Foreign exchange (Forex) trading algorithms require trading strategies involving the autonomous identification and exploitation of trading opportunities. Trading strategies are generally implemented in isolation, raising the question of whether diverse trading strategies can be combined in order to improve trading performance. It remains unclear, however, whether such a trading ensemble approach is capable of adapting to different trading styles based on predefined trading goals. Our contribution in this paper is, therefore, to evaluate the robustness of the notion of Forex trading ensembling in respect of different trading preferences when ensembling different trading strategies and when creating trading portfolios. An empirical analysis is conducted on four different Forex currency pairs during which five different trading strategies are ensembled and asset portfolios are created from the perspective of risk-averse and risk-seeking trading styles. The results indicate that an ensemble design is indeed capable of adjusting trading decisions based on these two defined trading styles. Ensembling from a risk-averse perspective reduces risk exposure while risk-seeking ensembling favours return maximisation with less emphasis on trading risk.