Nastyogre,
I'll stick with the big-then-small approach, for the following reason. Say you have a 'Mech with 16 points of front CT armour. Say you fire a Gauss Rifle and an SRM-6 at it; they both hit the CT location, and 4 of the SRMs make it.
if you do littles first, you end up reducing the armor from 16 to 14, 12, 10, 8, then -7, with one chance of critical.
If you do bigs first, then you go from 16 to 1, -1, -3, -5, -7 - four chances of critical.
While that's somewhat artificial, it does indicate the trend - the larger number of small hits produce a greater chance of making a critical in a hole made by the bigger weapons. Yes, when (say) firing ER lasers and LRMs - which produce middle-sized damage groupings - it doesn't tend to matter as much, especially when firing at heavy armour. But as a rule of thumb, big before little tends to pay off.
W.