Yeah you're right, the reason the HC's took so much damage is 2 of the types of units died real fast (10 vette/destroyer), so I think Wlerin tried to divide the ion damage by 6 to determine what the HC would take however it's more like divided by 4 because of the other two types of units getting knocked out so fast.

I understand that all damage is done simultaneously, but what determines the division in damage, such as if I have a stack of fighters and BSes. Do they split the damage done to them, by the amount of fleet composition % or...

1) Shields change the formula.

2) Without shields, every unit type (i.e. Fighters, Bombers, Destroyers, etc.) receive an equal amount of damage, and each unit type's total damage is divided evenly among the opposing side's units (though this doesn't matter until shields come into play).

3) As mentioned earlier, when one unit type is destroyed, the damage remaining is divided among the remaining types.

4) When shields come into play, units with lower shields or no shields will be hit with more damage, and not just because shields reduce damage. The actual % of total damage before shields changes. This is true of *all* shields, not just shields that are high enough to block all but 1% damage.
For example, if a fleet of Fighters and Cruisers were attacked by a fleet of just Fighters, the Fighters will receive 80% of the damage, and the Cruisers only 20% (before shields). (The actual percentage depends on the difference between the Fighters' attack power and the Cruisers' shields).

5) When shields are higher than the attacking unit's power, all but a small fraction of that attacking unit's power is redirected towards those enemy unit's with lower shields. Or, to put it another way, if a unit can't breach the enemy's shields, it won't try to. It will just target ships w/o shields or with weak shields.

That covers most of it, I think. Aside from the actual formula for how shields affect attack distribution, which i don't know.

The following input

Attack Force
[2009-01-16 11:12:27] <WlerinCZilla> Unit Start Quant. End Quant. Power Armour Shield
[2009-01-16 11:12:28] <WlerinCZilla> Fighters 300 164 3.6 3.5 0
[2009-01-16 11:12:28] <WlerinCZilla> Corvette 10 0 7.2 7 0
[2009-01-16 11:12:28] <WlerinCZilla> Destroyer 10 0 12 14 0
[2009-01-16 11:12:28] <WlerinCZilla> Cruiser 7 4.9 36 42 2.7
[2009-01-16 11:12:28] <WlerinCZilla> Carrier 4 1.9 17.4 42 2.7
[2009-01-16 11:12:28] <WlerinCZilla> Heavy Cruiser 4 3.5 72 84 5.4
[2009-01-16 11:12:28] <WlerinCZilla> Defensive Force
[2009-01-16 11:12:28] <WlerinCZilla> Unit Start Quant. End Quant. Power Armour Shield
[2009-01-16 11:12:28] <WlerinCZilla> Fighters 250 0 3.8 3.7 0
[2009-01-16 11:12:28] <WlerinCZilla> Ion Turrets 5 0 38.4 59.2 2.8


gives the following output with my battle calc:

Attack Force
Unit Estim Q. End Q. Power Armour Shields
fighters 163.34 164 3.6 3.5 0
corvette 0 0 7.2 7 0
destroyer 0 0 12 14 0
cruiser 4.86 4.9 36 42 2.7
carrier 1.86 1.9 17.4 42 2.7
heavycruiser 3.54 3.5 72 84 5.4

Defensive Force
Unit Estim Q. End Q. Power Armour Shields
fighters 0 0 3.8 3.7 0
ionturrets 0 0 38.4 59.2 2.8


With proper rounding, you get:
FT 164 (for units that do not have partials, you always round up)
CV 0
DE 0
CR 4.9 (4.86 is closer to 4.9 than to 4.8)
CA 1.9 (1.86 is closer to 1.9 than to 1.8)
HC 3.5 (3.54 is closer to 3.5 than to 3.6)

So, the results and what my algorithm predict are quite inline with one another ...

Thank you for that. I did not realise that AE only rounds up with integral units. That may have been where I miscalculated.

if you're looking at any calc other than kahar's you are likely not getting the right info about shielded distribution. the simpler way to look at 4 & 5 there is the basic dpu (damage per unit) model. calc the dmg the attacking unit type does to each defending type, add those up. distribution is dpu/ totoaldpu.

Hmm...since damage is all done at once. Then what do shields do? Block each individual unit's attack? Or just subtracts the shield amount from the total damage, and apply to the armor?

The first option. Shields block a certain amount of each unit's attack.

Say you have Fighters, with 4 power, attacking Cruisers, with 3 Shields. Only 1 (or 1.03)[1] power per fighter makes it across the Cruisers shields. If the Cruisers have 48 Armour, it will take 48 (or 47)[1] Fighters to destroy one (assuming there are no Fighters protecting the Cruisers).

The easiest way to make a formula out of this, at least that I have found, is to find
(Unit.AtkPower - DefUnit.Shield) / Unit.AtkPower (i.e. the percent of damage that crosses the shield) and multiply it by (Unit.Power)*(Unit.PercentTargetingDefUnit).

[1](The values in parentheses take shield bleed-through into account, and are thus more accurate. Shields only block 99% of their stated damage, or 99% of the attacker's damage, whichever is lower.)

Java calcs for the win

It's still helpful to know what's going on inside that java calc

Since the combat formulas appear to have been determined with sufficient accuracy, one should be able to then determine the most efficient fleet composition to use against any other composition.

Is there no tool with that capability available?


Yes, new to the game... thanks.

Yes, plenty of tools around that do that to varying degrees of accuracy. Only 1 that's 100% (or 99.9%) though I think.

Even then though you need to know tech levels, though I suppose you can play around with your own. There are some charts for what's best against what, but ultimately each unit needs its multi-dimensional chart based on the ratio between your and your opponent's shielding, weapon tech, and armour, as well as CC's.

But it's not that hard to understand the basics, hence the charts, which I'm sure you can find out there, possibly on the wiki.

I've been toying with a spreadsheet that predicts battle outcome, but you have to feed it differing attack forces in an effort to find your best outcome.

I was hoping a tool existed that allowed dumping in target tech, planet/ship defenses and yielding an attack force composition that would be most efficient versus the trial and error method I'm using.

The charts work fine if you're attacking one ship type, but the trial and error approach against larger compositions using battle calcs could be simplified by suggesting attack compositions based on your capabilities and/or priorities.

Judging from poor BR ratios, even from top guilds, few have this down to a science (or care to). <shrug>

Perhaps I'm just too much of a perfectionist

Against a large number of shiptypes, you can seldom go wrong with a fighter drop to start everything off. And then your next attack is greatly simplified.

Granted, that doesn't *always* work. And on rare occasions it's better to not even bother with the FT drop (i.e. when the amount of fighter armour is far greater than any other ship or base defense present).