Do More of What Works and Less of What Doesn’t

The admonition that traders ought to let profitable bets run while cutting losing one’s quickly has long since entered the realm of the cliché. It is sound advice, nevertheless. Applying such thinking oftentimes distils to trade construction: you should have a better than 1:1 risk to reward ratio, we are taught over and over again.

Rightfully so. It seems odd to risk losing more in order to gain less. Such a consideration also lowers the burden of getting it right on any given trade. Consider: with a 1:2 risk to reward ratio – a stop-loss distance of 10 points and a target of 20 points, say – being right on just one out of every three trades breaks even, and anything better is profitable.

Nevertheless, overlooking this part of the process is one of the most common trading pitfalls. Investors routinely find themselves in situations where a string of modestly profitable trades is wiped out by one or two big losses. A difference of perspective is perhaps in order.

This year, the routine annual review of my own performance seemed to suggest an interesting alternative. The headline statistics show a net return of 40.6 percent on 58 trades in 2021, of which 40 percent were profitable. That is, account equity grew by over 40 percent even as 6 out of 10 trades lost money.

How did this happen? The average winning trade made about 2.5 times more money than the average losing one gave up. Intriguingly, the average winning trade was also held for 4.7 weeks, while the average losing one was open for just 1.7 weeks. In other words, winners were held 2.8 times longer than losers.

Engineering a favorable risk to reward tilt seems more intuitive from this time-based angle: do more of what works and less of what doesn’t. Practically, this might mean that an already profitable trade deserves a bit of room to develop, or even merit scaling up. By contrast, a trade struggling to get above water may be cut quickly.

Be the first to comment

Leave a Reply

Your email address will not be published.


*