Warning: Turn your sensitivity meter down a little. This is a no sugar-coating, tell-it-how-it-is article (but rest assured it comes from a nurturing place).
All over the globe, trading gurus attempt to sell their wares (software, the ‘holy grail’ of trade set ups etc) using retrospective charting examples. Such powerful visual “evidence” is often used to persuade prospective FX clients that this vehicle is ‘easy’ to make profit with. With little work, little time, or whatever marketing buttons they are using to press to get a response.
So, hours of energy invested, often cash is exchanged and yet more often than not, with an off the shelf system in place (often just an entry system which we know is never going to offer a complete trading solution) traders are left feeling more than a little disappointed that such “guaranteed, easy riches” are not showing up in their trading account.
On an individual level we see similar.
Much airplay is given to the merits of back-testing and yet as with the aforementioned guru approach, you can just about find examples, if you look hard enough, of chart examples that mean this “next new indicator thing” is now the answer to replenish your now depleted finds. So, what happens, we have a system change, and yet results still often fall short of expectations.
There are 3 common dangers of the retrospective approach to creating (if you haven’t a trading plan already) or altering an existing plan that are worth highlighting.
#1 – Overstating the function of back-testing.
Let us be completely blunt. The purpose of back-testing is NOT, nor should ever be viewed as evidence that a trading plan, based on what ever system you are exploring, will work for you in the reality of live trading.
Back-testing does not generally consider:
a. The impact of economic data releases and revisions,
b. The political and general climate both globally and specifically in the countries that currency pairs relate to,
c. Individual investor behaviour re. timeframes, time of day that they trade, nor their ability (or otherwise) to act or inaction on a change of sentiment,
d. Unplanned events such as escalating conflict (or the threat of such),
e. The relationship and impact of other financial instruments of FX pairs e.g. equity and bond markets, commodities
So, why back-test at all if the evidence could be so flawed?
The answer is simple, back-testing creates evidence, not that a system will definitely work for you as a trader, but ONLY as evidence that a forward (or prospective) test may be worthwhile.
So, the bottom line is the function of back-testing is to justify the time and effort to prospectively test. It is after such a prospective test that system changes can be made/developed.
#2 – Failure to gather a critical mass of evidence
There are two issues here.
a. What constitutes enough evidence to move to the next stage of system testing. Quite often traders will make decisions on a limited amount of data e.g. one timeframe and one currency pair, over the last couple of months on which to make system decisions. Now you have read this it may seem obvious and may not need pointing out (but we will anyway) why this is insufficient information on which to base a “cross the board’ entry and exit system.
b. The second issue here is one of selective evidence gathering. A natural human response when excited by an idea is search for evidence to back up that idea. The potential danger with this is that we often tend in this search, to ignore information that refutes our idea.
#3 – The reason behind doing this may not be that your system is failing rather it could be a YOU issue.
System skipping is common amongst many traders and is invariably motivated by results that are not as desired.
Here is the danger. As much of what goes into creating trader results (some would suggest up to 80%) is due to behavioural issues (we have waxed lyrical about trading discipline previously) unless you:
a. Have a trading plan that is specific, measurable and comprehensive AND
b. Follow it religiously ‘to the letter”
then you are not really in a position to make a judgement on whether system could serve you well or is likely not to produce desired results. AND to add to this, as such behavioural issues have not been either acknowledged or addressed whatever system (based or retrospective charts or not) is more likely to produce equally disappointing results.
So, before you start on the journey of altering a system you should logically make every effort to have, follow and measure the impact of any system before you even consider changing it (or looking into what you may change it to). This MUST be your #1 priority before going down any path of system alterations.
So there you have it. You have a choice to take action of course on what you have read, If so, your missions going forward are:
a. Make sure you have a comprehensive plan that you follow. Then, and only then, should you begin to explore further development including the use of retrospective charts (or back-testing)
b. Recognise the SOLE PURPOSE of back-testing is to create evidence that a forward (or prospective) live test is justified.
c. Make sure you are basing any potential system change on a enough “balanced” data.
Previous: Trend retracement or reversal? – Can volume be a useful “clue”?