I found the solution as listed works well on the simulator but on the real HW I found that simply letting go of the string on the battery provided enough impulse to open the gate on the transistor on its own. So instead, I gently let go of the string and it works as designed. Alternatively switch T1 by hand back to closed position and then test the circuit as designed using the switch.
I also found that the impulse of suddenly switching on S1 generates enough torque to cause one or more components to fall over leading to a short circuit. I found holding S1 and T1 firmly at the base prevented this. This behavior I have seen on a few different circuits, and it would be good if the magnets were a bit stronger or there was a way to clamp the affected components down.