Recognized Market Picture(TM) explained. A snapshot as of December 12, 2018.
From your HEDGE21 data analytics team
There are good reasons why we did not pick Artificial Intelligence techniques to compute the HEDGE21 Recognized Market Picture RMP (TM), but rather a probabilistic model.
The big picture:
AI needs measurement data, mass data, Big Data. Even though the financial markets every second generate tens of thousands price ticks only for derivatives and in the US alone, relevant events such as the ECB Press Conference only take place every six weeks. A few hundred appearances by Jean-Claude Trichet or Mario Draghi since January 1, 2002, the introduction of the Euro, are not Big Data, but very small data.
- Looking back while driving forward: Future events are not hidden in past data. In today’s hyper connected world, our entire environment exposes and ever growing complexity, and complex dynamic systems have properties: You cannot predict their behavior.
- Spurious effects: All recorded data, when analyzed, produce spurious effects. These are correlations which are really existent in the data and which will be detected and exploited by an Artificial Intelligence with great certainty. In real world markets, correlations from spurious effects do not have any meaning. But if you placed your financial bet on a spurious effect – even without knowing, because many data scientist cannot explain the findings of their AI –, you might be in real trouble soon.
- Markets are political: AI is a number cruncher. But the numbers we collect represent a mere fraction of today’s crazy world. How do you tell a machine to take into account soft factors such as policies, sanction regimes, the Brexit chaos? Sure, you may point to the Efficient Market Theory: All information is factored into the prevailing price of an asset. But why would you then even consider using AI to crunch market data? AI is all about detecting patterns, e. g. private information in markets or the information imbalances which might give you an arbitraging opportunity.
Unlike AI, a probabilistic model of the markets is then a good model, if it provides a most accurate representation of the real markets.
The RMP(TM) market model generates hundreds of thousands of simulations of how a market can behave in the future. Compare it to the weather forecast. If the model simulates the market a day or two ahead, “forecast” is more accurate than for far-ahead times. Tomorrow, the EURUSD price is likely to bounce around today's price. But it may also occur a drift or even a trend in the one or other direction, depending on a scheduled event or even a Black Swan. A good market model has to consider all such eventualities. Therefore, RMP(TM) does not predict future rates, but rather calculates probability distributions.
Between the lines:
No, you don’t trade the RMP(TM). The RMP(TM) provides you with situational awareness and an input to your own decision making process how to deal with your investment risk.