fulfill your personal needs, which certainly defer from the rest of the crowd.

Are you a tech-oriented person or your seven-year-old handles your phone better than you?

Can you take on a complex software installation? Can you manage the execution environment yourself?

Would you rather use a platform-as-a-service kind of solution, which will certainly require to trust a third party with your exchange API keys and/or funds?

Can you handle the learning curve of an elaborate product offering great flexibility? Do you prefer a simple, therefore constrained solution instead?

Are you a developer? Will you develop your algorithm?

As you may now figure, there is no one-size-fits-all solution for crypto-trading platforms.

To get you up and running as fast as possible, I'm going to share a bit of my research on some of the platforms I've looked into.

By all means, take my views with a grain of salt as they certainly are subjective and may even be biased, in particular, because I do have a horse in this race, which I'll tell you about later on.

It's one of the few long-lasting commercial platforms out there and likely one of the market leaders, although it's impossible to tell in such a young market with little credible public information.

They started in 2013 specializing in bitcoin trading bots. Since then, they have amassed experience, which shows in their product.

They offer configurable trading bots which you launch from an interface that runs on your machine. The bots run in your local machine too, so you will likely need to hire a Virtual Private Server from a cloud provider to get bots to run 24/7, in a stable environment.

Some of their plug-and-play bots are black-boxed, so you don't know what they are doing. To the best of my knowledge, there are no published records or metrics of past performance either.

However, their flagship product, Haasbot, does support an array of possible configurations that allows you to create strategies with some flexibility.

The company markets the product to accomplished traders with enough experience to handle all the concepts involved in technical analysis and trading in general.

If you don't mind paying 0.066 to 0.187 BTC for a beginner or advanced yearly license respectively, you should check them out.

The software offers several strategies that users may configure to a certain extent. It certainly is not an open platform in the sense that it does not allow enough flexibility to create all sorts of strategies.

In turn, it constrains the use to specific use cases, which leads me to believe that the platform is oriented to unorthodox or beginner traders that lack the knowledge or understanding to create their strategies from scratch.

The software is sold in different versions for a one-time fee ranging from 0.04 to 0.25 BTC.

It combines an advanced manual trading terminal with a few bots automating strategies which may be customized.

In terms of assisting manual trading, it allows for placing orders with trailing stops, target take profit, and a few other interesting features most exchanges do not offer.

In terms of actual automated trading, the platform offers different kinds of bots, each for specific market conditions (bull, bear and sideways). These bots need to be configured with a choice of signals and trading parameters.

Again, like with Gunbot, the automated trading features seem to be constrained to specific uses, not allowing flexibility to freely create strategies.

Access to the service is sold on a monthly license basis ranging from 37 to 75 USD.

Named after the infamous Oliver Stone's fictional character in the 1987 Hollywood blockbuster Wall Street who openly defended that "greed, for lack of a better word, is good", Gekko is an advanced trading bots platform oriented to developers.

It's been around since 2013 and is probably the most popular open-source solution available at the time of writing. It started as side-project of Mike van Rossum, and has since evolved into a community effort, and also a business.

The platform is rather hard to install and manage, requiring advanced technical skills from users. However, there is a hosted solution, GekkoPlus, run by the software maintainer, made available for up to 40 USD per month.

The platform offers many strategies that may be configured, but creating new strategies requires advanced coding skills. If you are into coding or don't mind the technical challenge, you may want to check their Github.

This is my horse in the race!

We started in 2017 as a group of friends who'd had enough of making money for banks. We decided our time would be better spent creating the technology required to automate our own trading instead.

Then more people started joining us and the Superalgos Project emerged: an open-source project developing and freely distributing trading tools while building a Collective Trading Intelligence.

We recently pre-released an early-stage (alpha) version of the Superalgos, a piece of free, open-source software designed to help traders develop and automate trading systems, even if they have zero coding skills.

You may create strategies from scratch and with great flexibility, following the framework described in the Developing Your Own Trading System: A Step by Step Logical Guide article, implemented as the Superalgos Protocol. You may also start with open-source strategies maintained by the community.

Strategies are created in a visual setting, feeding your rules with simple formulas and statements and analyzing the effects of each rule directly over the charts.

You may use the built-in simulation engine with historical and up-to-date market data for backtesting and paper-trading. Once happy with the output of the simulations, you may deploy your strategies as fully automated bots, and trade live.

Because all strategies are based on the same protocol, traders may seamlessly share and exchange strategies or even specific parts or components, enabling collaboration among groups of friends, partners, and even large communities—a first step in the direction of building a Collective Trading Intelligence. 

Take a look at the Superalgos Platform and join our Telegram to interact with developers and the rest of the community.