exact combat resolution method?

[bvanevery]

What is the exact way that the game compares attack and defense values and obtains a result?  Oddly, it's not in the Datalinks!  For years and years I've played this game, relying only on my experiential grasp of the relative strengths of units, how they actually seem to perform in combat.  Not a mathematically precise specification of how combat is actually computed.  I've poked around the wiki for a bit but didn't a page on this.

I found a webpage which partially explains it, but doesn't specifically say how many "virtual dice" or rolled or what the unit hit points ? are regarded to be.

Regular combat

Combat is resolved by matching the attacker's weapons rating and the defender's armor rating, modified by special abilities, any applicable combat modifiers, and morale. The game rolls virtual dice based on these odds for as long as the units remain engaged. Fast units, like speeders or gravtanks can disengage from a losing battle and retreat.

Otherwise, the combat continues until one of the units is destroyed. Damaged units that survive combat suffer impaired performance promotional to the damage they've sustained.

I see that Reactor size determines hit points (10, 20, 30, or 40) but I'm not finding how much damage is done per die roll.

I seem to recall that some (all?) of the Civs had the concept of Firepower, which makes the amount of damage done clear.  For instance, Stone Age units had Firepower=1, Musketeers had Firepower=2, "modern" units I think starting with Cruisers had Firepower=3.  I might be going from memory on Freeciv with this, not commercial Civ games proper.  I might be wrong, but I think this is correct.  There were hit point increases that went along with this as well, so their version of "Reactors".  So when you get a hit on the enemy, you do your Firepower in damage. 

[Geo]

Ingame, there's an option to show combat odds before the fight starts. It shows data like hitpoints, attack, -and defense values. Perhaps that's enough for your needs?

[bvanevery]

I'll try turning that on and seeing what it says, to gain familiarity with it.  Historically I have not bothered, because in several cases, its assessment of the combat odds are somehow incomplete.  I know such-and-such unit will die if I attack, but it gives this gloom-and-doom about how it won't work.  Can't remember if that's about psi combat, or attacking Formers, or attacking while 1/3 or 2/3 move remaining, or what, but there are cases where I've found the odds given to be goofy.

I suppose I should also turn on the slow combat animations.  Maybe I can see how much damage is done.

My actual need is to either clone SMAC's exact combat system, or to contemplate the exact system so as to devise a "better" one.  For a new game.  I think in a new game, I'd like it to be possible for units to destroy each other, and for units to have a standoff where neither is destroyed.  At a first go I'm imagining tactical scale units fighting each other, i.e. little Plasma armored dudes with actual Particle Impactor cannons shooting at Silksteel armor and why does anything even matter?

EDIT: I watched my combats for awhile.  They did not help me solve the riddle of how many hits or damage per hit are done.  I did notice that an artillery attack could do what seemed like a lot of damage all at once.

[PvtHudson]

...in several cases, its assessment of the combat odds are somehow incomplete. ... Can't remember if that's about psi combat, or attacking Formers, or attacking while 1/3 or 2/3 move remaining, or what, but there are cases where I've found the odds given to be goofy.

It's about PSI combat against reactor above Fission. Assessment doesn't know PSI units health is adjusted so that enemy reactor doesn't play role.

[ih8regin]

I see that Reactor size determines hit points (10, 20, 30, or 40) but I'm not finding how much damage is done per die roll.

One. It's just that the animations happen only each third damage done to either participant, displaying an "explosion" over that unit's icon, this might confuse people.
Yep, in PSI combat the damage delivered is 10% regardless of reactor, or 1*reactor.

[Alpha Centauri Bear]

Here. Just for you.
http://alphacentauri2.info/wiki/Combat_Mechanics

[bvanevery]

Thanks for that!  Good to have the fuzzy hazies set down in print somewhere.

For Disengagement, I'm pretty sure you have to have more than 50% health remaining to be able to disengage.  I think once you get knocked down to 50%, you disengage.  If you started at 50%, you die.  This gleaned from attacking many things twice.  I've never had to attack 3 times.

I'm still trying to decide what "betterness" in combat systems means for a future game.  I've made several posts on Reddit about it.  I've noted the "rounds rolling over and over again" problem to others, the way it skews the odds.

Someone told me that Civ 1 did not have hit points!  So a strength 4 attacker vs. a strength 1 defender really would have 4:1 odds, or 80% attacker vs. 20% defender victory chance.  So they claimed.  They don't have to be right.  But they claimed the proportion of victory is:

A = attack strength
D = defense strength
attacker victory odds = A / (A+D)

More recent Civs seem to just add or subtract things from attack and defense strength, to arrive at odds.  They don't express anything in terms of %XX modifiers, and there isn't the more complicated summing vs. multiplying we see in SMAC.  That's probably less brain friction for figuring out what's going on.

[Alpha Centauri Bear]

For Disengagement, I'm pretty sure you have to have more than 50% health remaining to be able to disengage.  I think once you get knocked down to 50%, you disengage.  If you started at 50%, you die.  This gleaned from attacking many things twice.  I've never had to attack 3 times.

There is a good description of disengagement somewhere I saw on the net. As far as I remember you need to start with higher speed than attacker and you need to drop to 50% of initial health you had before the combat + other retreat conditions.

I'm still trying to decide what "betterness" in combat systems means for a future game.  I've made several posts on Reddit about it.  I've noted the "rounds rolling over and over again" problem to others, the way it skews the odds.

Do you refer multi round combat?

Someone told me that Civ 1 did not have hit points!  So a strength 4 attacker vs. a strength 1 defender really would have 4:1 odds, or 80% attacker vs. 20% defender victory chance.  So they claimed.  They don't have to be right.  But they claimed the proportion of victory is:

A = attack strength
D = defense strength
attacker victory odds = A / (A+D)

They claimed, yes. The actual probability is still skewed even for single round Civ 1 combat.

[bvanevery]

Yes, multi round.

The dominant consensus of opinion among 4X players and game designers who answered my queries, is that players generally do not need explicit odds calculators like you and I have been worrying about.  They're mostly capable of intuiting a system over time, if it's simple enough.

To the extent that players do want explicit odds declarations, some reminded me about hex aiming reticle approaches, where a very few numbers appear directly over the unit(s) to be attacked.  The "big portraits and lists of all factors that takes up a lot of the screen" approach that SMAC uses, might actually be atypical.

So I'm not rockin' out with a "better" system just yet.

[Alpha Centauri Bear]

You are mixing two questions as usual. Or people you asked misunderstood the question.

You ask them whether they want to see a single final number representing their winning chances vs. having a final number with detailed breakdown of its components (bonuses). Then the answer could be either depending on whether they are casual or seasoned players. Latter care more about details.

However, ask them whether they want to see a single final number representing their winning chances vs. not having this final number but having some raw input data those they need to convert to final number using some complex non-linear algorithm using their brain and/or calculator instead. I doubt anyone would prefer raw data.

That is what I did. Converted raw data into a final answer that everybody wants. Now computer does computations, not you. Don't tell me that doesn't make you game easier.

[bvanevery]

You ask them whether they want to see a single final number representing their winning chances vs. having a final number with

No I didn't.  My post was mostly about ragging on SMAC's stupid attack vs. defense system, the nonsensicality of armor magically attacking a plane.  I told them that to do better, you end up with a complex 2D visualization of all the outcomes.  Or worse!  Never, ever, ever, did I ask anyone if they thought having a single number for "odds" was acceptable.  Because it's not acceptable to me.  I want to know more than that.

but having some raw input data those they need to convert to final number using some complex non-linear algorithm using their brain

My original post was about more realistic simulation of combat.  It wasn't about providing odds indicators for maximum mental convenience.  The issue of whether all these extra factors really mattered, did come up eventually anyways.  Like, would a scalar system suffice?  Do I need both weapons and armor, in a 4X TBS game?  It's debatable.  We debated it; I'm still debating it.

Converted raw data into a final answer that everybody wants.

There is no one single number. There is a spectrum of odds spread, in a system where either combatant can die, and either combatant can be left standing at the end of combat.  How wounded either is, is material to the outcome for planning purposes.

[Alpha Centauri Bear]

Ah, you mean generally in all kind of 4X games? Then I don't have anything to say on that. It is too broad for me.

[bvanevery]

Yep, I was posting about the theoretical consequences of any possible system.  I was using SMAC as my concrete example.

An interesting thing I only realized today, is that having combatants not fight to the death, automatically turns it into a 2-dimensional grid of outcomes visualization problem.  Has nothing to do with whether the underlying system uses weapons, armors, attacks, defenses, any of those more complicated factors.  Even a simple scalar system comparing 1 number to 1 number, would have this result, if combat is not decisive to the death.

Doesn't matter what the inputs to the system are.  The output is a 2D grid of results.