I think that the only way you can make a worthy comparison like this is to simultaneously analyse an identical longer timeframe for all currency pairs considered. Then you can select an appropriate weighting to give to any one currency based on it's performance under your system over an extended period of time.
I don't think you can have confidence in your changes if you make them only looking at a small 'bad' period of time - what are the repercussions on the total system over the whole test period - try to look at 3 or 4 years worth of data.
Also are you confortable changing your balance from being 50%GBP, 50%USD to something more like (from eg1) 30%GBP, 30%USD, 20%JPY, 10%AUD, 10%EUR to try to fix a single drawdown period. What does this do to your entire system in the long term? Does it provide better % return, less drawdowns, smoother equity curve, sensible diversification - are you spreading yourself too thin over too many markets?
Would 33%GBP, 33%EUR, 33%JPY by using GBP/USD, GBP/JPY, EUR/USD, EUR/JPY be more appropriate by giving no one currency any more influence on your total return than another? Not all currency pairs will be profitable all the time.
There are many things to consider with diversification - I am no expert on what they all are. But my advice is to think long term. One or even six months of flat or negative return on one currency pair is insiginificant on it's own when you are analysing 4 years worth of trades with a properly balanced, risk managed system.
I don't think you can have confidence in your changes if you make them only looking at a small 'bad' period of time - what are the repercussions on the total system over the whole test period - try to look at 3 or 4 years worth of data.
Also are you confortable changing your balance from being 50%GBP, 50%USD to something more like (from eg1) 30%GBP, 30%USD, 20%JPY, 10%AUD, 10%EUR to try to fix a single drawdown period. What does this do to your entire system in the long term? Does it provide better % return, less drawdowns, smoother equity curve, sensible diversification - are you spreading yourself too thin over too many markets?
Would 33%GBP, 33%EUR, 33%JPY by using GBP/USD, GBP/JPY, EUR/USD, EUR/JPY be more appropriate by giving no one currency any more influence on your total return than another? Not all currency pairs will be profitable all the time.
There are many things to consider with diversification - I am no expert on what they all are. But my advice is to think long term. One or even six months of flat or negative return on one currency pair is insiginificant on it's own when you are analysing 4 years worth of trades with a properly balanced, risk managed system.