Are computer models beating the market?

One of the greatest Israeli mathematicians in the field of game theory, says that during all the years of his career he came across many computer models that claimed to beat this or that market, but the people who built the models were never able to prove to him that they actually succeeded in doing so

It is possible that the ultimate ambition of many of the most brilliant among us is to build a computerized model that knows how to make money from money without additional help from us. "The golden fleece of the profession", "the ultimate ambition" and other nicknames were given to this field. It is about building computer models, which will operate without human contact and will be able to do what is called in the professional parlance: "beat the market". You'd be surprised how many there are trying to do this. Some build models that focus exclusively on the foreign exchange market. Some focus on stock or commodity markets and some dare to try and predict the movement of many markets. As there was a development in the power of computers and the ability of a private person to sit at home and flow data from countless sources into increasingly powerful processing systems, so did the popularity and attempts of programmers and even laymen to build a model that would crack the movement of the markets. The main idea was and remains that the model will be able to give a forward forecast in a period of time that will be enough for the model builder to run a speculation that will allow him to make a lot of money for sure.

Do computer models without human touch, or in fact without additional human thinking, really work?

It is important to emphasize already at this point that what is written here refers to completely automatic models, which claim to be able to see the future, without additional thinking on the part of humans. There are many computer models that are "decision support models". Decision support models are very effective, make the work of the investment manager considerably easier, allow for quick filtering of a lot of material, and manage to locate parameters and investment options that the investment manager might have missed. These models are able to sort out the residue from the pulp and have semi-automatic capabilities. Still, at the end of the road, after the model has given its recommendation, a decision is required by the investment manager regarding whether to invest, in what amounts, at what times, and so on. More advanced models can provide answers to these questions as well, but the final decision is still up to the investment manager.

One of the greatest Israeli mathematicians in the field of game theory, a field in which Israelis are considered very strong (Nobel laureates Uman and Kahneman as the most cautionary examples), told me that during all the years of his career he came across many computer models that claimed to beat this or that market, succeeding in predicting such an index or another and so on. But the people who built the models were never able to prove to him that they were indeed able to do this - to predict the direction of the markets accurately and over time. Moreover, he does not believe that this is possible...

I have personally come across quite a few examples of such models. Since it is known in the Israeli capital market that you can contact me regarding advice on setting up hedge funds - I met a series of brilliant Israelis with different and interesting ideas. Some of them presented me with computer models that are supposed to hit this or that market without human intervention. I take my hat off to these people. The amount of work that many of them put in was tremendous. One told me that he did not sleep day and night and worked continuously on what he considered the "discovery of America" ​​in this field, no less. Among them was a mathematics professor from the Weizmann Institute, an aeronautical engineer who specialized in the field of air-to-air missiles in the defense industries, an accountant for a large international company, and more.

Unfortunately, I am not aware of any of these models being successful in the long run. There were models that did great for six months to a year, then market conditions changed and the model stopped working. There were others that on paper their models worked great but when they started trading real money all of a sudden it didn't work. There was a model that did not take into account all transaction costs - the differences between Ask and Bid, bank fees and so on. There were many explanations and many reasons for the failure of the models. I have also heard many stories about such and such successes, but I do not know of any cases that made money systematically and especially over time.

At the beginning of 2006, I was introduced to a new immigrant Jew from the USA, a man in his late fifties who settled in Jerusalem, stopped working because he didn't need money and developed a super model - which he claimed predicted all markets. This person is quite famous on Wall Street, articles have been written about him, he appeared on CNN and some of the largest investment houses in the US have used and are using his services. He used to charge tens of thousands of dollars and more from every investment house for his market forecasts and he probably had many successes in the past.

In any case, in a series of meetings I held with him, he offered his good services for financial vaccines, in forecasting various foreign exchange markets, commodity markets and bond markets. He told me a lot about his computer models that are able to predict almost anything. I must point out that he had bombastic claims, bordering on madness. His models, he claimed, were able to predict dramatic falls in stock markets because they are able to predict, among other things, times of major terrorist acts and even wars - no less. I expressed to him my skeptical opinion regarding these issues. I explained to him that there is a saying in the field of exact sciences that says that extraordinary claims must be accompanied by extraordinary convictions. At the same time, I hoped in the depths of my heart that I had indeed encountered a person who cracked the ability to predict the future. What could be better than that? You don't have to try harder, you just have to quote what his models say and report it to our customers and they will make a lot of money and pay us a lot of money. A winning model without a doubt!
He dismissed my disbelief with a wave of his hand. He already met my shoulders, "You just can't understand" he said. Of the trends his models predicted would occur in 2006, there was one particularly dramatic prediction. His models predicted at the beginning of 2006 that the Dow Jones would fall by 30% (the Dow at the time of the forecast was at 10,000 points) and would reach a level of 7,000 points during the months of September and October 2006.

This prediction did not really come true. For those who happened to forget, the Dow Jones reached a level of 12,000 points at the end of October... Worse, his claim that his models are able to predict major attacks and even wars was completely falsified on July 12, 2006, when the Second Lebanon War broke out, which his models did not predict at all... not that anyone Someone else predicted the war, or someone else managed to predict long-term what was happening in the financial markets, but apparently for the payments he received, and especially in view of his exceptional claims, it was necessary for the models he built to work better. In retrospect, he had some beautiful predictions, but also painful failures. Since you cannot know in advance in which area his models will be right and in which area they will succeed, it was not possible to make money from his predictions, at least not in 2006...

There is a brilliant mathematician whom I have known for many years. This man managed to get very rich from his activities in Israel. He did this by spotting anomalies in the markets. By detecting anomalies in the markets, money can be made for sure. You locate things that others didn't see, or missed, or analyzed incorrectly, you buy the commodity or sell it short and wait. Sooner or later the market exceeds the anomaly and then the prices of the securities adjust to the level they should be at and you sell or buy back the goods and make money.

In any case, he accumulated most of his wealth during the nineties of the last century, when the markets in Israel were not developed enough. Since the beginning of this century, when the intensive activity of foreign investors in the local markets was increasing, his successes were decreasing. At the same time as this activity, he has been working for nearly 20 years on a super model that will be able to crack all the markets at the same time, find all the connections between them and know how to predict all the market trends in advance. He is not doing this to make a huge fortune, he is already very rich anyway. His goal is much more ambitious - a Nobel Prize. So far there have been several times when he thought he was very close. I would chat with him and he would tell me very excitedly that this is it, he cracked the matter. After that I would stop hearing from him. Something went wrong with the super model. I wish him success not so that he will be extremely rich, but so that another Israeli will win the Nobel Prize.

I heard that abroad there are successful hedge funds based on computer models without human contact. It is possible. They haven't been able to convince me yet.