Is back-testing a reliable method for evaluating technical strategies?
Forex trading is risky. To reduce risks, traders employ different strategies and
analytical tools, but even then, there is a degree of uncertainty which cannot be
entirely eliminated in the chaotic environment of trading. Even the best technical
strategies can at times lead to significant losses, and the only way to avoid such
losses is ensuring that money management tools and risk controls are applied vigorously.
But some traders find even those approaches insufficient for establishing the safety
of a strategy. Many are misguided into thinking that we can discover methods which
generate excellent returns in all circumstances, and then turn to back-testing to
isolate and identify them. Applying their strategies to historic data recorded over
the course of decades in some cases, and traders believe that they can refine their
strategies in such a way that perfect harmony with the price action of the past is
acquired. Is that possible? To an extent, yes. Is it useful? Unfortunately, no.
There are many errors in the reasoning that attaches too much importance to back-testing.
First of all, there is no evidence to back the belief that prices somehow repeat
themselves. Triangles, trends, and ups and bottoms are constantly encountered in
trading, but each time we find them in a different formation, a different order,
with a different degree of volatility. Indeed, the rules that generate the price
action are unknown, and as the market action is chaotic, I we cannot expect the same
rules to generate profitable outcomes at all times. Since back-testing applies the
same strategy to past events, and past events are extremely unlikely to be repeated
in the future, the results generated have little relevance for future price action
and profits.
In addition, back-testing ignores many issues which are specific to the broker. Server
failures, software crashes, slippage, and many other unpredictable events can complicate
trade decisions greatly and necessitate human input. The issue is even more complex
if we consider that not all back-testers use automated tools. In the cases where
emotions have a role, back-testing is almost entirely useless. Finally, although
the market can show calm and stable conditions overall, problems related to your
own broker, such as widening spreads, and lack of liquidity may create an environment
that is very different from what is assumed for back-testing.
The various ways of forex broker comparison can grant us insight into the efficacy
and reliability of a firm, but no broker will offer the perfect tools which will
not fail, or generate the unrealistic returns claimed by traders who depend on the
results of back-testing. Back-testing is fine as an educational tool, but it not
a way to discover successful strategies which will help us profit in the unpredictable
environment of the real market.
article ends...