I understand, but the main reason I don't like it is that it asks
the game to do something that you can't normally do with co-operative
deals: correctly predict future behavior. In your example, I would
only actually do the attack if I knew for sure I was going to be paid
off later. A less objectional mechanism would have the ENFORCED
orders executed for as long as they matched, and once they
diverged, to flush to the END (i.e. as long as you think the other
guy is sticking to the plan, you stick to it too, but you can't read
his mind to find out if he will stick to the plan tomorrow).
The other reason I don't like it is that it give too much scope
for "bullying". A Big Meany can dictate a whole month's worth of
ENFORCED orders to a small fry, with an "ATTACK smallfry" on day
31. If the small fry doesn't do what I say, they die. This may
be fun for a certain kind of Big Meany, but clearly ruins the game
for the small fry.
Finally, I think trust works almost as well, and is a lot simpler.
-- Bron Campbell Nelson email@example.com or possibly uunet!sgi.com!bron These statements are my own, not those of Silicon Graphics.