Mechanical Forex

Immediately after you open a new trading position you face the challenge of properly managing it to obtain a profitable outcome. Part of this management exercise is to decide whether or not to use a stop loss and/or a take profit, a decision which will heavily impact the outcome of each particular trade and the general soundness of your trading strategy. On today’s post I am going to talk about why using a take profit is probably a bad idea and will also go into why so many traders – especially more novice traders – have historically advocated in favor of its use. We will talk about the main consequences of using a take profit in terms of both actual trading and the strategy development process.

–

–

What is a take profit (TP)? A take profit is mainly a limit order you place on your account such that you exit a position at a predetermined level of gain. If you enter a trade and then place a take profit 50 pips in favor of your trade you will then see your trade exit at this point when this level is reached. The main consequence of using a TP is that you will exit the market at a predetermined point and any subsequent profits or losses from this position will be averted. The take profit sets a fixed expectation of maximum gain for every position and this is the main reason why it’s weaker compared to other position management methods.

It is however no mystery that many people advocate for the use of the TP. You can see several articles about this online (here, here and here for example). Generally traders who support the use of the TP will put forward arguments such as having difficulty interpreting phrases like “let your profits run” and having a far easier time psychologically with the use of fixed profit targets. In reality what the take profit does is to provide a limit to the expectation of gains and setting this expectation somewhere can be in fact rather tricky. Deciding what amount of money should be expected for each trade means that you must rely on some measure of past trades (optimize the TP) or have some predetermined relative notion of how far price will go (for example if it will stop at a high/low, MA, etc), which may also require some optimization (what high/low, which MA, etc).

–

–

Once you have a historically optimum value for a TP you have run into an important problem which is that if this value changes your entire profit structure expectation will be broken. This may be very hazardous for your strategy if it tends to happen towards the negative side. Imagine that you have decided to always place the TP at the high/low of the past 10 hours but suddenly price is not reaching this level after your entry signal but on average reverses towards your SL 8 pips before reaching it. Your TP will not be triggered and you will now have a net losing system that is probably far worse than your initially designed strategy. Expectations for trades are not constant and having to set a fixed expectation leads to a sub-optimal trade management structure.

Another very important point is the probability that your historically chosen TP has just worked due to chance alone. During the past two years of data mining experience I have seen that strategies including a TP always have a much higher data-mining bias, far more than what you would expect from including an additional parameter. This is because there is an innate probability to reach a given TP level just due to random chance and by establishing a clear cut level where trades are exit you instantly put all this profit due to randomness into your system statistics. If you do not use a TP then most randomness will fall towards negative outcomes which reduces data mining bias, when the TP is used data mining bias increases a lot and it therefore becomes much harder to distinguish something that comes from a real historical inefficiency from something that just comes due to chance.

Establishing a trade management mechanism that is successful but does not include a TP is the real challenge. This is what generally constitutes the “let your profits run” advice, it does not mean to trade without any expectation of trade evolution but to take advantage of how trades are expected to evolve without having to “cut off” profits at arbitrarily chosen levels through the use of a mechanisms like the TP. In practice this means to establish a mechanism that properly evolves the expectation of adverse movement such that this expectation can be moved into profitable territory with time. This is what I have established with function based trailing stop mechanisms in the past, which are trade management methodologies that work great without the problems associated with TP mechanisms. The fact that these mechanisms do not “fall apart” by changes in how trades evolve is also important, since the expectation of adverse excursions is always controlled any reduction in the expected level of profit will fail to be catastrophic (contrary to what happens with the TP), additionally and probably more importantly is that the lack of an “upper cap” in expected trading profits also allows you to take advantage of being “lucky” when price moves in your favor much more than expected from your initial trade analysis (something you clearly miss with a TP).

–

–

In summary a TP can always be replaced with a profit trading management method that controls adverse excursion expectations instead of setting profit targets “in stone”. Moreover avoiding a TP reduces data mining bias and – given adequate management techniques – opens up the opportunity to far more robust and capable trade management. If you would like to learn more about automatic system generation and how you too can generate your own systems using the power of GPU technology please consider joining Asirikuy.com, a website filled with educational videos, trading systems, development and a sound, honest and transparent approach towards automated trading.