Monday, November 16, 2009

Knowing Before Knowing, The dangers in Hindsight

A few days ago, one of my new pupils in forex told me that he had been going through some automated trading system websites and that he payed close attention to the disclaimer shown at the end (shouldn't we all pay attention to that disclaimer). He wanted my opinion about hindsight and why it is important to consider it when trading either a manual or an automated trading system. Since this is an extremely important subject I wanted to dedicate today's post to the discussion of hindsight and how it can affect our trading negatively, specially when evaluating discretionary trading strategies. What is the problem of hindsight ?

To begin with, having hindsight is knowing something before it happens, that is, knowing the future. When we are evaluating our automated or manual trading strategies we have the benefit of hindsight. We all know the EUR/USD reached a high near 1.60 and during the next year dropped to near 1.29. We know that the pair then rallied towards 1.5, we all know this. What is wrong about this ? Well, it depends. I thought that I was barely affected by hindsight until I discovered that on an statistical study of a system I was doing on the EUR/USD I had been subconsciously supressing short positions before 2008 and long positions after mid 2008. I was in fact getting better results because I knew what was going to happen and I was subconsciously looking for profitable results. This is an inocent, bad consequence of hindsight, I was overestimating profitability.

Of course, there is a much more evil side to hindsight. Expert Advisor sellers looking to get very high profits in backtesting results can effectively manipulate their systems so that they "know the future" and trade in an exceedingly profitable fashion on historical data. Obviously if the system has any "hint" of the future it will most likely achieve very high profitable returns. This is the main reason why we cannot trust mere backtests from EA sellers, we always need live tests that show consistency with the backtests so that we know that the system behaves the same in real live trading conditions as it does in simulation.

Then it gets worse when systems are designed around hindsight. As an example, it has become popular to trade grid systems on the EUR/CHF or the AUD/NZD based on the hypothesis that this pairs are range bound. How do you know they are rangebound ? As a matter of fact, had you made the same hypothesis 5 years ago you would have gone through some very bad range extensions that would have wiped your account. Backtesting strategies that are based on a range that we know now but previously ignored obviously yields profit, since this range did happen. But could you have forseen that range somehow in 2000 ? 2005 ? You couldn't have. The same applies now, could you forsee the range that will be in place in 2020 ? NO ! Will there even be a range then ? As you can see, grid trading is a perfect example of a system designed around hindsight that assumes the future will be the past, nothing could be more wrong.

Hindsight is a very powerful force. When you evaluate or program either your automated or manual trading systems you should be very careful to leave hindsight aside. You should always consider that a healthy system must have no assumptions either about the past or the future, good long term profitable systems trade as market conditions change and have no assumptions about absolute ranges, past trend directions, etc. A good system trades the market as it plays out and assumes nothing about where it is going or where it has gone, it simply trades when the market shows a certain scenario it has been programmed to react to, then it takes advantage of it.

If you would like to learn more about hindsight free automated trading strategies designed to tackle varied market conditions through ATR based techniques and other adaptive tactics please consider buying my ebook on automated trading or subscribing to my weekly newsletter to receive updates and check the live and demo accounts I am running with several expert advisors. I hope you enjoyed the article !

No comments:

Daniel Fernandez's Expert Author Email Alerts
Sign up to receive email alerts of Daniel Fernandez’s latest articles from EzineArticles.com!

Email Address: