Cryptocurrencies are no longer just a niche concept — they’re a driving force in today’s economy. What started as an experiment has grown into a worldwide phenomenon, influencing the way people think about financial systems and opening new investment possibilities that could affect the future. 

Dr. David Enke, interim associate dean of the Kummer College and Curators’ Distinguished Teaching Professor of Engineering Management and Systems Engineering at Missouri S&T, focuses his research on using artificial intelligence (AI) to forecast the price and volatility of traded assets and securities, with a particular interest in Bitcoin.  

What exactly are cryptocurrencies and blockchains? 

A cryptocurrency is a digital currency that uses cryptography for security. Unlike the United States dollar, or other fiat currencies issued by governments, cryptocurrencies are typically decentralized with all transactions verified and recorded on a “blockchain.” 

The blockchain is a decentralized and distributed digital ledger that records transactions across multiple computers, preventing the ledger from being altered by a centralized authority. The distributed blockchain ledger is essentially what prevents counterfeiting, copying or double spending the digital cryptocurrency. 

What makes Bitcoin unique? 

Bitcoin was designed as an alternative to fiat currencies, serving as a store of value and medium of exchange. Some investors see it as “digital gold” for maintaining value and purchasing power over time. Unlike a government printing more of their currency, thereby devaluing its worth, the amount of new Bitcoin mined or produced is designed to decrease over time, even becoming more scarce than physical gold. Only 21 million Bitcoins will ever be mined, with over 19.8 million already in circulation. Other cryptocurrencies such as Ethereum and Solana have different approaches for growth and validating their blockchain ledger that can give them more functionality and speed of transaction, but potentially at the cost of security and being a good store of value. 

How is Bitcoin affecting investing and the economy right now? 

As a result of being scarce and limited, Bitcoin can provide protection or a hedge against inflation and thereby serve as a store of value. This is particularly important for people living in countries with high inflation or unstable currencies. Since Bitcoin is decentralized, individuals have more control over their financial transactions without relying on a third party, such as a financial institution or broker.  

This is of importance to those who are unbanked or without access to traditional financial services. Being decentralized, Bitcoin allows transfers across country borders without the need to pay high fees to banks or others providing remittance services, allowing for quick money transfers at lower transaction cost. 

With the introduction of new Bitcoin Exchange Traded Funds (ETFs), Bitcoin is also growing as a new asset class that can help diversify existing portfolios beyond traditional assets like stocks, bonds, commodities and real estate. Since 2011, Bitcoin has been the top performing asset in 11 of the last 14 years when compared to broad range of common asset classes and stock indexes. 

But it’s not all positive. In those three years in which Bitcoin was not the top performing asset class, it was actually the worst due to its high level of volatility. While this might not be an asset you would want to “go all in” with, it’s increasingly being considered as an asset that could be part of a well-diversified portfolio. Some companies and countries are even using or considering Bitcoin as a potential reserve asset to help diversify their balance sheets beyond existing fiat currencies. The United States government also just established a Strategic Bitcoin Reserve. 

How might Bitcoin and other cryptocurrencies affect the economy in the future? 

Bitcoin and cryptocurrencies have been instrumental in the rise of decentralized finance, which is leading to new ways to lend and borrow money, trade assets, value non-fungible tokens (NFTs) and support innovation in financial technology, also known as “fintech.”  

For instance, while NFTs have been used to place value and authenticate ownership of digital assets, they can also use smart contract technology, or code embedded in the NFT, allowing an original artist or musician to continue to earn a percentage of all future asset sales. For example, if a previously sold painting of an artist gains value and is later sold again at a higher price, the original artist could get a percent of this new value.  

Cryptocurrencies and NFTs could affect royalty rights, ticket prices to sporting events, the selling of stocks and bonds or even digital land in the Metaverse. The potential to transfer existing financial transactions and processes onto the blockchain is tremendous.  

How do regulators currently treat cryptocurrencies, especially Bitcoin? 

The classification of cryptocurrencies by U.S. regulators is complex and varies significantly based on the specific cryptocurrency and regulator. For instance, the Commodity Futures Trading Commission (CFTC) has declared Bitcoin a commodity, treating it like other commodity assets such as gold, crude oil or corn. For tax purposes, the Internal Revenue Service (IRS) treats Bitcoin as property rather than a currency, meaning transactions involving Bitcoin are subject to capital gains taxes, similar to other assets like stocks.  

The Securities and Exchange Commission (SEC) has stated that Bitcoin does not meet the criteria to be labeled a security, such as stocks and bonds, and therefore does not fall under their regulation. On the other hand, the SEC does believe that many other stated cryptocurrencies, initial coin offerings and tokens do resemble securities and therefore do fall under their jurisdiction. With the new administration recently coming to office that is believed to be more supportive of cryptocurrencies, the regulatory environment in this space is expected to undergo change and will be updated for this new asset class. 

You mentioned that you are using AI to predict asset prices. What types of models and data do you use to make forecasts? 

Most of our current work for forecasting the price of traded assets involves the use of deep learning artificial neural networks for making price predictions. Artificial neural networks are models that mimic the way the human brain learns, but just like for humans, these networks need input data to be trained and learn.  

For traditional stock price forecasting this usually involves past stock prices, past trading volume and other data about the range or momentum of price movements over a certain timeframe. By providing years of past daily data as input to the artificial neural network, along with what the price was for the next day as an output, the artificial neural network will start to learn a relationship between input and output values. Of course, with computers, there is the old adage of “garbage in, garbage out,” so it’s important to make sure one is providing data that will help the model provide an accurate prediction. 

How do you know what data is best for training your models? 

This is where experience of the domain can be helpful but also limiting. For instance, as someone who studies stocks and investing, I might think I have a good idea of what is important for driving stock prices, so I will use this data to train my model. But as it turns out, some of the data that I thought was important might not be that useful.  

Once trained, the model might find patterns in the data or use the data in a different way than typically used by a human investor. This is where proper feature selection comes into play.  

In addition to using artificial neural networks for price forecasting, we can also use them along with evolutionary algorithms, statistics and classification models for selecting the data with the most predictive information. This increases the performance of the forecasting model since it is now using information that will help it make a better prediction, instead of trying to use less useful data that we thought was important but is in fact just confusing the model. 

Is there anything unique about forecasting the price of Bitcoin compared to a stock or stock index? 

There is, and this is what has made trying to predict Bitcoin prices so interesting. For traditional stocks, research has tended to focus on using better prediction models, or ways to make the models more efficient by focusing on the key inputs, most of which are well known. With Bitcoin, we have an entire new dataset we can work with – the blockchain ledger that records all Bitcoin transactions.  

In addition to just specifying the time and how much Bitcoin was traded at a particular price, we also know information about what electronic addresses traded with each other, even if the people behind those addresses are anonymous. With this data, we can begin to learn more about those coins that are actively traded, or had previously been dormant, such as a long-term holder who is now selling.  

We know how much of the supply is currently in active trading circulation over various time frames, and whether the coins came from traders with small positions or those with much larger positions, often called “whales.” We know how many unique addresses have big or small positions. We can determine the average size of a transaction over a certain time period. We know when coins are taken out of “cold storage” and placed on exchanges where they are more likely to be traded and not held. We also know how many people have positions that are in profit, as opposed to being at a loss, each of which might impact whether they are more likely to buy or sell. 

All of this new data gives us a fresh way to look at trading and provides our models with new training information. 

Are your forecasting results improving with all this new data? 

Yes. Using traditional data such as price, trading volume, price ranges and momentum tended to produce Bitcoin price forecasting results that were similar to other asset classes in terms of accuracy. By using blockchain data and the wealth of information that can be derived from this source, the Bitcoin price forecasting results have improved significantly. However, there’s always room for improvement, which provides for an interesting and challenging area of research. 

What are students currently working on in your research lab? 

Within the Laboratory for Investment and Financial Engineering (LIFE), students are working on a number of projects. In addition to using machine learning and on-chain data for forecasting the price of Bitcoin, they are also developing price forecasting models that use technical indicators that identify positive or negative price momentum, as well as overbought or oversold conditions in stocks and cryptocurrencies.  

One recently graduated doctoral student also used AI to analyze specific price pattern images to see if they can be used to identify upcoming bullish or bearish price moves. Another student is using machine learning to forecast asset price volatility to not only help make a better price prediction, but also provide a way to manage financial risk, an area of importance to the company where he works. Finally, one student is working on a project for the Department of Transportation (DOT) that is investigating whether AI can be used to conduct a better life-cycle cost analysis of highway pavement maintenance, hopefully allowing for timely and cost-effective maintenance decisions. 

What advice do you have for investors? 

If you’re thinking about buying a cryptocurrency, do your research and be careful. There are numerous cryptocurrencies that are available for purchase, with new ones being created every day. Most will end up worthless, and even those that have a real use case and seem to have staying power will be volatile in price. In fact, purchasing most cryptocurrencies is more akin to gambling than investing. However, some will survive as investments and will no doubt have an impact on how we think about the way finance, investing and commerce is conducted in the future. 

As of this publication date, the price of Bitcoin is around $83,000. 

About Missouri University of Science and Technology   

Missouri University of Science and Technology (Missouri S&T) is a STEM-focused research university of over 7,000 students located in Rolla, Missouri. Part of the four-campus University of Missouri System, Missouri S&T offers over 100 degrees in 40 areas of study and is among the nation’s top public universities for salary impact, according to the Wall Street Journal. For more information about Missouri S&T, visit www.mst.edu.