A trading system or method?

On this site I offer a few e-mini trading systems and two trading methods. The systems are purely mechanical and you can use them to trade the S&P 500 e-mini futures (also known as ES) or (in the past) ER2 (currently TF), which is the Russell 2000 e-mini futures, two very popular e-mini instruments. One of the methods is for trading stocks, the other is (or was) for trading breakouts in ES. I now offer a more comprehensive set of methods for trading day e-mini futures, KING. All have been originally designed by me.

As you see, I make a purposeful and clear distinction between systems and methods. (The latter can also be called strategies.) Any competent trader or system developer certainly realizes that the failure to distinguish between the two makes him or her rather sloppy and that’s at best. As a matter of fact, in my opinion, that makes them quite unprofessional. Yet, this happens so routinely among vendors that it finally prompted me to write this article.

One would think that it is easy to understand what the system is. It is just a strategy that can be subjected to systematic and objective evaluation (testing). Not every strategy can be elevated to the status of the system and this happens for two simple reasons.

Namely, the strategy may not be amenable to systematic testing, either for some intrinsic or only practical reasons, or cannot be formulated in a completely objective way. The latter category includes strategies that rely on pattern recognition which is almost always at least partially subjective. That’s just the nature of the beast and pretending it can be otherwise would be dishonest. The reason is this: whenever you deal with patterns, there is small but not insignificant a chance that you run into something that may keep you seriously wondering if this is really what you are looking for. This is so simply because some cases are not as well defined as others and some are defined so poorly that you will have a very hard time deciding if it still is the pattern you are interested in. In other words, patterns exhibit some continuity and it is not always possible to sharply distinguish proper patterns from improper ones. One example of such a strategy is my method of trading breakouts in ES. As rigorously defined as it is, it can still leave a tiny bit of room for subjectivity, which is why I may sometimes say that “it virtually leaves no room for guesswork.” (I said so in the context of the breakout trading method. That this method to trade breakouts in ES is really very precisely defined can be seen from the number of trades it admits per month; it is just a few.)

My original method of trading stocks cannot be raised to the status of the system for the other reason, namely it cannot be subjected to the systematic testing using the past data because the relevant data is not tracked. You could easily do this as you go, but that would not be as complete as one would like it leaving the past completely unexplored not to mention that it would involve quite a bit of time to do this for the entire universe of stocks that could potentially meet the method’s criteria. Since I have not found enough time to do this and since what my clients receive from me is tools with which to seek the best stock candidates and not a complete methodology of how to do this, I would feel very uncomfortable implying that this strategy of mine is anything more than just that. That does not mean that it is inferior to other strategies on the market. As I argue below, quite the opposite can be true, not to mention that unlike many other strategies, mine is also fundamentally sound.

In this context, I would like to note that a somewhat similar situation arises for George II. As you can notice, its performance had been examined only back to January 2006 when I released it in March the same year. This is not because I wanted to hide its more distant past, but simply because this past cannot be completely reconstructed, the reason being that the historical intraday data we need for this purpose is provided in the form that does not allow for totally complete backtesting. (It is be possible to approximate the data needed, but that would make the backtesting process a bit "patchy," so to speak.) You can easily test this system as you go and that is what I have been doing since January 2006. However, it is much easier to do this for one trading instrument or two than a much larger universe of stocks.

The fact that the strategy history cannot be totally reconstructed should not necessarily be viewed as disadvantage. In fact, it can be just the opposite. It is rather unusual to seek this kind of strategies and so they remain largely unexplored offering potentially higher edges than the strategies every trader and their grandmother have known for ages.

As I noted, the distinction between systems and (merely) strategies tends to be largely ignored by vendors. I believe that this is so for two main reasons. The first one has already been hinted at: they simply are not competent enough to discern what hardly is just an academic subtlety. This is not at all surprising as many of these people are just marketers who jumped on the vendor’s bandwagon knowing full well that trading ideas, tools, and strategies in particular, sell very well. And what’s the best way to sell a strategy than by wrapping it as a system. Doesn’t that sound much more glamorous than the plain ‘strategy’? You bet, it does! Who cares whether this is a system or strategy? If your objective is to sell this stuff to the first sucker to emerge on your website’s horizon, why would you indulge yourself in such silly hairsplitting? And that’s what I meant by the other reason.

Yet, if you are the designated sucker, you better make sure you know the difference I am talking about here or else you might end up with “the best system on the market,” according to the sales hype, that is not even a system, much less so the best one! As a matter of fact, how can you determine the ‘system’ quality if the ‘system’ is not objectively defined? The answer is: you cannot!

Most of what masquerades as systems out there is really strategies. Now, there is nothing wrong about it as long as the buyer is informed that what he is offered is a strategy and not a system. I would not bother writing this article if this indeed were the common case.

Any system that relies on pattern recognition is unlikely to produce the results that are claimed if any results are quoted at all. The latter approach can actually be more honest, particularly if the strategy in question is not very strictly defined. The results from such ‘systems’ are not easy to duplicate. The more often a system like that trades the more lax its rules are and the less likely you are to come up with the results quoted by its seller. There are ‘systems’ out there that trade 5-10 times a day. If they are based on pattern recognition you are buying yourself disappointment and frustration and not genuine results. The results from such ‘systems’ come from after the fact inspection of charts and they are likely to be different from the results you would have obtained had you traded the system in real time.

It’s no wonder then that Internet trading boards are crowded with negative posts from the people who thought they were buying a genuine objective trading system and ended up with just a trading strategy whose results depend heavily on the way you interpret things. This obviously varies from one person to another and thus you cannot expect to get the same results, even in principle, much less so in practice.

The moral from this somewhat elaborate article is quite simple and can be summarized in two points: a) to avoid disappointment make sure that you know if what you are buying is a trading system or a trading method and b) avoid vendors who do not differentiate between the two as such vendors tend to be incompetent or dishonest (usually both).