The Shield Problem(s)

[hrvojej]

But then we have the massive discrepancy between win/loss ratio of human vs. computer controlled unit with supposedly similar stats...

Did you do your tests on medium difficulty? If so, the medium difficulty level might be slightly biased towards the player, and therefore the player wins more often. I though that the best way to counteract this was to use the very hard difficulty for the tests, since if anything it should be biased against the player (even though from what I've gathered from the forums on VH the battles are in fact balanced). This is also a problem if all of the test runs are done by player controlling only one side - the effect of difficulty level cannot be discarded, which is not the case if you alternate which side you are controlling.

Knowing that animations truly are what most units draw their power from. Has anyone taken the time to look at the attack and kill animations of the same units, with and without a shield stat? It may be that the shield stat somehow slow's them, or somehow adds an interupting parry movement that lock's them from really attacking.

I watched closely all my test runs, and though I have no numbers to show for this, from what I've seen the animations were the same. The French simply got slaughtered really quickly after the charge.

EDIT:
For missiles, the things are a bit more complicated. I think they actually have to hit the model of a man in the target unit unit before the att vs. def is even compared, and I think that the hit rate is dependent on more than just stats. Also, you have a lot more variables to consider: the general scatter of the projectiles, the speed of the target unit if it's moving, the actual density of the soldiers in the target unit, the spatial arrangement of the firing unit (perhaps the shooters are blocking each other), perhaps even angles (I seem to get almost no kills from those 90 degree 'high shots'), etc. While some of these things could be affecting the melee as well, it is my impression that they would more pronounced when trying out ranged combat. Hence, if the effects are small, whoever tests this would have to be very careful to control for as many of the things as possible and also to do a lot more test runs because of the larger inherent potential for variability. Of course, another thing that would help would be to push the stats to the extreme in an attempt to increase the effect size.

Cheers,