At the start I was using systems with a discretionary overlay to try and improve results. Always underperformed the system, so this year I have moved to 100% systematic with much better performance.
Instead of the sharpe, you might want to use the sortino ratio, which uses a downside deviation instead of standard deviation, since in general upside volatility is something you want.
tiocsti i am conservative and assume anyone can loose what he can gain. even if an equity curve shows strong difference between upside and downside vola i assume they are the same. this is just from experience that sooner or later every upward distorted equity curve distorts down as well. i am aware that "cut losses, let profits run" speaks against that, but especially trend followers, who literally live from this rule, do not show distorted vola over long periods IMHO. nevertheless sortino ratio, which divides return by the standard deviation of down observations, and mar ratio, dividing return by draw down, are return/rik figures and therefore better than return as such. to me the whole zoo of ratios is to some extent part of what i call financial marketing ... finally it is a question of taste maybe. peace