IF ONLY; two small words which are spoken by thousands of traders and investors every week. ‘If only I had cut that position’; ‘If only I had bought some shares of Moonbound Mining’. Many traders are blessed with 20/20 Hindsight vision and hindsight will always be the scourge of traders. In this article I will approach hindsight from a couple of angles and hopefully offer some techniques to help reduce the effect that hindsight undoubtedly has on trading.

Unfortunately hindsight is also used by many ‘educators’, financial planners and trading system vendors to substantiate their claims and to ‘prove’ certain facts. For example we have all been told that if we had bought stocks say 20 years ago we would have made X amount of money. But in reality it depends upon which stocks we bought. The Australian Stock Market is for example, at its all time high but this does not mean that all stocks are trading on or even near their high. Ansett Airlines, formerly a major airline has gone bankrupt as has HIH Insurance, previously one of Australia’s most prominent Insurance firms. Shares in Telstra and AMP, two more blue chip stocks, and probably the most widely held stocks in Australia have fallen by 50% or more over the past 4 years or so. Even during the biggest bull run that mining shares have witnessed in almost a generation, Sons of Gwalia, a 100 year old, well established and known, predominantly Gold mining firm has gone bankrupt. It is probable that many investors in Australia have held at least one of these stocks and so the reality is that when financial planners hold up a chart of the performance of the ASX 200, many or most investors will not have matched that performance. In years to come though, the stocks that no longer exist will not show up and people will assume that they can get the returns their financial planner suggests.

This point actually hits upon another bias, called survivorship bias which is the view that because we only see the winners or survivors our view of the performance of whatever we are looking at will be distorted. Survivorship bias can therefore be seen to relate to hindsight as well.

In Jan/Feb 2005 one of Australia’s leading technical analysts wrote the following statement about Aristocrat Leisure which was trading at $8.59 AUD:

“Investors who climbed aboard after the dramatic fall from the $7.30 high on 31/8/01 to the 76 cents low on 30/5/03 have been well rewarded.”

This statement is a clear example of someone using hindsight to try to appear to be smart. The fact is that when this company’s share price was trading at 76 cents it looked like the company was going bankrupt and those who bought shares at this time have been rewarded for taking a big risk. Sons of Gwalia, HIH Insurance, Enron and WorldCom shares fell sharply too but those who bought these at low levels were not rewarded and because they no longer exist our technical analyst will not be reviewing them. I am sure though if she were presented with a graph of these latter stocks and if she knew which they were she would argue that they were clearly in a downtrend and should be avoided. The fact is though that the graphs of these latter stocks and that of Aristocrat Leisure were very similar and it is only with hindsight that we can try and find some differences. It is vitally important as traders that we understand this so that we can avoid the claims and views of people who make such statements.

After the fact, traders, particularly technical and mechanical ones can examine the past data or graph and apply the multitude of indicators, patterns etc that are available and find out which one(s) worked best over the preceding period or which ones ‘predicted’ certain moves. This approach and other back testing techniques are said by their supporters to filter out the techniques which ‘work’ on a contract. However those who follow this approach are actually more likely to be practicing data snooping. Heavily influenced by hindsight they can easily discard poor performing indicators and find the ones which have performed well in the past but using these will only be successful in the future if the future exactly resembles the past; almost an impossibility. The question we need to ask ourselves is, ‘Could we have found these winning indicators before the period that we tested them on?’ Otherwise the work that we are doing is of little use and we are relying on hindsight to establish our trading techniques.

While researching my book, ‘Technical Analysis and the Active Trader’, I discovered that this question has been posed and tested by some very thorough studies. In all the studies that I found the answer was the same; a trader could not have known before the period which indicator would perform the best in the future. I find these studies among the most interesting that I examine in my book but until now they have mostly remained within academic circles. The studies are far more thorough and robust than any system vendor would conduct and we should take note of their results. For example Professors Sullivan, Timmerman and White in their paper for the London School of Economics Financial Markets Group entitled ‘Data Snooping, Technical Rule Performance and the Bootstrap’ tested nearly 8,000 technical rules on 100 years worth of data on the Dow Jones Industrial Index. One test they conducted was to use only available information to construct a trading strategy; that is at any time they first found out which tool or indicator (out of the nearly 8000) had performed the best until that point and then used that one for future trades. They found that this approach performed poorly and concluded, “…investors could not have known ex-ante the identity of the ex-post best performing trading rule”.

So there is significant evidence that this back testing approach which has become very popular among private traders and has made many a trading software vendor wealthy, is no where near as reliable as those who use it believe. It is effectively a hindsight and data snooping based technique which has the added concern of making its followers over-confident in their trading. It is likely that most who use these techniques will fail even though they believe they have a ‘proven’ system.

What I find particularly amazing is that in many countries and on most financial products there must appear a statement along the lines of ‘Past performance is not necessarily a representation of future performance’. There is a good reason for this, it is true. Yet systems which are ‘proven’ using past performance are becoming more and more popular among private traders partly due to their failure to understand the influence of hindsight in the results (as well as slick marketing).