Those who have followed the discussion on the Conquest 441 accident at Renmark would note that much of the comment revolves around the possibility of a practice engine failure after take off gone wrong.
The following link sent by a reader on the thread is a relevant ATSB report on a practice engine failure in a Beech 1900.
https://www.atsb.gov.au/media/24342/...000492_001.pdf
I commend it to all multi-engine aircraft flying school instructors even though it deals with practice turbo-prop engine failures and not those on light piston twins typically used by flying schools.
It is probable that most flying schools engaged in initial twin engine endorsement training use a standard mantra of mixture up - pitch up- throttles up - gear up - flap up - identify failed engine by dead side dead leg - confirm with throttle - and finally feather.
All the time the simulated failed engine throttle is closed hard against the mechanical stop - or if the mixture lever was used to cut the engine - the throttles are both still forward at take off position. Either way, the propeller of the "failed" engine is windmilling and producing significant drag until the pilot gets through the engine failure mantra and finally either feathers the prop (assuming a real engine failure) or the instructor sets the throttle and pitch lever to a guesstimate of zero thrust.
With a windmilling propeller, airspeed will decrease unless the pilot deliberately lowers the nose and loses height in order to maintain safe single engine flying speed. He will probably be criticised for deliberately losing height to maintain airspeed especially if he is under the hood in simulated IMC. In other words he can't win
It could be upwards of 15 seconds going through the engine failure mantra until he gets around to "feather." All the while the "failed" engine is producing lots of drag. See the ATSB report. It may not be as severe as in a turbo-prop simulated failure but nevertheless it is bad enough to cause rapid speed loss. All this below 500 feet means the risk of loss of directional control is substantially increased as long as the prop windmills.
To mitigate (don't you just love that buzz-word) that risk or manage that "threat" as in Threat and Error Management (love that buzz-word too), CASA note
perhaps it would be safer, if instead of closing the throttle against the stops to simulate engine failure, the instructor would reduce the throttle to the approximate zero thrust position on the quadrant and announce "simulated engine failure."
There is no need to increase the risk of mishandling by allowing the huge drag from a windmilling propeller to ruin your whole day while you get around to muttering the mantra and eventually get around to setting zero thrust.
There must be a limit to faking realism in asymmetric training. The current acceptance of reducing airspeed causing increased yaw and drag associated with failure to set the throttle to zero thrust to simulate engine failure, needs to be reviewed.