Judging Trading Strategy Survivability: Looking at Systems like the Military

The most important topic when dealing with trading strategies is to gauge their ability to survive to future market conditions, their ability to yield profit through an unknown trading future. Although it is obvious that we can never have certainty about a system’s ability to survive, there are various trading strategy characteristics and tests that may help us increase the probability of a future positive outcome. With this in mind it becomes key to device ways in which this probability to survive can be measured, ways in which the different paths through which a system can fail can be judged and analysed. Today I want to give you a view of robustness analysis based on the concept of survivability used by the military to judge combat systems. This concept includes clear criteria that may help us gauge differences between the different probabilities of survival from our different trading strategies.

One of the key areas where the ability to survive to future and unknown conditions plays a key role is in the design of combat systems. When you’re designing a new fighting machine it is obvious that you will want it to survive to fighting conditions that will be unknown, with threats that might be from the very mundane to the very odd. Combat systems are always designed to be able to face several threats that are known to be present and to withstand threats that may not have been considered at the time. The idea here – as it is the case for trading systems – is to test a system under a set of well known conditions with the hope that the system will be able to survive a whole new range of stress outside of its design experience.

How do the military judge the ability of a combat system to survive? They have created three important categories that define the ability of a combat system to withstand “combat stress”. The three key concepts here – that make up survivability – are susceptibility, vulnerability and recoverability. The first concept encompasses the inability of a combat system to be hit by a weapon (how susceptible it is to receive damage), the second talks about the inability of the system to withstand the hit and the third talks about post-hit effects; how the system behaves after the hit, how it recovers form the hit and how it escapes and evacuates from hostile situations. All of these concepts are very easily transported into the world of algorithmic trading techniques, giving us a unique view of a system’s robustness.

First of all we need to think about what a “weapon” means for our trading system. A weapon is not necessarily a drawdown period – as it would be intuitive to think – but it is instead a set of market conditions for which the system is not prepared, hostile data for which the system has no previous training. A weapon – the unkown set of market conditions – does not necessarily cause a hit (drawdown) but it is the prerequisite for one to happen. We can then measure susceptibility as the ability of a system to avoid drawdown under a new set of market conditions (the ability to avoid being hit). We can test for susceptibility in several different ways but the most straightforward is to introduce random distortions in data and see how these changes the system’s ability to avoid drawdown conditions.  Another one could be to introduce patches of random data through the set and see how the system reacts to this. We can also simply examine the back-testing data to measure the in-design susceptibility of the strategy (how many times we went into drawdown).

The next measurement of survivability – vulnerability – can be considered the ability of a system to control the extent of the drawdown period ones it goes into one. When the system is “hit” – suffering a drawdown period – we can then measure the drawdown depth and “length to deepest point” as measurements of the system’s vulnerability, a measurement of how bad hits are when the system actually takes one. A system that goes into less drawdown periods – less susceptible – can be more vulnerable if the few drawdown periods it gets into are deeper. The combined view of suceptibility and vulnerability gives us a clear idea of how frequently and how deeply our system is bound to go into losses. We can also test vulnerability through distortions by making the system face prolonged periods of random data to see if it can control its drawdown or if it goes through a very bad losing period.

The last measurement – recoverability – talks about the ability of our trading systems to go into drawdown and then reach a new equity high plus their ability to stop trading when market conditions become hostile. It is key to remember that recoverability does not only measure the ability of a system to “recover from the hit” but it also measures the ability of the system to successfully exit “hostile conditions”. This means that the measurement should include both the ability of a system to reach new equity highs – drawdown period lengths – and the ability of the system to exit once conditions start to become hostile (drawdown period depths and lengths become too big). The above random distortions can also be used to play with recoverability, testing implementations that would allow the system to exit as the conditions become too hostile (too many weapons attempting to hit the target — not even necessarily hits!).

As you can see, combat systems have a lot of things in common with trading systems when evaluating survivability as the concepts used to evaluate a system under unknown conditions apply to both worlds. Do you think the above concepts are applicable? How would you apply them to your trading systems? Please leave a comment with your opinion :o)

If you would like to learn more about automated system building and evaluation 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 “Judging Trading Strategy Survivability: Looking at Systems like the Military”

  1. Fd says:

    Hi Daniel,

    I think your analogy only makes limited sense. It is true that in trading you’ll better have a strategy. Whatever you are doing a sound strategy always will help you to be successful, which is also true for a great variety of jobs or situations. The nature of trading is that you are ready to accept whatever the market is offering to you. You can’t defeat the market, you can only try to evade that the market defeats you – if you want to see the market as an enemy.

    My preferred analogy would be to compare a trader to a meteorologist. If he is right with his forecast he will wear suitable clothes next day, if not he has an increased risk of getting a cold or bathing in sweat.

    Best regards,
    Fd

    • admin says:

      Hi Fd,

      Thank you for your comment :o) I like the meteorologist analogy, very commonly used in game theory (as I am sure you know) it does fit the view of trading as a game where we are competing to find competent forecasters, not to avoid obstacles (which is what the military analogy tries to depict). In any case, thanks a lot for offering your opinion :o)

      Best Regards,

      Daniel

Leave a Reply


Fatal error: Call to undefined function mcsp_html() in /home1/danielfp/public_html/mechanicalforex/wp-content/themes/modern-style/comments.php on line 71