A simple approach to combat

[agent_smith]

Memetix and others, thank you for keeping the ball rolling

[Darhuuk]

I have some questions about this as well. Take a look at this picture:



You can ignore the shaded squares, they show the true combat range. We are interested in the combat range + 1, which is the ant's influence. The influence is shown in each square using red & green numbers (top & bottom left of each square). A square should be safe for a green ant if the green number > the red number in that square.

As you can see, left of the bottom green ant should be safe & right of the top green ant should be safe as well, since the influence fields suggest 2vs1 in both locations. However, if we move the green ants there and the bottom red ant moves up 1 square, both of the green ants die (as well as both of the red ants), so these squares are not safe at all.

Another example is bottom green moves left, top green stays where he is. Again the fields suggest this should be safe, but again, both will die if the right red ant moves left and the bottom red ant moves up.

Am I calculating something incorrectly here?

[Memetix]

[Darhuuk]

[carlos.guia]

[dshawul]

I think in the example both green and red ants die which is not so much of a loss. It would be interesting if it is possible to suicide with this method. a1kon's method has this problem ,which is probably the case for all combat strategies as well,since we have an imperfect information game here. I think a good combat strategy should decide for a group of ants simultaneously to minimize suicides.
YMMV

[a1k0n]

You could fix it by adjusting the influence map once you commit an ant to a move, removing its influence in all the directions it's not going. Then subsequent ants will be forced to go along with its choice (sort of -- it still doesn't represent "if ant B doesn't move here, ant A dies").

[Memetix]

[Weir_Doe]

[bluegaspode]