As I said, it's only offered as food for
thought.
Should have explained properly, the text is part of a briefing related to exactly what Chimbu was talking about, the art of 'nutting it out' before blasting off. That is; OK, it's short, it's hot and old girl is heavy, so if one quits just as I get to the trees, how do I get home for tea with Mum??.
Using rotate speeds of 90 knots and the gear retracting, the aircraft is certainly in a position to rapidly accelerate to 95 knots. On a minimum length strip, the chances of stopping from 85 knots without damage are limited. The chances of accelerating from 90 to 106 are definitely limited without planning and forethought.
There are no available details for OEI 'second segment' acceleration. The difference in acceleration capability between that available on two engines, compared to one engine provides an element of uncertainty. The manufacturers provide detail of the expected single engine climb performance. The CAO require a 1% gradient at all heights up to 5000 ft in ISA conditions. If the aircraft cannot climb below Vyse, then, the aircraft must have room to safely stop after a failure or a "suitable" OEI landing area accessible.
It is most hazardous to attempt, at tree top height, an acceleration segment from below Vsse to a SEBAC (Vxse) speed. On a hot, turbulent day, attempting to feather and secure an engine, whilst achieving a correct asymmetric balance and attempting to find performance for acceleration is difficult from a strip which only allowed acceleration to 80 knots before the DER.
The manufacturers indicate that in high ambient temperatures, to achieve adequate engine cooling in the critical case, the rate of climb may require a reduction of 50 fpm, (an increase in speed).
PA 31 example. At 3175 kg the expected rate of climb is 200 fpm at 25°c. At a Ground speed of 106 knots (still air) and a 50 fpm deficit, the achieved gradient is 1.4 % climb. To achieve a circuit height of 1000 ft AGL there needs to be ::11 miles of obstacle free gradient, straight ahead! (1000 x 100 / 1.4%). OC gradients are based on a take off surface splay (safe area) distance limit of 3000 metres or 7500 metres.
I think the point being made is that without a good knowledge of what you have to work with, the laws of physics and dumb luck will only get you so far. Murphy rules. The boys discussing the SE turn back are saying the same thing - know your aircraft, know the limits and think about it. That can't be bad.