Saturday, February 01, 2014

The Pareto Principle

In the early 1900's, Italian economist Vilfredo Pareto created a mathematical formula to describe the unequal distribution of wealth in his country, observing that twenty percent of the people owned eighty percent of the wealth. Since then, this basic principle has been applied to many aspects of life, business, management etc.

I even unknowingly referred to this principle in my e-book, in which I stated the belief that 80% of all trades taken would cancel each other out - small losses and small losses. It was the remaining 20% of the trades that generated the overall profit.

In his Market Wizards interview, Richard Dennis expressed an even more extreme example of the principle - that 95% of his profits came from just 5% of his trades.

Below is the latest distribution chart of all trades recorded here (click on image to enlarge). This sort of distribution is typical of a trend following system:

I then reviewed the 270 trades in that sample, and sorted them into the orders of the returns achieved (expressed in R) to see if the principle held up. The actual results almost exactly mirrored Dennis's experience - rounded to the nearest percentage point, 6% of all trades taken accounted for 94% of all profits.

In basic terms, out of 270 trades, just 16 of them accounted for virtually all of the profits (+145R) made over the 19 months.

This is a difficult concept for many traders to acknowledge or appreciate. They prefer to seek sanctuary in a high win rate, and take lots of small gains. The danger with that is, when a loss takes hold, they wait for it to move back into a small gain position before closing. How far do they let a trade go into a loss in the hope it turns round? This is how many of the automated forex systems touted around are engineered - they work until a big loss comes along, causing a major drawdown or even an equity blow-up.

I have seen this change from the original 80:20 split evolve, by keeping track of the performance metrics. Since applying the changes to my stop methodology following the poor performance in 2011, the overall win percentage achieved has dropped. In contrast, the profit factor has increased - again expressed in terms of R, currently the size of the average win is just over 4.8 times bigger than the size of the average loss. 

What is basically means is that overall profitability of the system has improved, even with a falling win percentage.  This has been achieved by cutting losing trades more aggressively than before. So, the number of losing trades has increased, but the magnitude of those losses has reduced. Remember that the expectancy of system is based on both the win percentage and the average win and loss sizes.

So what conclusions can you take from this?
  • Once again, this is a stark reminder as to why good risk control is so important in achieving longevity and success in the markets;
  • It also reinforces the need to avoid cutting winners short, for fear of losing those profits, as it is the big profits made that generate the overall positive expectancy;
  • It forces the trader to accept that he will incur many losses, and discipline in keeping them as small as possible;
  • It also explains why, if you are unable to accept this sort of trade distribution, that trend following is not for you.

No comments:

Post a Comment