Bookworm, I've been following this thread since leaving these two people to their fun, cos I find the name-calling, patronising assumptions and so on to be a bit dull. I mean, I already went to high school, you know? But I thought their might some pearls of wisdom to be salvaged, or wheat amongst the chaff. (Choose your metaphor
And there are.
To your specific question re VMCA vs VMCG: "If the aerodynamic forces after an engine failure at that particular speed will roll the aircraft after its wheels have left the ground, why do you think that you will be spared if the wheels are still on the runway? "...well, I wonder if there may be a pretty simple explanation.
It seems to me that you'd much less likely to roll if you're still on the ground, with a high speed engine failure that occurs before rotation, primarily because the "wheels on the runway" act as a stabilising factor against the development of a roll, or to be precise, the wheel that is under the failed engine, and therefore holding up that wing to some degree. Sure, you'll yaw, although nosewheel steering & differential braking may control that to some degree as you retard the working engine...but when on the ground, you have the the main gear to at least partially oppose a rolling motion, no? When practicing high speed engine failures, yes, with a very experienced instructor, I recall a startlingly fast yaw, until the good engine was retarded - but no hint of a roll. And even if a roll did occur and overpower the stability of having wheels on the ground, if you're still on the ground you have a wingtip to hit the ground first, which means that it could be ugly - but probably not as ugly as ending up completely inverted when airborne, where you'll presumably have a serious ROD, even from 20 feet or so, to compound the problem. So the most likely worst case, I would guess, would be departing from the runway - but right side up, and rapidly regaining yaw control as the power from the working engine winds down. Provided you cut power immediately, of course.
Once you're in the air, you've got no wheel on the ground under the failed engine to act as a stabilizing factor, and oppose the rolling motion, so a roll, it seems to me, would be far more likely to occur. And once the roll gets going, it could indeed, perhaps, be difficult to stop given the minimal effectiveness of controls at that speed.
Could it be that simple? The fact that the standard T/O technique is to holding a twin on the ground until VMC or faster would imply so, dontcha reckon?