A successful trading strategy is based on market inefficiency, a unique feature that gives an advantage over other traders. It is important that this inefficiency is stable.
Every market inefficiency has a lifespan. A trader’s dream is to find such inefficiency that will generate profits indefinitely.
Such inefficiency exists.
The behavior of any system becomes easier to predict when the system enters the realm of extreme, atypical values. The trader’s task becomes: 1) finding and formalizing such a pattern; 2) developing a technological solution to exploit it.
The fact is that from the realm of extreme values, the system always tries to exit. It’s similar to gas, where gas molecules are events. Any local compression leads to the movement of gas and the maintenance of average density. Algorithmic trading in such a situation has a decisive advantage over conventional trading, as an algorithmic trader can programmatically find those very extreme deviations in Big Data.
Imagine, for example, that you conducted 100 backtests of your strategy for the past month, differing in key settings (entry time, indicator period, stop-loss size, etc.) and generating, thus, the results of 100 independent sets of settings. Out of 100 sets, some turned out to be near-zero, some demonstrated significant profits, and some – losses. You can choose one set to trade in the new month! Which one would you choose?
Statistically, most people choose the most successful set. It seems that the streak of luck is bound to continue, and the chosen setup will not lose its effectiveness. This is how many come to create “paper grails” (that work nowhere except in backtests) through over-optimization (the main scourge of algorithmic trading).
Everything works the other way around when it comes to interventions in the realm of extreme values.
• If you take an ABNORMALLY GOOD set (with no losses or with minimal losses) – in the new month, you risk facing a pullback from the zone of extremely positive results. This is often referred to as: “Over-optimized!”.
• If you take an ABNORMALLY BAD set (with no profits or minimal profits) – in the new month, the system is likely to move into the range of moderate good performance, or even show one of the best results.
Of course, this logic alone is not enough. You need to do many other things: ensure that the abnormally bad series has ended (the goal is to enter at the end of it, not at the peak!), determine the marker for a probable reversal, and much, much more…
At EDVI trade, we have done the main thing – determined which market inefficiency can and should be utilized. Why this works will be explained in future posts.
