A Reddit thread from a multi-engine trainee flying a Beechcraft Baron surfaces a training issue that instructors and examiners see constantly: mishandling the recovery phase of the Vmc demonstration. The pilot describes two failure modes. In the first, retarding the operative engine to idle and simultaneously releasing rudder input causes the airplane to yaw abruptly during the pitch-down. In the second, holding rudder while reducing power and pitching down causes the aircraft to roll or "dip" the opposite direction. Both symptoms point to the same root problem: the recovery inputs are not being sequenced and coordinated correctly, and the pilot is likely mismanaging one control input while overcorrecting with another.
The Vmc demo is one of the more misunderstood maneuvers in multi-engine training, precisely because it is often taught as a memorized control sequence rather than an aerodynamic concept. The correct recovery — reduce power on the operating engine, decrease angle of attack (pitch down) to reduce AOA and restore aileron/rudder authority, and simultaneously roll toward the operative engine while maintaining directional control with rudder — has to happen essentially together, not as isolated sequential steps. When a pilot yanks power to idle and lets go of rudder at the same instant, the airplane still has residual asymmetric thrust and P-factor effects at the stalling AOA, and the sudden change in yawing moment as rudder effectiveness returns can cause a snap in the opposite direction. Conversely, holding rudder too long while pitching down and reducing power creates an over-yawed condition where the rudder is now producing more side force than needed, causing an opposite excursion in roll or heading once the aircraft accelerates. The "dip" the pilot describes on the second technique is almost certainly this — rudder held past its needed authority as speed builds and AOA decreases, causing the airplane to skid or roll away from the intended recovery heading.
For working pilots, particularly those who came up through the CFI/multi-engine instructor pipeline, this thread is a reminder of why Vmc demos have such a fraught safety record in general aviation flight training. Vmc rolls — the uncommanded, often unrecoverable roll that occurs when a light twin is flown at or below Vmc with asymmetric power — have caused numerous fatal training accidents over the decades, which is why the FAA revised the ACS to emphasize the demonstration of the onset of loss of directional control rather than a full stall/Vmc event, and why many training providers now favor performing the maneuver at higher altitudes with a more conservative recovery trigger (first indication of loss of directional control or stall warning, whichever occurs first). The maneuver sits at the intersection of aerodynamics, muscle memory, and startle response, and pilots who treat the recovery as "power to idle, then pitch, then roll" in a rigid sequence rather than as coordinated, simultaneous corrections are the ones most likely to experience exactly the kind of secondary yaw/roll excursions described in this post.
More broadly, the thread reflects a persistent gap in multi-engine and complex aircraft training: many CFIs teach the mechanics of the Vmc demo without deeply explaining the underlying aerodynamics of why each control input matters and how they interact dynamically. This is relevant well beyond the Baron trainer world — turboprop and light jet pilots transitioning from single-engine backgrounds, as well as those returning to multi-engine currency after time away, often carry the same misconceptions about sequencing power, pitch, and rudder during an engine-failure-at-low-speed scenario. Operators running Part 135 or Part 91K multi-engine fleets, and flight schools producing the next generation of airline-bound CFIs, have a vested interest in ensuring Vmc and engine-out training emphasizes coordinated aerodynamic recovery rather than rote control movements, since the instincts built during initial multi-engine training tend to persist — for better or worse — into every subsequent asymmetric-thrust emergency a pilot will face in their career.