How is battle combat rating figured?

Seems to be that attack, defense, combat speed, and HP are put into some formula.

You can't put much stock in the rating right now. Even a higher rated force can get destroyed if they're all low attack power units trying to attack a single high defense monster (as they don't do enough damage to get past the defenses).

It's something like...

( AtkSpd x Attack Strength ) + Defense + ( sqrt (Current HP) )

...but is subject ot change as the battle system is tweaked. 

Ok thanks!  That is exactly what I was looking for.  I assumed that it would be tweaked since it was a beta, but that's a great starting point for me when considering how to build/analyze units.

Good to hear, because the current calculation method is far from being perfect.

hmm altough somehow logical this would lead to the creation of high attack 0 defense/low hp mass-one-hit units and huge hp/defense 0 attack single city-defense units...?

Probably not. A high attack/0 defense unit would get destroyed by a peasant horde, which could simply swarm it into the ground.

While a 0 attack/high defense unit would make an alright defender, anything that *can* pierce its defenses would kill it with zero risk. I'd agree that high defense units would make effective defenders against weaker soldiers though, due to the inability to do much harm.

They attack more than once, and both sides need both attack and defense. A unit with high def and no attack might be able to survive a while but it's not going to do any damage and eventually will be worn down, even by a unit with 1 atk. Don't forget there's an actual battle taking place under that results screen.

that means its not like 10 attack vs 5 defense = 5 damage, therefore also a 1 attacker can harm a 10 defender over time?

You forgot about the various dmg types & resistances. So if X unit is highly resistant against blunt dmg [let's say 90%+ res.], unit Y equipped with a mace [considering that the mace does blunt dmg only] won't be able to do any serious dmg on it, even if X unit's DEF is uber low, and Y unit's ATT is uber high. Not even if it's a 20x unit Y vs. 1x unit X situation.

i was not aware that such a differentiated combat system was in place (or will be)...?

Dmg types are in. [Note: Thankfully...because without dmg types we would end up having a primitive & dumbed down combat system. Good example: Civ4.] You can read about it in this thread. The devs haven't talked about the resistances so far, but if we have dmg types, it wouldn't make any sense, if we wouldn't have the appropriate resistance types. Correct?

No. It's very easy to get defense up into the 20s, but I think the best melee weapon that can be researched right now is like 7 attack, and the best I've seen from quests is 10 attack. With those sorts of numbers and a flat reduction system battles would never end as everybody would be doing 0 damage, but that's not the case. I actually think defense is dodge chance not damage reduction, I dunno I haven't paid much attention to it. I'd imagine attack is a mix of both damage and hit chance. There's a lot of misses going on if you look at the combat report.

dodge chance might makes sense since dexterity increases defense, but at the same time unlogical cause armor increases defense too which can hardly increae dexterity - maybe a combination of damage reduction and defense - but then i am also wondering if certain attack types are worse / better against certain armor types (e.g. are magical attacks also targeted against defense or rather against wisdom/intelligence?)

In D&D and Diablo (among others) armor is chance to avoid damage. The logic I guess is that armor helps deflect blows or outright prevent it. I would bet the influence here is D&D, dex and armor increase dodge chance, and strength increases damage and hit chance.

Then again the few times I looked at the combat results I also saw a lot of 1s. So I don't know if something is reducing the damage to 1 or they just only do 1 damage a lot of times (glancing blows?).

Many months ago in a thread I can't find, there was a discussion on how this worked at the time. Short version: it works like it does in Galciv 2. (Or it did back then, it's entirely possible it's been totally changed.)

Longer version: Defense is not dodge, it's mitigation. Attacker does a roll between 0 and their attack stat. Defender rolls between 0 and their defense stat. Damage done is Attack Roll - Defense Roll, or 0, whichever is greater.

So a unit with Attack 5 hitting a target with Defense 10 could do 5 damage. It could do 0 damage. It could do anything in between. Most of the time it will do minimal damage, and if the defender roll is >= 5 (which it will be about 50% of the time), the attacker will do 0 damage.

Damage types don't seem to be in the current version, but they are scheduled to go in. That will mix things up some.

In regards to the examples above, a high attack/0 defense unit will pulverize other units, but also get pulverized in return. A 1 ATK/1 DEF/5 HP peasant can damage it, but will probably die in one attack. 50 of them will win the day though, simply due to their ability to inflict damage faster then the attacker can pick off each peasant.

Conversly, a 0 ATK/high DEF unit would still lose to the peasants as it could never kill any of them, but it'd take a very long time since it would only ever take damage if it rolled a 0 on its defense roll AND that peasant rolls 1 on its attack roll. A 1 ATK/high DEF unit would almost certainly slaughter the peasants due to its much higher chance to inflict damage. (As the peasant only has two possible defense rolls, you'll see zero a lot more often.)

i am really looking forward that they implement special skills, attack type, resistances, status etc. etc. to give the battle much more tactical depth.... i would really like the battles to be like tabletop game battles (complexity wise)

Sadly no absolute combat rating, that doesn't take into account the opponent, will do the trick for estimating accuratly the odds of a battle outcome. What you need is how many turns needs my army in average to kill the other army vs how many turns need them to kill me to estimate if you would can win the fight with eventually a bit of luck.

Your attack against ennemy defense gives an average damage per swing.

When you factor  your combat speed, you will get an average damage per turn.

When you finally factor the opponent HP, you will get an average number of turns for killing the opponent.

In addition or instead of the combat rating, could we have a way to figure out the the % chance of success vs a given opponent.  Say though the game running a few mock auto-resolves in the background?  

The problem I'm finding is the one talked about here, you have a bunch of high defense, low attack adventurers.  They can have a combat rating of 50+, but can't actually hit an enemy with any kind of defense.  This situation makes the combat rating less than useless because it actually gives you the impression that you have a chance of winning when in reality you don't.

Later,
LAR

Edit: Thinking about it a bit more, that might help the AI figure out which battles it can win also.

not a bad idea...the combat rating is abit missleading the way it is...on the other hand i also find it realistic that even 50 peons with spears can not really become dangerous to a 30 feet tall dragon because they can simply not harm his thick skin...

Good point.

Maybe a relative rating would make more sense. When you have a stack and mouseover another stack to attack, show a combat rating that is the relative strength of your army vs their army. If you know what is fighting what, you can compare average attack to average defense, and come up with a numbre that will be saner.

i think the devs said that using terrain will have influence on the battle (so attacking from a hightend position etc.) - is it also beneficial to attack the back / flanks of enemies - i think this would immensly increase tactical options, especially if troop-units (grouped) can not simply turn from one moment to another (a true regiment consisting of several 100 soldiers couldnt do that either)...?

I have done some calculations based on the algorithm exposed Tridus and assuming that when a number between 0 and N is rolled, all integer values from 0 to N have the same probability. The results are roughly the following (if my calculations are correct):

If you are facing an opponent with 0 DEF, you will roughly do an average damage  of your ATK/2 per swing.

If you are facing an opponent with the same DEF than your attack, you will roughly do an average damage of your ATK/3  per swing

if you are facing an opponent with a DEF equals to 2 times your ATK, you will roughly do an average damage of your ATK/6 per swing

if you are facing an opponent with a DEF equals to 4 times your ATK, you will roughly do an average damage of your ATK/12 per swing

I really hope they go towards MoM style combat calculation rather than the GalCiv one. GalCiv method is horrible with high ratings, the variance is just too huge. With an attack of 100, you have the same odds for each result, 1% each, creating a straight line of probability. In other words, a Civ effect where a spearman can kill a tank relatively often.

With MoM every point of attack has a 30% of scoring a hit and every point of defense has a 30% chance of mitigating a hit. A similar system is used in many boardgames and rpgs. The resulting probability is a bell curve where it's a lot more probable that a result of 30 is scored than a result of 1 or 100. Those extreme low- and high-ends are still possible, but you won't see those spearmen killing tanks unless they get really, really lucky.

Any update on this after the tweaks and changes?