If I understand correctly, you're implying that an extended choppy environment will inevitably drag the robot into a losing territory, yet in the long run, the mix of trends vs chops is such that the robot cannot afford to be left out of a position when the big swings materialize. It strikes me that the most recent trade could have been quite profitable already (rather than criss-cross above and below the entry price) if it were to have taken multiple ATR-size profits off the entry limits (e.g. 73.14->75.00 =~ +2.5%) but that would have required the robot to change its plan (it simply doesn't know the stats associated with that scenario). As you say, statistically there's that big swing chance awaiting that will dwarf all these small losses to chop so sticking to the robot plan, as tested, is the best course of trading action.
Trader D