In any case, with the usual jet pitch attitude, I suggest that it is folly to attempt to eyeball a critical departure.
I was never suggesting that a performance A aircraft would be required to visually avoid obstacles in the take-off funnel.
What I was trying to explain is that for example in the case of a departure with a track change of more than 15 degrees, the maximum width of the take-off funnel is 1200m if the pilot can acurately maintain the requried track and 1800m if they can't. In the absence of suitable navigation aids, a limitation of VMC can give the pilot the ability to maintain the required track visually so that the funnel is smaller.
Imagin a departure along a tight valley with a 30 degree turn 1nm after the runway end. No navigation aids. If IMC is permitted then the funnel expands until it is 1800m wide - taking in quite a bit of the hills each side of the valley. If the procedure is limited to VMC then the pilot can (a) determine the 1 mile point if there is a good landmark and accurately track the valley centerline thus reducing the final size of the funnel and the height of the obstacles that are in it.
Hope that explains my point better.
Agree with the statistical situation but ensuring that the net requirments are met makes it very unlikely that the gross performance will be less than the net - assuming all factors are as planned of course!!