Robotics and AI in Trading

Talking about robotics and AI might cause images from The Matrix to flash through your mind, where some computer program has outcompeted everyone around it. Just the idea that there is development to make a machine that can think and act better than a human implies that humans will become unnecessary, and thus phased out.

In real life, there is a difference between bots and AI and this article seeks to explore both robotic automation and AI as it relates to trading, and give a primer on these technological advancements.

Bots in Crypto

As much as 60-80% of trading volume on some exchanges comes from trading bots. Big players with money invest in high frequency traders (HFT) whose prime function is to notice inefficiencies in the markets, and execute trades measured in microseconds. Most of the trading bots in crypto are used for arbitrage, which is why prices stay relatively correlated amongst all the various exchanges and currency pairs. 

There are some bots that are for rent, and there is a long list of online companies that offer them. They let subscribers build their own bot using open-sourced code. While businesses claim that their bots are ‘easy to use and are beginner-friendly,’ they often have a long list of bad reviews that complain of long response times and poor support, as well as negligible returns.

What happens is, new traders get excited by the promise of passive income, and rush into a monthly subscription, only to find that the bot is not as easy to operate as promised. The irony here is that the only ones making passive income are the businesses offering the bots.

That is, except in cases where individuals who are tech-savvy enough to download the open-sourced code and program trading strategies for the bot to execute. These people can see returns of 1-15% each day.

Bots can be a useful tool to help traders, but there is a lot of know-how the trader needs to have beforehand. Namely, programming as well as trading.

Something else to be aware of is the multi-level marketing, where returns are mostly generated from referrals and not any trading the bot actually does.

Bot Manipulation

Bots can also be used to manipulate the crypto markets. We’ve seen countless examples of spoofing, wash trading, stop order hunting as well as pump and dumps on illiquid markets. Manipulators don’t necessarily need bots to do this but automating the process makes it easier in a 24/7 market. 

Another thing to be aware of is that bots can cause flash crashes. If you place a lot of open limit orders, then you probably never noticed. But if you trade margin (especially with leverage) then maybe you’ve gotten liquidated from rapid price fluctuations. This sort of thing is not isolate to the crypto markets, as trading bots have even caused hundreds of millions, even trillions, of dollars of loss in traditional financial markets. 

If you know how to program bots, you can take advantage of that crash and profit (that even helps with market liquidity, by the way). But if you’re trading with leverage, you’ll likely get rekt.

Now, automation is not all bad news. It can also help to provide liquidity to markets. Many market makers that fill the order books use algorithmic trading. 

Systems of Artificial Intelligence

Now, none of the above is news. Bots have been in the financial markets for years before they came to crypto. What is new is the use of artificial intelligence in trading.

What happens with AI is that it is able to analyze massive amounts of market information and make its own predictions on what will happen in the market. While bots are more like tools to help a trader to be more efficient, AI doesn’t need to have a human next to it at all; it can be its own independent trader.

Moreover, “artificially intelligent systems can analyze large amounts of data at speed and improve themselves through such analysis.” So as time goes on, they’re only going to get better. They ‘read’ information online, make mistakes, learn from their mistakes, and then improve.

Currently, the returns on these AI systems are not that impressive (roughly 2%), and seem to be isolated to hedge funds, who are hiring multiple kinds of researchers and developers to teach the AI to learn faster. While AI has been in development for a long time, its use is still relatively new.

As of yet, these AI systems have not started to trade cryptocurrencies. But, there are many new ICOs that are looking for funding to make improvements to their AI. Some of these projects want to train AI in pattern recognition, but need billions of data sets to make progress, so they offer to pay people in crypto to answer simple binary questions. Other projects seek to amalgamate data sets on a blockchain, making them available for purchase for anyone.

This article won’t give any specific project names – you can do your own research. The point here is that people will eventually be phased out from certain tasks. By 2025, it’s estimated that more than 100,000 jobs will be lost to AI in the banking sector alone and are going to make life more difficult for traders.

AI and automation are here to stay. There’s no possible way to outcompete them since they are designed to be better than the experts. The only thing to do now is watch them, learn from them, and understand them so that you don’t become redundant. That, and make a ton of money so if you do get outcompeted, it won’t matter that much to you anyway.

Dominic Rose

Dominic Rose

Dominic first entered the crypto space in 2015, and is now looking to make trading a successful career. His current interests involve writing on the subjects of blockchain and artificial intelligence. Outside of crypto, he is a serial entrepreneur.
Dominic Rose

Latest posts by Dominic Rose (see all)

Facebooktwittergoogle_plusredditpinterestlinkedinmail