I have been seeing differences in troop losses from identical attacks, here's what I've found attacking 10 lvl 6 pict camps with:
Lvl 8 poison edge
Lvl 7 healing potions
Lvl 80 knight
5000 swords + 100 wagons
Each attack was progressively further away with a total of 5100 troops here are the survival numbers

1. 4849
2. 4856
3. 4851
4. 4853
5. 4859 (+ a plan)
6. 4848
7. 4860
8. 4854
9. 4861
10.4858 (+ a plan)

I am supprised about these and although I don't agree with alot of the hammering some people give kabam, their standard response that there will be a thread explaining battle mechanics coming soon is not acceptable. We have been waiting for months so at least help us work it out for ourselves.

I have accepted that zero losses are impossible but how can I attempt to minimise losses when I cannot trust my experiments. Are there any other variables I should be taking into account.

Any advice would be much appreciated...