Updated: Oct 17, 2020
by Mitesh Shah
The Currency of Cryptocurrency
Soho, Noho, Tribeca, Nolita… There’s no denying New York neighborhoods have character. Or that they tend to be populated by characters. The newest cool neighborhood? Dumbo; a portmanteau derived from ‘Down Under the Manhattan Bridge Overpass’. What was once an industrial area famous for factories and easy access to the East River, Dumbo is now the place to be in the five boroughs.
Companies eager to show they are relevant, techies who want the imprimatur of awesomeness; a diaspora of Wall Streeters and wannabe artists comingle every day at the dozens of bars, bodegas, and bistros that have moved into the area to chase the opportunities that always come with cool.
I stopped into a bistro for a bite the last time I was in Dumbo and discovered that while the proprietors accept credit cards, debit cards, PayPal, and Bitcoin — they do not accept cash. The waiter looked at the $50 I tried to hand him as if it was raw sewage. As the poet laureate of Rock and Roll once said, “The times, they are a-changin’.”
We are now living in an age when cryptocurrency has become just another form of currency — and in some places, apparently, it is even more readily accepted than cash. From the time the enigmatically pseudonymous Satoshi Nakamoto published his now infamous paper in 2008, Bitcoin: A Peer-to-Peer Electronic Cash System, and released the first Bitcoins into the wild the following year, we have seen the concept of cryptocurrency catch fire.
In the span of a decade, the total market cap of cryptocurrency soared from precisely $0 to reach a highpoint of an $819 billion market cap in January 2018. What has proven to be the most volatile commodity in history then proceeded to fall from that lofty height to (at the time of this writing) $139.7 billion; accounting for a staggering 80-percent-plus slump in just a year-and-a-half. In the course of that meteoric rise and equally sudden fall, scores of the most informed and shrewdest crypto-traders did more than merely a little well for themselves.
It is no small wonder why the volatility — and correspondent opportunity — of cryptocurrency has caused it to become one of the most exciting options now available to investors. As the newest and fastest-growing form of currency, crypto is, in nearly every way, indistinguishable from the currencies that have fed what has become the largest and fastest-growing market in history: the currency foreign exchange (Forex).
The Extraordinary Opportunities of Forex
In a seminal article that appeared in Sloan Management Review twenty-eight years ago, Abraham S. George and C. William Schroth made a compelling argument that corporations should take seriously the idea of Managing Foreign Exchange for Competitive Advantage (1991). Their readers responded — and, not coincidentally, the size of the Forex market has since exploded.
The Triennial Central Bank Survey of the Bank of International Settlements reports that over the past two decades the Forex markets have quintupled; from $1.1 trillion traded per day in 1995, to $5.4 trillion when the triennial survey was last conducted, in 2016. By comparison, all of the stock exchange activity worldwide is approximately $2.8 billion per day, or about 400 times smaller.
Traders have been drawn to Forex for a number of reasons. Chief among them, liquidity, availability (Forex trading is available 24 hours a day from 5:00 pm ET Sunday through 5:00 pm ET on Friday, including most U.S. holidays), the absence of transactional costs — and perhaps most importantly, perfect transparency. With the factors that drive currency fluctuation being so openly and readily available, it is not possible for any player in the market to possess material nonpublic information.
Unlike stocks, the factors that affect Forex movements are more finite and more readily definable. Inflation and interest rates; weather and war; recession, speculation, trade policies, political stability, and performance. Any perceptive investor need only watch the news to get a high-level idea of where currency rates are headed. The rub is knowing which of the many factors and proximate proxies — and in what combinations — serve as cues to which ways prices may be headed. The computational complexity of resolving the vast number of possible combinations is where Artificial Intelligence comes to the rescue.
The Rise of the Robots — Artificial Intelligence and Finance
True story: A fellow stormed into a Target store in Minnesota a few years ago to confront the manager. He had a fistful of coupons — and a look of rage on his face. “My daughter got this in the mail!” he told the stunned manager. “She’s still in high school, and you’re sending her coupons for baby clothes and cribs? Are you trying to encourage her to get pregnant?”
The poor manager was shocked and confused, but he took the mailer the fellow was thrusting at him and, sure enough, it was addressed to his daughter — and it contained ads for maternity clothing, nursery furniture, and a collection of pictures of adorable infants.
“I’m so sorry Sir!” The manager said, “I will look into this and will get back to you right away!”
Our hero, the store manager, then called corporate Target to complain. He recounts what happens and shares his exasperation with the analyst they connect him to. “Not only did you enrage this customer, you’re wasting thousands of dollars of our money!” he raged, “these circulars aren’t free — and this is nothing more than a waste!”
The analyst excused himself as he typed out a query before coming back to the phone. Confirming again the name and address of the recipient he told our manager friend, “She’s pregnant. Well, actually there’s an 87 percent chance that she’s pregnant. And her delivery date is sometime in late August.”
True to his word, the manager called the customer back. Before he could explain that their algorithm must have gone awry, however, the now much calmed father interrupted and said, “I had a talk with my daughter. It turns out there have been some activities in my house I haven’t been completely aware of. She’s due in August. I owe you an apology.”
Okay, so how did the Target analyst know? Well, it turns out that if a customer purchases 25 products — and she does so within a particular sequence, over a given span of time — there is a high probability she is pregnant. And no, it’s not any of the products you may think (or having not purchased any of the products you may think). It’s things like large bottles of unscented lotion. And supplements like calcium, magnesium, and zinc. It’s cotton balls, hand sanitizer, soap. Big purses, washcloths — and a bright blue rug.
Yes, but how did the Target analyst know? Ah, sorry. He knew — or, more accurately, the algorithm was able to inform the analyst so accurately because they cheated. They began not with prediction, but with retrodiction. They started at the end, and then worked their way back.
Target, you see, has a baby registry. And by analyzing the millions of prior purchases of tens of thousands of women, a sequence popped into place. Well, it didn’t so much “pop” as it became fodder for very smart Machine Learning algorithms that are able to find a pattern in otherwise discordant data. By starting with the outcome and working backwards, Machine Learning models can be developed that are able to generalize and when they see similar patterns manifest again, predict what will likely happen.
That is precisely the sort of thinking that undergirds much of Machine Learning; the ability to predict based on patterns. It is what enables Amazon and Netflix to make recommendations, and it enables Siri and Alexa to understand what you say. It is what is behind Uber’s self-driving cars — and it is what drives algorithmic trading.
By considering vast collections of data that are far beyond human comprehension, Machine Learning algorithms learn to spot patterns — and they get progressively better at predicting what should come next. Those Machine Learning models that Target used to predict pregnancy? Their accuracy is now up to 94% — and they can guess the delivery date more accurately than an OB/GYN.
The Robots of Wall Street rely on Machine Learning in very similar ways. Whether enabled by signals similar to those that are relied on by fundamental or technical traders (or both) they find patterns that precede upward and downward inflections. That lawsuit was settled for less than expected? The drone attack in Saudi is impacting oil production? Those are the sort of simple relationships you and I can easily see. But separating the signal from the noise — when those signals are not obvious, and the world presents a continuous cacophony… Now we are in the provenance of truly smart machines.
Cryptocurrency and AI
Artificial Intelligence (AI) is about making machines that can learn, think, and act. In that respect, and for all practical purposes, the goal of AI is to mimic the intelligence of smart human beings. And just like people, the process begins with data ingestion. While humans rely on their senses, AI relies on its sensors and other input mechanisms. Through a process known as Signal Processing, data is ingested from keyboards, counters, sensors, newspapers, websites, and conversations you have with your Alexa. That information is then translated into the binary language of zeros and ones.
Once data has been converted into a usable form, the real workhorse of AI kicks in. Machine Learning (ML) is what puts the intelligence in Artificial Intelligence. ML consists of a collection of algorithms for clustering, categorizing, connecting, and predicting — based on the data that is being assessed.
The responsibility for taking action on what Machine Learning learns is often then relegated to robots. Whether the types that resemble Robbie the Robot from SciFi fame, the Roomba that cleans your house, the Robotic Process Automation (RPA) used for mundane clerical tasks, or the Algorithmic Trading Bots used by quants that have taken over the securities markets.
Among the intelligentsia there is no longer talk of the rise of the robots; they have already risen. As the cyberpunk writer William Gibson has said, “The future is already here — it’s just not evenly distributed.”
Unsurprisingly, Forex has followed the trend toward using AI. Indeed, like all other securities and tradable assets, the volume, velocity, and variety of the variables — and the complexity of possibilities — makes instantaneous and perfect analysis impossible without the assistance of AI.
To be sure, there are those companies that are making dubious claims to offering insights derived from AI — but frauds have likely been a part of every market since markets have existed, But the obvious need for inclusion of AI, ML, and Artificial Neural Networks (a component of ML) has been known since the mid-1990s. It’s just taken until recently for the computational power to catch up with the algorithms to make the implementation of ML in trading practicable.
In 1995, Woon-Seng Gan and Kah-Hwa Ng published a paper titled Multivariate FOREX forecasting using artificial neural networks in which they showed the ability of artificial neural networks (ANN) to forecast the foreign exchange (FOREX) rates of major currencies. A 2004 article from Cambridge University by Michael Dempster and Vasco Leemans (An Automated FX Trading System Using Adaptive Reinforcement Learning) went a step further, contending AI would — inevitably and fundamentally — forever change foreign currency markets. Their prescience proved particularly profitable to early adopters who heeded their call. Fast forward to today, and it has nearly become investing malpractice to presume anyone can juggle all the relevant variables that need be considered in making major Forex investments. And even if any human were able to juggle all of those variables and disclose the invisible patterns, they could never do so with the extraordinary speed and accuracy of the smart machines that are now needed to take advantage of arbitrage and scalping opportunities.
As part of an annual contest, the Japanese media giant Nikkei recently used Artificial Intelligence to predict Dollar-Yen exchange rates. Using signals that included Natural Language Processing (NLP) to ingest vast amounts of data from the Nikkei’s own articles, trends, commodity prices, and market indicators, the platform predicted the value of USD/JPY over the course of a month with greater accuracy than the company’s best analyst. A similar study by I Know First found an AI system was able to predict correct values 77.75 percent of the time over a seven-day period; the predictive equivalent of climbing Everest in an hour.
This is something that will be discussed in future articles as a part of the series.
The complexity of the factors that contribute to market moves has transcended human perceptual and processing capabilities. Success in the trading arena now requires the smart use of smart machines — and in particular, the use of AI. Just as Artificial Intelligence has proven essential to currency trading, it is now only a matter of time before that same power is brought to bear in trading the newest currency available for exchange: digital currencies.
Mitesh Shah is the Founder and Chief Executive Officer of Omnia Markets, Inc. Mitesh extensively researched and studied the digital currency industry and blockchain technology, participated in leading industry meetings, conferences, and summits, and networks with dozens of industry leaders. Mitesh also provides regular podcasts on the trends and developments in the industry and communicates with the U.S. Congressional staff members on developing future regulatory framework for the industry.
Omnia Markets is based on the belief that Sir Francis Bacon was right when he said, “Knowledge is power.” To help investors better inform their investment decisions, we have developed the first database of cryptocurrencies available in the global market; a secure decentralized cryptocurrency platform for collecting, aggregating, and filtering information that is based on Blockchain technology.
Omnia Markets is committed to solving investors' greatest concerns by providing unparalleled insights into the cryptocurrency market and individual coin offerings through a proprietary suite of analytical tools backed by Artificial Intelligence, Machine Learning, and Predictive Algorithms.