Another good question. One unit in a stack deals 1-2 damage; a hundred of the same units will deal damage equal to the sum of all units, calculated separately (judging by the thread about the damage calculation program, L&M). This means that the spread (statistically) will be much smaller. As a result, it doesn't make sense to divide a hundred archers into 2-6 stacks (I used to do this to reduce the spread). Is this correct?

This is still an open question.
I tried to verify this when playing as the Dungeon faction, but after gathering some information that was sufficient for me, I stopped.

If the damage of N units was the sum of the damage of each unit, then 10,000 troglodytes would almost always deal (without modifiers) 20,000 damage, and the standard deviation would be 81.

However, it has been established that the damage per troglodyte is always a multiple of 0.1 (one-tenth) and in 90% of cases is between 1.7 and 2.3, regardless of whether there are 100 or 1000 troglodytes in the stack.

So, the physics here are clearly different. Troglodytes fight coherently :D

And, of course, there is no coherence between stacks. So, it turns out that dividing a stack may be beneficial in this sense to reduce the spread...