Regarding the statement that when defense is higher than offense, as I described, it is indeed the case; I checked it. We even had an entire conference call on Skype where I ultimately refuted the physics engine. Specifically, we were discussing how much damage these Archangels would actually deal, but I don't have a save file of them.

Well, respect, there is indeed an error in the physics engine. But either you wrote the wrong formula here, or it's also incorrect. In my test save (attached), 10 Archangels with an attack of 38 deal 400 damage to Archangels with a defense of 49; both heroes have experience-based offense and armor. According to the physics engine formula, it should be 435, and according to your formula, it's 373. The correct damage of 400 is given by this formula: 10*50*(1+0.3)*[1-(11*0.025)]*(1-0.15) I checked several variations with a negative attack-defense difference, and it seems to work. In a party with desynchronization, it will be: 16*50*1.3*(1-0.175)*(1-0.15*2.05)=594, If the angels are on defense with a +6 defense bonus, then: 16*50*1.3*(1-13*0.025)*(1-0.15*2.05)=486 P.S. We got a little off-topic, sorry to the organizers, it's all Captain's fault; perhaps it would make sense to move this somewhere else.