> If the axles are not physically tied, there is no way for any of the four wheels to know how much traction it actually has until it breaks free. Then, of course, it knows "too much" torque and it can apply traction control. But the damage is done. It has broken free, and any negative consequences have already happened prior to the computer being able to make any further calculations.

> Again, if physically locked together with large gears and drive shafts, there is no possible way for one wheel to break free in that manner, unless of course something is broken.

The real question is can all the wheels go to zero traction at the same moment and thus the computer thinks all is well. I think the odds of that are low enough to not worry about.