The impossibility of predicting the future

I am providing my reflections on this industry after several years of study, experimentation, and contemplation. These are personal opinions that may or may not be shared by others.

The dream of being able to dominate the markets is something that many people aspire to, but unfortunately, it is very difficult because price formation is a complex system influenced by a multitude of dynamics. Price formation is a deterministic system, as there is no randomness, and every micro or macro movement can be explained by a multitude of different dynamics. Humans, therefore, believe they can create a trading system or have a systematic approach to dominate the markets precisely because they see determinism rather than randomness.

When conducting many advanced experiments, one realizes that determinism exists and can even discover some "alpha". However, the problem arises when trying to exploit this alpha because moments of randomness will inevitably occur, even within the law of large numbers. But this is not true randomness; it's a system that becomes too complex. The second problem is that it is not possible to dominate certain decisive dynamics that influence price formation. I'm not saying it's impossible, because in simpler systems, such as the price formation of individual stocks or commodity futures, it is still possible to have some margin of predictability if you can understand when certain decisive dynamics will make a difference. However, these are few operations per year, and in this case, you need to be an "outstanding" analyst.

What makes predictions impossible, therefore, is the system being "too" complex. For example, an earthquake can be predicted with 100% accuracy within certain time windows if one has omniscient knowledge and data. Humans do not yet possess this omniscient knowledge, and thus they cannot know which and how certain dynamics influence earthquakes (although many dynamics that may seem esoteric are currently under study). The same goes for data. Having complete data on the subsoil, including millions of drill cores, would be impossible. This is why precursor signals are widely used in earthquakes, but in this case, the problem is false signals. So far, humans have only taken precautions once, in China, because the precursor signals were very extreme, which saved many lives. Unfortunately, most powerful earthquakes have no precursor signals, and even if there were some, they would likely be false alarms.

Thus, earthquakes and weather are easier to predict because the dynamics are fewer, and there is more direct control, which is not possible in the financial sector. Of course, the further ahead you go in time, the more complicated it becomes, just like climatology, which studies the weather months, years, decades, and centuries in advance. But even in this case, predictions become detrimental because, once again, humans do not yet have the necessary knowledge, and a small dynamic of which we are unaware can "influence" and render long-term predictions incorrect. Here we see chaos theory in action, which teaches us the impossibility of long-term predictions.

The companies that profit in this sector are relatively few. Those that earn tens of billions (like rentec, tgs, quadrature) are equally few as those who earn "less" (like tower, jump, tradebot). Those who earn less focus on execution on behalf of clients, latency arbitrage, and high-frequency statistical arbitrage. In recent years, markets have improved, including microstructure and executions, so those who used to profit from latency arbitrage now "earn" much less. Statistical arbitrage exploits the many deterministic patterns that form during price formation due to attractors-repulsors caused by certain dynamics, creating small, predictable windows (difficult to exploit and with few crumbs). Given the competition and general improvement of operators, profit margins are now low, and obviously, this way, one cannot earn tens of billions per year.

What rentec, tgs, quadrature, and a few others do that allows them to earn so much is providing liquidity, and they do this on a probabilistic level, playing heavily at the portfolio level. Their activity creates a deterministic footprint (as much as possible), allowing them to absorb the losses of all participants because, simply, all players are losers. These companies likely observed a "Quant Quake 2" occurring in the second week of September 2023, which, however, was not reported in the financial news, possibly because it was noticed only by certain types of market participants.

Is it said that 90% lose and the rest win? Do you want to delude yourself into being in the 10%? Statistics can be twisted and turned to say whatever you want. These statistics are wrong because if you analyze them thoroughly, you'll see that there are no winners, because those who do a lot of trading lose, while those who make 1-2 trades that happen to be lucky then enter the statistics as winners, and in some cases, the same goes for those who don't trade at all, because they enter the "non-loser" category. These statistics are therefore skewed and don't tell the truth. Years ago, a trade magazine reported that only 1 "trader" out of 200 earns as much as an employee, while 1 in 50,000 becomes a millionaire. It is thus clear that it's better to enter other sectors or find other hobbies.

Let's look at some singularities:

Warren Buffett can be considered a super-manager because the investments he makes bring significant changes to companies, and therefore he will influence price formation.

George Soros can be considered a geopolitical analyst with great reading ability, so he makes few targeted trades if he believes that decisive dynamics will influence prices in his favor.

Ray Dalio with Pure Alpha, being a hedge fund, has greater flexibility, but the strong point of this company is its tentacular connections at high levels, so it can be considered a macro-level insider trading fund. They operate with information not available to others.

Therefore, it is useless to delude oneself; it is a too complex system, and every trade you make is wrong, and the less you move, the better. Even the famous hedges should be avoided because, in the long run, you always lose, and the losses will always go into the pockets of the large liquidity providers. There is no chance without total knowledge, supreme-level data, and direct control of decisive dynamics that influence price formation.

The advice can be to invest long-term by letting professionals manage it, avoiding speculative trades, hedging, and stock picking, and thus moving as little as possible.

In the end, it can be said that there is no chance unless you are an exceptional manager, analyst, mathematician-physicist with supercomputers playing at a probabilistic level, or an IT specialist exploiting latency and statistical arbitrage (where there are now only crumbs left in exchange for significant investments). Everything else is just an illusion. The system is too complex, so it's better to find other hobbies.