[chrispedersen]

Originally Posted by thejeff

OK, so the test:
I started a tiny game with Arco & Man, starting with 9 provinces so they'd share a border without having to fight indies.
First turn moved the starting troops to the border and built a commander(M) there for Man.
Second turn built a commander(A2) in the border province for Arco.
Third turn both attacked across the border.
Arco led by one of his starting commanders(A1) and A2, Man led by M.
All ordered to retreat.
The battle took place in Man.
I repeated this with the same commanders. The next time it was also in Man. The third in Arco.

This led me to think it's not that simple. Then I ran another try hosting with debug on. The following was interesting:

host: move
  findpath from 6 to 8... distance 7 ms 514 bestland 8 ok 1
from 6(sea0) to 8(sea0) (water-1 boat0)
  findpath from 8 to 6... distance 7 ms 514 bestland 6 ok 1
from 8(sea0) to 6(sea0) (water-1 boat0)
  findpath from 6 to 8... distance 7 ms 514 bestland 8 ok 1
from 6(sea0) to 8(sea0) (water-1 boat0)
      moves validated
  Move interrupted from 6 to 8
  attack move: Bune moving from 8 to 6

Bune is the Man commander. 6 is the Arco province, 8 the Man one. This battle took place in 6.

The next try had the same findpath and move validation, but ended with:

  Move interrupted from 8 to 6
  attack move: Leonteus moving from 6 to 8
  attack move: Kake moving from 6 to 8

Those are the Arco commanders. The battle took place in 8.

This doesn't seem to be based on commander id or the order of the moves.

Very interesting.

It does look as if moves are validated by commander, although what determines the province the fight occurs in, and the order of the battles is not clear. ( But it isn't cmd order).

Its also interesting that the distance isn't 1, its 7. Does this correspond to the relative sizes,do you think?