A very fair point and in fact one I considered bringing up. In fact, if I am not mistaken, the defeat suffered at the hands of the Cimbri you mentions was one of the most decisive, and costly (in terms of casualties) in the entire history of Rome.
Even with Arausio, do not forget that the Romans were MASSIVELY outnumbered. I am not entirely sure they were "marian style" troops. If I am not mistaken, Marius was "allowed" raise legions that drew from the head count (ie. landless), but as far as i know, this was not adopted immediately. Marius raised his legion in 107bc, battle was in 105bc. Aruasio troops may well have been the pre-marian type troops.
This was not the only diaster, so, in any event....
I convinced myself that leadership was not a significant issue because leadership is always important (“ten men wisely led will beat a hundred without a head”…sorry couldn’t resist). From my perspective, it is difficult to take into account the influence a good leader would have on his troops when modding EB. For instance, the example you mentioned re: Marius and Caepios + Caepio (Arausio). BUT if you take into account that they were well armoured, typically/comparatively well trained (due to the Roman Infantry Tradition), and professional; and if you compare them to the infantry fielded by other states around them (taking leadership out of the question), man-for-man they were quite superior.
From that time period, the vast majority of defeats were caused by almost purely by bad leadership. I also agree that the opposite is true. The great victories were due to great leadership. But if you look at when they went toe to toe with other heavy infantry, more often than not, they were quite successful. Arausio was a disaster because the 2 leaders led 2 separate armies and would not even camp together due to the fact that one of them was a pleb and other a patrician.
I completely agree that there was no standing army, and no standardisation of training. And these are very good points. For these reasons I am inclined to agree that they should be slightly weaker than imperial cohorts, but no pushovers themselves.
Bookmarks