There's some flawed logic because copper resistance rises enough with temperature that current division between unequal lengths in parallel would equalize enough to prevent only one of the four from overheating to that degree. There most certainly was some other factor causing that failure....
Jehu's video where he had a wire in a small parallel bundle completely burn as it was the shortest, and took the majority of the current (least resistance). ...
Also I fail to see why someone would have made a multistrand cable from different lengths. It didn't appear they actually took a close look or dissected the crimps. Maybe all 4 strands were equal and all but the burnt one had pulled loose out of the crimped lug leaving only the "short" one to carry all current.
But copper parallels very well due to its positive resistivity coefficient. Lenght tolerance isn't normally a factor for wires and bars.