That doesn't sound a very plausible scenario.
Firstly, it assumes a crew cycling the flaps while on approach - hardly conducive to a stabilized approach.
Secondly, if they were approaching with the flaps not fully extended and hoping that they would extend somehow, then they should have been considering the flaps failed landing distance - which since they overran it seems unlikely.
Had the scenario of flaps retracting during the final stages of the approach been possible, then there wouldn't have been a need to add power - if anything, the cleaned up config would need less power. It'd also be needing more AOA, though - something you'd think might be noticeable.
Depending on the system architecture, having the handle in the wrong position might be the input to a config warning, rather than the surface itself. Either are plausible. Would need to know the system specifics to know if thats a plausible suggestion.