Trading strategy simulation software has tended to overlook the effect of interest rates on currency trading, as that is not something that affects many other markets.  However, for a longer-term strategy, the effect can be particularly significant, hence the ‘carry trade’.

If one were to hold a long AUDJPY position overnight, then that position would have a positive yield, or ‘carry’, overnight as the position was rolled.  This tends to give carry trade currency pairs an underlying trend, often characterized by sharp corrections, not dissimilar to the price action of a stock market.

It’s therefore vital to know if a strategy is working because of an underlying interest rate differential, as these can change dramatically over time and even invert.

Trading System Robustness

When testing a trading system, one sign of robustness is that it works across a broad range of instruments.  However, when testing a currency strategy across a broad number of currency pairs, it’s important to appreciate why it may show very different results and to really understand the similarities, as well as the differences between each currency pair.

For example, currency pairs with strong interest rate differentials are more likely to show trending characteristics.  However, other currency pairs may trend even with little interest rate differential; the underlying reason for those trends will likely be different and needs to be considered.

Therefore you have to look at the results of any simulation to determine whether there is a valid reason for the results being different, before being able to truly decide whether a system is robust or not.

One also has to look at the price action itself and compare the equity curve of the system to the price action.  It may be that the system worked particularly well in a trending market, or a sideways market, if it was mean reverting in nature.

Just because a system works on one currency pair and not another, does not mean it isn’t robust.  It may just mean that one currency pair exhibited a strong trend during the test period and the other did not and that may have been due to a shock news event such as 9/11, an underlying interest rate differential or a steadier shift in market fundamentals.


The currency markets share many similar characteristics in terms of being a global market, with similar ebbs and flows in volume and ranges, as each centre opens and close.

However, they all have their own individual characteristics in terms of time zones when news may affect that currency, liquidity, volatility and interest rates.  Even though two currency pairs may share many of these characteristics, even data releases for one currency will not always occur on the same date and time as another.  Therefore, at any given time, hardly any currency pairs are identical in nature but their differences are usually quantifiable.

Therefore, when testing trading strategies, all these factors should be taken into account to be able to determine whether a system, and any given set of parameters, is truly robust.

Caspar Marney