An important subject and one needing lots of airing in places such as this. May I summarise the basics as discussed so far ...
(a) performance can be approached theoretically, in which case the numbers are all good fun and can be run to whatever accuracy is considered pleasing to the eye. Alternatively, one can look at the real world and note that the distribution curve shows up a degree of variability in achieved performance when compared to calculated. He who blindly puts all faith in charts beyond a reasonable scatter accuracy lives with the potential for excitement .. ie a little bit for mum and the kids in usually a good idea.
(b) the discussion so far shows clearly the variability in knowledge within the Industry on matters technical.
(c) the line pilot generally will not know whether the go or stop case is limiting or closer to being limiting for the particular takeoff on the day in question. Consider quizzing your ops engineering folk for runways which you might consider to be more exciting than others.
(d) Vmcg is a real limit and can make for exciting times if abused. Consider that the transition from easy directional control to loss of control may manifest itself in the space of several knots and the rate of onset of control problems is fast as the failure speed decreases.
Keep in mind that real world Vmcg (as opposed to the certification animal) is very crosswind dependent (typically 0.5 kt/kt for twins increasing to something in excess of 1 kt/kt for 4-engined beasties). This is made more interesting when one considers that the certification value usually is based on zero wind (7 kts under the old British standards). The certification value is there more for line in the sand purposes but, if you happen to be near or at the limit on the day, that doesn't help you much.
Taking an arbitrary decision to reduce the go decision to a speed significantly below V1 may be hazardous, especially for low weight, low altitude, low OAT takeoffs. Keep in mind that our simulator experience is a two-edged sword and that the great majority of simulator failure training exercises are done under non-limiting handling conditions .. this can create an unjustified comfort zone .. this is especially the case for aircraft where Vmcg is not usually a critical takeoff limitation.
(e) history suggests that more people have trouble with high speed aborts than they do with high speed continued takeoffs.
(f) current thinking (and this has been the case for quite some years) is that if you haven't initiated the stop BY V1 (ie the first abort action has been executed - normally brakes full on) then you keep going (unless there is a real perception that the aircraft will not fly). Particularly with older certifications initiating the stop after "deciding" at V1 (if the ASDA is, in fact, limiting) might be a bit worrying. Keep in mind that, in the real world, the bit of the runway on which you are hoping to stop is the bit on which umpteen thousands of tyres have left a thick film of rubber over the past however many months/years since it was all cleaned off .....
(g) the pilot who doesn't monitor his own instruments during the takeoff is asking for a surprise sooner or later.
(h) there is some fat in accel-stop scheduling (which varies with certification basis) but not a great deal (amendment 42 sought to redress this problem to some extent).
(i) the calculations should be made for the ambient conditions, not standard day. An earlier poster's reference to WAT conditions left me a little confused .. perhaps he might add some additional commentary ?
(j) being able to justify one's actions at the investigation is an important part of the plan. If you think you won't be able to, then maybe you ought not to be about to do whatever it is you are contemplating ... ?