Profit from artificial intelligence. The story of Cristian Păuna

Initially published by wall-street.ro
Click here to read the Romanian version.

During the most severe stock market crash in history, a simple post on a professional social network tells us that someone is making a profit using artificial intelligence programs. Finding out that everything happens in Romania, we became curious to find out who is behind this activity and how such a thing is possible. From previous documentation on the internet, we found out about the existence of Cristian Păuna, engineer, economist and computer scientist, the one who made a software that automatically makes investments on the capital market. He is also the author of some unique mathematical methods included in this software. Moreover, we find that the system is the basis of a professional service that is offered to investors who can increase their capital in their own investment accounts, using these programs. As the story became unique and interesting, I invited Cristian Păuna to an interview about money, mathematics, programs, stock exchange and profit:

Are you the author of automated software that makes a profit on the stock market?
Yes, I am the author of automated software that makes a profit on the stock market. Very concisely, you said it all in one sentence. Thank you very much! We can stop here.The rest are details that don't even matter anymore... I am kidding. In fact, I know that details matter the most in this area. Yes.I made a profitable software. It is a dream come true. From a young age, I dreamed of making a money machine. But who doesn't dream of it? I have been working on this project for twenty-two years. In 2010 it became functional; since 2011 it is on the market, and starting with 2019 it is also in Romania, a fact for which I am very proud.

How did this story begin?
I was born in Bucharest, in a family of artists. My parents, both music teachers, guided me on the same path, but the attraction to technology and commerce was much higher, native I would say, probably taken from grandparents. I discovered money around the age of seven or eight. I have sold various things in the neighborhood, candies, badges, later pictures, or pages from various magazines and all sorts of other rare stuff at that time. That's when I found out how profitable it is to sell when more people want to buy. I remember every day the enthusiasm of the child with his fists full of coins earned from negotiations with other children. It was fantastic! I think that's how I started.

Was money important?
Money is an instrument. Like any tool, money is important only when you don't have it, and you need it. In the beginning, the money had only one purpose for me, that of being able to buy a movie ticket at the Favorite Cinema on Drumul Taberei neighborhood. I went there for a long time without telling my parents. The ticket cost two lei and forty bucks, and I didn't have it every week. After a while, I managed to make a profit from the stuff merchandise. I was collecting the surplus in coins, in a milk bottle. I never managed to fill the bottle, because there were always various other needs, but the usefulness of money and the need to obtain it have become a constant activity since then. I have always considered that the money itself is not important, but the problems it solves.

You said you were dreaming of a money machine.
As a child, I imagined that I would have to invent a device that would collect a coin from each one, similar to the model of the balance with coins from the past.But I wanted the device to offer something new and exciting, so that people do not regret, to return alone and pay again. I had heard that there are millions of people in the world, and a simple calculation gave me a ticket to the cinema for life.But when I found out that whatever I invented would automatically become the property of the communist state in which I lived, I was very disappointed. Since then, I have probably a big problem with injustice.

I understood that you liked math.
At the age of twelve, I became convinced that mathematics never lies.As everyone had disappointed me until then, I became a great enthusiast of the exact sciences. As a teenager, I studied aviation high school in Băneasa.There I was lucky to have Mr. Florin Vulpescu-Jalea as a math teacher, a dedicated professor, of great talent. He didn't give me too high notes and always told me "you can do a lot more." In fact, I think that only once I got the legendary “bravo maăăi!”, the encouragement you received, with a very serious and pressing tone, when you had a really special idea. He is the one who showed me the practical usefulness of mathematics. I use daily notions and subtleties in-depth and long-discussed with him, and I still find special ideas among the notes taken at that time.

When did you discover the computer?
Passionate about technology, I took an exam at the Aircraft Faculty of the Polytechnic.It was 1990 when I met the computer, which later became my best friend. I think I was lucky again. In the university, I had access to the latest technology of those times. Although then I did not realize what I was doing, now I can praise myself that in 1991 I was already optimizing computational models for aircraft engines, and in 1992 I sent for review the first articles in the magazine of Applied Mechanics of the Romanian Academy, papers that were accepted and published in 1993. I was twenty years old, and I had a million unapplied ideas. On the aircraft faculty, I have studied special mathematics and numerical methods with Professor Corneliu Berbente. I always remember how in his classes, I was thinking about how to use those methods to make money instead of airplane engines.I was poor; I was living from a tiny student scholarship and math meditations. That's when I started applying the Fisher transformation to the dollar exchange rate, and I found that the model worked quite accurately.

When did you start producing software?
In 1992 I managed to buy a 386 computer at home. It was a luxury at that time and a great advantage for me because I didn't have to go to work on the faculty's computers anymore, where I couldn't run programs that weren't about engineering. In the spring of 1993, I managed the first cyclicity simulations on the exchange rate. Also, in 1993, I took an exam at ASE, and I started doing Cybernetics in parallel. Here I got acquainted with the database systems and with the specific methods of realization of the economic and management programs, notions that were missing in the Polytechnic. During a legendary assembler exam at ASE with Professor Ion Gheorghe Roșca, he convinced me that "I have a call" to make software. He is the third teacher who has changed my life.We didn't work together, but it was very convincing. Thinking of his words and encouraged by the enthusiasm I feel every day when I code, in 1998, I canceled all my university activities, and I have opened my own software company with the money from my last university scholarship.

How did you start, with what project?
In April 1998, I wrote the first lines of code for SuperCont, the business intelligence system I developed every day. I had it in my mind for a few years, and it was born very quickly. It was from the beginning a centralized, real-time system, specialized for various fields of activity and has since then reporting and event notification automatic procedures. In particular, the software was designed for collecting and analyzing financial data and stock quotes, the way I have continuously tested and improved the mathematical models I use today. At that time, stock market data was a scarce and expensive resource. For a long time, I have used the free quotes available on the internet, but they were with 60 minutes delayed. However, the data could be used to optimize and test the models I was working to develop. I have always been convinced that any price evolution can be mathematically modeled and therefore, can be predicted at least for a short period of time. After many years of work, I also got the confirmation. theServer, today's integrated investment software, it is the result of a huge number of attempts and failures to automate this activity.

When did the investment system become operational?
In 2010 I managed to implement the first fully automated algorithmic trading software.It's called DaxTrader, and it is still one of the most profitable I've ever made. In the meantime, another 49 programs have been added, all are integrated in theServer. Using special artificial intelligence procedures, each of these programs identifies opportunities in several capital markets, automatically builds and transmits trading signals in the brokerage informatics system. He turns them into stock exchange orders, executes them, thus making a profit. Everything happens simultaneously in several capital accounts with which the system is connected from an IT point of view. This is how I made the money machine, the one I dreamed of as a child.

So is it profitable? Can money be made on the stock market?
Of course. All you have to do is do what most people don't do, measure the risk well, be patient and, most importantly, know when to stay aside. I use mathematical algorithms for all this. Artificial intelligence is much smarter than me, and anyone who thinks he can do with his head what a few servers do together. The system I have built performs today, on average, 6.6 million instructions per second. With this volume of calculation, the software has made 72% profit in the last twelve months in my partners' accounts. It means an average of 6% per month, a net profit of 3% for the investor, which is a performance.

It is, therefore, a very complex system.
Yes. As I said, differences bring profit. My software is not a simple trading program, as you can find a lot to buy on the internet. theServer includes today a number of 142 advanced mathematical models that identify entries in 17 capital markets. There are now a total of 2006 procedures looking each second to find market entry opportunities. Once an investment is initiated, another 454 algorithms look for the most suitable exit from the market, meaning the most profitable closing of each position. The system includes algorithms optimized for short, medium, and long-term trading, and has integrated automatic risk management and capital exposure optimization module.

What's the secret? Where does the money come from?
Obviously, there are several factors, I wouldn't necessarily tell them secrets. First of all, I use a particular method of risk management that allows the precise automation of all these processes. Then, the data-mining algorithms used to identify opportunities in the capital markets are based on mathematical methods that I have developed starting from optimization models known in engineering, and that I have adapted and optimized specifically for each financial market. They are new, original, and very profitable methods as it was proved. I have already published some of these methods in academic papers, on the occasion of my doctorate, which I am now completing. Some models have already attracted the attention of the university world, being awarded in several international conferences. Indicators like Price Cyclicality, Volume Cyclicality, or Price Probability Predictor are functions that work wonders, and you won't find them in any trading platform. All these allow us to build limit conditions that exclude high-risk entries, indicate exits before the price to changes its direction, or tell us exactly when it is better to stay away and wait. Many of these models are presented in detail on the site pauna.biz.

There are also losses, I assume.
Of course. Losses are part of our lives. It's all about knowing how to lose, how to manage a loss, and when it's better to accept a smaller loss not to make a bigger one in the future. There is no perfect trading model. The unexpected news and events ruin the continuity of mathematics from time to time. Today, my software makes approximately 85% profitable trades and 15% negative positions. The big difference between these proportions allows me to close a negative position only after other positive transactions have already covered that loss. Thus the capital balance does not decrease. During these periods, we do not make a profit, but we do not lose money, we only waste time. I remind you that software is an investment system, which gives us the time to resolve all negative positions without being constrained by any particular rush. This is, in fact, one of the peculiarities that essentially contributes to profitability, obviously organized together with a judicious risk allocation.

So when there is a crisis do you just waste time?
From my point of view, crises are long-awaited opportunities. In the crisis, the share price drops significantly, usually unforeseen, and often completely out of control. However, if the capital exposure remains within the acceptable limit, and if you know how to identify the time or range of minimum prices, you have the chance to buy with minimal risk. In such periods many are scared, or they have to sell because they have exceeded their risk limits. In any crisis, it is unlikely only the time when the price will rise, not the fact that this will happen. In a well-organized investment activity, patience brings a substantial reward. Today we are after the biggest stock market crash in history, caused by the panic generated by the well-known Coronavirus pandemic. There was an unprecedented stock market crash, in which, in about 30 days, the quotations decreased between 25% and 50%. Making classic investments, those with nominal shares lost even 50% of the capital, a sad reality brought only by the bad investment orientation. In my case, in the first period of this crisis, the software covered some of the negative positions, and in the second period, it made a profit. Two months after the crisis broke out, we were on profit in all the accounts operated by the system, and in the large accounts, the profit made in that short period had already exceeded 5%.

What do you call large accounts? Does account size matter?
The system is designed to operate at full capacity for accounts with over one million dollars or euros. It is a system aimed at professional investors and was specially designed for this purpose. The software also works for smaller amounts, obviously with lower efficiency, because not all available algorithms are used. The minimum amount of capital for which the service is still operational is one hundred thousand dollars plus the used risk. The value of invested capital is very important, of course. The higher the profit rate, the higher the capital exposure, even if the capital exposure rate is the same. The explanation comes from the fact that in the case of a larger amount, the coverage of negative positions is done much faster, the capital being thus less time stuck in unprofitable investments.

How much is the risk? What does an investor risk using this software?
The investor can set a minimum risk level of 10% of the capital. Once the amount of capital invested increases, the risk can also increase to 20% or 30%. An additional share of 10% of the capital, or 20% for professional accounts, is used for negative position recovery operations. So the risk allocated can be between 20% and 50% of the capital, depending on the amount invested and the option of each investor. It must be said that the risk is not a loss. Being engaged in an investment, and having the time necessary to settle each unprofitable position, the risk materializes in a capital exposure as the sum of open positions at any time, positions that are closed in time, one by one, profitably or at zero. The only situation in which capital exposure can turn into a loss is when the investor wants to close all positions immediately and to withdraw the capital without having the patience to smartly close each investment. This is possible at any time but entails a decrease in investment efficiency. However, after the profit accumulated in the account becomes higher than the assigned risk, the investor no longer risks anything from the initial subscribed capital, thus completely eliminating the stress.

Where is the capital? How does the investment work?
The most significant advantage is that the money is permanently in the investor's account opened at an independent brokerage house. Only he can deposit or withdraw capital from his account, being the only holder. The brokerage house is an authorized and supervised financial institution that manages the capital, and it is responsible for the account security. The investor purchases separately an IT service, based on theServer software, which identifies opportunities in different capital markets and transmits the corresponding signals in the brokerage house's IT system, in correspondence with the investor's account. With the explicit approval of the investor, the broker automatically takes these signals, turns them into stock exchange orders, executes them, and deposit the resulting profit into the investor's account. The entire activity can be tracked in real-time both on the broker's interface and through SuperCont, the real-time business intelligence system with which theServer is integrated. For the service provided, the investor has to pay an invoice periodically, in case a profit has been made. He can withdraw part of the obtained profit for the service payment, or he can pay the invoice from other sources. In this investment, the capital is permanently liquid, deposited in the investor's account. Moreover, the investor has at his disposal the profit made at any time and not at the end of the activity, as is the case with other types of investments. The activity can be stopped at any time, unconditionally, if the investor is not satisfied with the results. Our service contract has no restrictions and can be terminated by the beneficiary at any time without penalty. Thus, the capital is not blocked for a certain period of time and is permanently available to the investor whenever he wants.

With which markets does the system work? Does it also work on the Bucharest Stock Exchange?
The system operates 24 hours, 5 days a week, on the most liquid international capital markets. Frankfurt, London, Zurich, Paris, New York, Sydney, Tokyo are some of the connected markets.Unfortunately, the Bucharest Stock Exchange is not attractive, from my point of view. In addition, we largely depend on the technical capabilities offered by brokerage houses. My system trades contracts for differences, these being today the only one financial instrument that offers the possibility to perform long-term hedging operations with low commissions, and that allows the automation of the activity with a time response of the order of milliseconds.80% of the investments are made in the stock indices of the main mentioned stock exchanges; the remaining 20% are investments in currencies, gold, and German treasury bonds.

Are you the only author of this software?
Yes, this software is my accumulated work for the last twenty-two years. I am both the author of the software, and all the mathematical methods implemented in these programs.I am the sole owner of the copyright and related rights.At the moment, SuperCont and theServer are being developed by the Romanian company Algorithm Invest (algoinvest.biz).

So you are on the Romanian market?
Yes, from 2019, I am also in Romania. theServer was originally produced and developed in Germany, where it has been subjected to the rigors of the German investor for years. Still, most partners are foreign investors, but for reasons related to the high level of taxes, I have decided to move the software company in Romania, even the profile of the Romanian investors is not necessarily an advantage. The service generated by the software I produce can be accessed through several subsidiary companies, some of which are also represented in Romania. I don't want to give a name, so as not to offend anyone. You can find them on the internet.

Do you have a software, and the investor buys a service?
Yes, my software is not for sale. It is leased to several companies that have the right to use the soft for the benefit of their customers. These are investors who open their own capital accounts at independent brokerage houses, investors who access our investment service to generate profit in those accounts. The cost of the service is part of the profit made. There are no hidden costs, fees, or other expenses. If the service provided does not generate a profit, the investor has nothing to pay. As our interest is to have something to invoice, we thus ensure the profitability of the entire investment.

How much is the profit made, and how much does the investor have to pay?
The efficiency obtained in the last twelve months in the professional accounts is 72%, gross profit. The cost of the service is usually half of the realized profit. Thus, the investor has an annual profit of 36%, a net result superior to many businesses, achieved without any effort, and with limited and controlled risk. If the investor is a Romanian company paying VAT, it can deduct the VAT included in the service invoices, thus increasing the profitability by 9.5%. The above figures are for professional accounts of over one million. For smaller amounts of capital, the efficiency decreases accordingly but still remains at attractive rates. A minimum account has achieved an efficiency of 48% in the last twelve months, so it offered the investor an annual profit rate of 24%, meaning an average of 2% per month. All these figures include the tax that each party involved has to pay to the state, a tax that can be lower if the investor is a firm. Please note that these figures are not estimates; they are profit values ​​already realized in the partners' accounts.

Are there differences between Romanian and foreign investors?
Yes, there are very big differences. A German or Swiss investor must first cross the psychological threshold generated by the fact that the service is based on software developed by a Romanian person and developed by a Romanian company. Believe it or not, this aspect is very important to them. After solving this mental block that really exists, the foreign investor is a mature investor, who knows that he has to take a risk, who appreciates that the risk is measurable and controllable, and who is very satisfied that he has the chance to achieve profit without doing practically anything, except to follow of the activity. The Romanian investor is, first of all, scared that someone wants to cheat him. After it is clarified that the money remains in his account and no one can touch it with anything, the Romanian investor is the one who, as a rule, wants a huge profit with a tiny risk, zero if possible, which is practically impossible to be offered. I must say that our financial education has great shortcomings, even among those with a lot of money. Personally, I am still very surprised by the nonchalance with which the so-called "investors" propose to you that you take the risk, and they have only to collect the profit. Not infrequently, I encounter such situations that lead, of course, to the impossibility of signing a contract. However, I also had very pleasant surprises. We are on the Romanian market for a year, and we already have contracts signed for several million euros with Romanian investors, whom I thank a lot for their trust! I think we have a very good start on the Romanian market.

Can you tell us who your customers are?
I would really like to be able to tell you the names of my partners because some of them have famous names, especially those from abroad. However, I cannot do this because all the contracts we sign are strictly confidential. Confidentiality and the fact that we treat the relationship with each partner personally, are important aspects of this activity. However, it is possible to find out about our partners as soon as possible. Some of them, satisfied with the results obtained in their accounts, started to recommend us to others, which practically leads to the declassification of the existence of some contracts.

Can the system be tested?
Yes, of course, but the system does not work with demo accounts. The service can be tested on a real capital account with a minimum amount of one hundred thousand dollars plus the assigned risk. The test can take as long as the investor needs to clarify the efficiency obtained, before making plans for a serious investment together. Testing the system does not imply that we voluntarily produce a profit on the investor's account. In this case, also, the cost of the service is half of the profit made during the test. I want to remind you again that the money is in the investor's account, which excludes from the beginning any problem related to the trust.

Are you making a bigger profit than known investment funds?
Hmmm, some mutual funds came to buy the software. It's just that I don't have any software for sale. The only way an investor can benefit from what I have achieved is by purchasing the IT services generated by this software, services sold through subsidiary companies. Looking back, in percentage terms, in the last five years, we have really made a higher profit than many investment funds. If you make this comparison today, when many funds report losses, our figures may become even harder to believe. But do you know what is wonderful? The fact that we call the profit made in the partners' accounts, so for each case, there is at least one person who can confirm the profit. There is another big difference between our activity and that of investment funds. In order to participate in a financial investment fund, you must send your capital to its account, while we make a profit directly in your account, an account to which only you, the investor, have access and control. If you look closely, this is the main advantage, certainly more important than the value of the profit itself.

You have achieved something special. Where do you see yourself in ten years?
Today I am exactly where I saw myself ten years ago, and in ten years, I see myself also here. The software is profitable but not perfect. There is still a lot to do. I work daily to add new algorithms in order to increase efficiency. A year ago, I was at 53%, and two years ago, I achieved an annual profitability of 48%. I'm sure "I can do a lot more," as my favorite teacher always told me.

Another challenge is the optimization of all these algorithms. To adapt the soft to the volatility and market behavior changes, computational models are optimized weekly by an automatic machine learning server. Optimizing the optimization server today is challenging. Calculating the optimal value for a number of about 20,000 coefficients involves a complexity that you can't even think about without frowning. Simplifying these procedures by statistical or stochastic methods, while maintaining the efficiency of the system, is a new research topic, a new challenge I am working on now.

In addition, the system currently operates in 17 international capital markets. Their number can be increased to identify new opportunities. This step is just a matter of the level of commissioning in those markets and the computing resources available. The addition of new servers is already in the offing in the near future. System response time is also a factor that can significantly contribute to profitability. In the last year, the delay of the system was between 39 and 282 milliseconds between the time of receipt of the quotation until the time of transmission of the signal to a capital account. Combined with the delay in data transmission in the brokerage IT system, the total delay is less than a second today, but this can be substantially improved in the future, which will, of course, lead to further increase efficiency.

Many ask me why I don't live on a sunny beach among palm trees. I tell everyone that the first person who can relax is the investor. He has nothing to do. Once the capital account is opened and connected to the server, the software will do everything automatically. The investor can follow the profit evolution by the phone and can do it from any beach, without any stress.
---



error: Content is copyrighted !!!