When You Should NOT Trade a Given System: Five Red Flags To Avoid Trading a Strategy

There are countless strategies we could run on our live trading accounts but it is obvious that from all of these strategies only a handful will be worth running and therefore we need to be very selective with which strategies we would want to run and which ones we would not want to run in our accounts. How do you determine which strategies might be worth the trouble? Definitely there are some “rules of thumb” for choosing strategies, such as avoiding martingales and strategies with unreliable or excessively broker dependent results but there are a few other tips which are not that obvious that will help you decide whether or not you should allow a trading strategy to be executed on your Forex trading account. Through the following paragraphs I will share with you five red flags which tell you to AVOID running that strategy on your account.

As I mentioned before there are probably many obvious criteria you need to take into account when running strategies that I will not mention here (probably because I have mentioned them extensively before on this blog). Things such as having long term reliable backtests which are as less broker dependent as possible and using strategies that do not exploit backtesting faults  is a must, using strategies that are only profitable due to the fact that they postpone risk through the use of unsound money management is also a must, things such as D’Alemberts, martingales, etc should be completely avoided. However – once you’re past these obvious characteristics – there are many other things you need to consider. Here are five red flags that tell you to AVOID trading a system:

1. You cannot trade it manually. If you have a system which you cannot replicate completely by trading manually then you should NEVER put it on your live account. The corner stone of automated trading success is understanding and if you’re trading a blackbox whose results you simply cannot reproduce from a manual perspective then you’re simply doomed to failure. You should be able to look at the same chart as your EA and come up with the same signals with the same lot sizes, TP, SL and exit targets by using the same logic as the EA. The way in which your system trades should make absolute sense to you and you should be able to replicate its moves. If you cannot trade something manually then you should NOT trade it, how could you be successful with something you don’t even understand?

2. You don’t know the development process. A key aspect of succeeding with an EA is not only knowing how it trades but knowing how each parameter the EA trades got the value it now has. How do you know that the EA is not extremely curve fitted? How do you know that the way in which the parameters were deduced is good? You not only need to know how a system trades but you should also know the exact process which led to the achievement of the current parameter sets and what possible pitfalls may arise from that process. Knowing an EA inside out – including its end-development stages – is a key aspect of automated trading.

3. You have no Monte Carlo simulations. After spending so much time learning and running Monte Carlo simulations and seeing in live trading how valid and important the values obtained from these simulations are I would never ever run a system for which I lacked this information. Monte Carlo simulations reveal problems in strategies with unsound money management that could have just “been lucky” in a test and they also show the probability of a system behaving “worse” than in the back-tests while still being bound by its core statistical nature. Monte Carlo simulations are useful to define worst case scenarios, possible profit scenarios, robustness, broker dependency, etc.

4. You have performed no stress testing. One of the things which I consider most important before running a strategy is having information about stress tests of the system. Increasing the spread to see its effects, introducing 5% random parameter variations and doing out-of-sample tests are all ways to ensure that a system is resistant to things that attempt to “mess” with its profitability. When I run a system I want to satisfy at least some core aspects of robustness in order to ensure that I am not running a piece of garbage which might just fail tomorrow with some brilliant settings run today. As I am trading not to get impressed by back tests but to make money having stress testing information from the systems I use is absolutely vital for success.

5. You have performed no reliability tests. Perhaps something that is very important but few people do is to perform strong reliability tests of your systems. Run a platform with 20 instances, crash your platform, unplug your computer, do all things you can to try to make the system “lose sync” or take wrong action on a demo account. By carrying out reliability stress tests you ensure that your trading platform will be “up to the task”, being able to handle these things when they happen in real trading. Using a system that “forgets trades” or that suffers from instability problems when running under the conditions you want to use it is a definite “no-no”.

So as you see the above red flags are quite important although they may not appear to be obvious to the large majority of traders. Certainly if you run a system you totally understand, a system for which you know the entire development process, which you also have Monte Carlo data for and which you have stress tested both for trading and stability purposes you will have a much BIGGER change of achieving success. The key to success in trading – as I say so frequently – is understanding, so make sure you perfectly understand everything you run on your live accounts, make sure you know all its characteristics and make sure you stress those systems like there was no tomorrow!

If you would like to learn more about automated trading and how you too can get a true education in system evaluation, design and stress-testing please consider joining Asirikuy.com, a website filled with educational videos, trading systems, development and a sound, honest and transparent approach towards automated trading in general . I hope you enjoyed this article ! :o)

Print Friendly, PDF & Email
You can leave a response, or trackback from your own site.

2 Responses to “When You Should NOT Trade a Given System: Five Red Flags To Avoid Trading a Strategy”

  1. Maxim says:

    Daniel,

    A really useful post, thanks for sharing those “flags”.
    Will one be able to reproduce Sunku trades manually?

    Maxim

    • admin says:

      Hi Maxim,

      Thank you for your comment :o) Well you could certainly reproduce Sunqu trades if you use the NN as an indicator but you will never be able to reproduce the whole logic (how Sunqu spits a value) because an NN is – in essence – a complete black box. This is why we need to exercise GREAT care when moving onto this field! (our understanding of how decisions are being made is limited). This is why testing of NN systems needs to be carried out in a “moving window” type of walk forward analysis were we run a 10 year test with a retraining of the network every X days. Thanks again for your email,

      Best regards,

      daniel

Leave a Reply

Subscribe to RSS Feed Follow me on Twitter!
Show Buttons
Hide Buttons