Being English (although I am now a Naturalised Aussie too) I am a huge soccer fan. This season my club Liverpool have enjoyed a wonderful run of success in the European Champions League, the most prestigious club competition in the world and they will play in the final which takes place before this edition hits the newsagents. Imagine my glee therefore when I was sent the following statistics in an email before the Quarter-finals which took place a few weeks ago:
LIVERPOOL FC ** Too much of a coincidence?? You decide.
1978 – Welsh Grand Slam. Pope Dies. Liverpool win European Cup
2005 – Welsh Grand Slam. Pope Dies. ???
1981 – Prince Charles gets married. Ken and Deirdre marry on Coronation Street. A new Dr Who is appointed. Liverpool win European Cup
2005 – Prince Charles gets married. Ken and Deirdre marry on Coronation Street. A new Dr Who is appointed. Liverpool win European Cup??
An amazing set of coincidences and I hope we will see the happy ending of Liverpool being crowned European Champions. Actually to add further intrigue because both Liverpool and their opponents in the final, AC Milan, wear red in their strip there was a coin toss to decide who would wear their red strip in the final. It was decided that Liverpool would wear red and Milan white, which introduced yet more coincidences. Liverpool had previously won 4 European Cups, always wearing red and always against opposition wearing white!
While a few friends of mine headed straight to the bookmakers on receipt of the first set of information, and may well make money from their trade, I expect that most readers would see that there are no real correlations in these events. Of course we can look back over history and find events which have coincided on perhaps a number of occasions but few of us really believe that there are any links between Welsh Rugby, Prince Charles? marital situation, a popular television soap opera and Liverpool Football Club. So we will view these events as coincidences and nothing more and we would not develop a betting strategy based on this research. Furthermore if my friends do make money should Liverpool win the cup, we would all admit that their profits were not the result of their analysis but purely due to Liverpool playing well in all or most of their matches. In fact, as Liverpool have had to beat Italy?s best team and the runaway champions of England just to reach the final we could argue that just getting to the final was actually a very unlikely event for Liverpool.
Application in Trading
So how does this relate to trading? Well, the type of analysis that I have been describing is a form of the representative bias which I explained in past editions of YTE. To make decision making easier, some people estimate the likelihood of something happening based on how closely it resembles something they have seen in the past. Hence they try to look for patterns and correlation in events which may be very different.
Unlike our sporting example, though, with the markets it is far more difficult to show that events which look the same may not be correlated. Plausible stories are created which explain the ?psychology? of chart patterns and other technical trading tools and how that ?psychology? is seen to repeat. Yet the reality is that our markets are far more complex than this and I explained in my earlier articles how, when studied, chart patterns and technical tools and indicators more often than not show poor reliability.
In fact, in August 2003 US Federal Reserve Chairman Alan Greenspan warned bankers precisely against looking for repeating patterns, trends and correlations. While summarizing Greenspan?s speech in his book ?Bull?s Eye Investing? John Mauldin writes, ?It is as if we see that Trend A and Trend D seem to move in lock-step and draw a conclusion about the connection. But hidden, and not in the model, are Trends B and C, which are the real connection between A and D.? Furthermore, none of us may even know what Trends B and C are and these relationships may constantly be changing. So any conclusions that we might reach will only be huge simplifications or rules of thumb; any profits which might be made by someone thinking in this way may be despite the analysis not because of it.
Representation taken further
Among other things, Greenspan was warning the industry against using representativeness. Not only does this heuristic affect chartists but there are an increasing number of mechanical and quantitative traders using similar approaches at banks and hedge funds. In addition the Value-at-Risk, risk management system which many if not most banks use, also contains an element of this flawed way of thinking.
One reason why Greenspan is so worried by this way of thinking is that it leads to over-confidence which leads to poor trading decisions and in certain situations can be destabilizing to our markets. LTCM would be one example of traders who were over-confident because they had meticulously studied past correlations and performance. For LTCM not only did they run the risk that the past correlations were not as clear as they had seemed, but they had not weighed up the magnitude of the situations when their trades went wrong.
Good trading requires more than pattern recognition or studying past data and applying technical tools to them: anyone can do these and there is little or no edge for traders who use these techniques. Good trading decisions should involve an understanding of context and good traders will be able to use this information and weigh up possible future outcomes, outcomes which obviously may not be on the graph or in the past data.
In fact, a recent study by Acker and Duck highlighted a further problem with studying correlations. They found that if different reference days are used for correlation studies the results can be very different. For example when studying the degree of correlation between monthly returns of the UK and US stock markets between 1998 and 2002, the authors found that if one reference day (day 12 in the month) was used the correlation was approximately 70%. But if day 23 had been selected then the correlation between the two markets was 90%. So what conclusion can we draw about the correlation of the two markets? The answer is very little and a trader who used say the 70% figure because that happened to be the result on the day he tested would, if profitable, claim that the profits were a result of his analysis. Yet another trader may have performed similar analysis yet just on a different day and he/she may have lost money. Survivorship bias means that we will not see or hear from the second trader but the first may go on to start up his/her own trading courses to teach others how simple analysis can lead to success.
At the beginning of this year I am sure we all read about how years ending in 5 are usually bullish ones for equities and we hear similar stories about performance in pre-election years and post-election years. Often plausible stories are created for example how pre-election years are bullish because the incumbent leader creates a bullish set of circumstances to help him/her achieve re-election. However the cause and effect in these situations is far from so clear cut and what is really happening is that people will look at the performance of markets in say pre-election years, find that more often than not stocks rise, and then create the plausible sounding story. Like some financial journalists, they are always looking for a story behind a move, when there may not even be one. Decisions based on this type of analysis should have no place in a trader?s armoury and certainly if any trader that has ever worked for me had used such explanations for a position he/she would find themselves on breakfast duty for a considerable period.
I can assure readers that my views are in no way premeditated against the techniques that I don?t like, they are the result of working with and witnessing at first hand, hundreds of traders. I have not read books by these people, I have not attended seminars by these people, I have watched them trade and traded with and against them. I can also state that being a good trader is not necessarily about intelligence or intellect; one of the best traders that I have encountered left school at 16 with few qualifications whereas two of the most highly qualified and most highly intelligent individuals that I have worked with made extremely poor decisions, mainly using mechanical and quantitative methods based on their math and programming skills. It is traders such as these that Greenspan was referring too.
And if Liverpool have won the European Cup I hope I don?t have to wait until the next time Wales win the rugby Grand Slam for them to win it again.
© YourTradingEdge Magazine, all rights reserved 2005. This article is reproduced with the permission of the publisher, and originally appeared in the July/August 2005 edition of YourTradingEdge