ibra,
Most assets perform differently due to the characteristics of the large players in the market, e.g. cable is notorious for stop hunting behaviour, because stg is not a huge currency, but the city of London, & London banks are very large and can get away with it at certain times. The carry trade is significant in some pairs, some currencies are strongly linked with commodity movements, AUD & XAU, CAD & Oil etc. Not to talk of long bias on equity indices, stocks etc., different seasonal & time of day effects.
I would not be overly concerned if a strategy only applied to a small few assets, so long as it makes sense, i.e. you can explain why it might work in one currency and not in the others. If you are trawling through strategies and happen to find one works on one currency only, without a reason why, it is much more likely to be a selection/data mining bias.
Given all the above, I'd be surprised if a strategy can be discovered that genuinely performs well across more than a handful of assets. Look at the zs: z5 only works on 1. z3 on 5. z1 & z2 with 11 are unusual, but in my time the number of assets in z1 and z2 has reduced.
