Minuteman,
(a) Turning away from an obstacle is an option .. but not a requirement. Consider that the radius of turn complicates the analysis and there is a climb gradient decrement (usually somewhere in the vicinity of 0.6 percent for a 15 degree bank turn). In many cases there is just not enough room to lay out the turn prior to an obstacle and the only way out is to constrain weight enough to clear it vertically. I can remember one difficult runway where I spent some considerable time trying umpteen different (and variously interesting) escape tracks before it became apparent that the various penalties ended up predicating a straight ahead climb over the obstacle as being the best payload option ....
(b) if the aircraft is anywhere near WAT limits, you will see a measurable distance penalty if you drive the second segment higher .. no hard and fast rule as to where it becomes worse off for the fourth segment .. just a matter of doing the sums until one gets a good enough feel for the particular AFM to be able to foretell the story in advance. As the second segment gradient is not constant .. as the aircraft goes up .. the gradient capability goes down ... with the very shallow gradients we are looking at .. a small gradient penalty for a given third segment height can result in loads more distance run.
(c) If clearance was compromised, no operator would consider accomplishing drills in the second segment! ... you exhibit a refreshing degree of idealism, good sir ......