are combat odds worth anything?

[Somebody613]

I've seen somewhere a VERY DETAILED EXPLANATION on how C2C combat works, and it's extremely "first strike luck" based beyond anything else.
Fact is, "strength" gets recalculated after each round, and this means that a single "lucky shot" CAN twist the DIRECTION of the entire battle drastically.
To make a crude example:
Say, both units roll for "hit" each turn, and the math is "if I hit, I do damage=(me/you) of points", where "strength" is the same as "health".
So, let's pit two "3 points" units against each other for two rounds, but with different starting "hits".
Y means this unit "hit", N means it "missed".
1. 3/3 =(Y/N)> 3/2 =(Y/Y)> 2.33/0.5 =(Y/Y)> 2/0.
Unit 1 wins and keeps "2 points" intact.
2. 3/3 =(Y/Y)> 2/2 =(Y/Y)> 1/1 =(Y/N)> 1/0.
Unit 1 wins and keeps "1 point" intact.
3. 3/3 =(N/Y)> 2/3 =(Y/Y)> 0.5/2.33 =(Y.../N...)> 0.5/0.
Unit 1 wins and keeps merely "0.5 points" intact.
And it all gets determined by the FIRST "hit" more than anything else.

 

[BlueTemplar]

[...]
Also, even if the odds displayed were for just the first round of combat, that still doesn't change much; any increase in odds for an underdog victory in the first round is matched by a reduction in their odds in losing the first round, yes?
[...]

I'm not sure I understand what you mean ?

The thing is that the % is a little less reliable for the battle overall because if the unusual happens in the first round and the underdog gets an abnormal upper hand, the odds can swing dramatically in their direction from there.

Odds recalculate round after round while outcome is a projection of all the possible end results of all the branches of possibility from there.

It would help if we stopped calling "chances to hit (and/or not get hit ?) in a single combat round" as "odds"...
in fact I'm not even sure what you are talking about in the context of the C2C engine if you say that there in the context of a normal (non ranged/first/stealth strike) combat round both or neither of the combatants can hit ? (Since the round can result in a win, loss, OR two kinds of "draw".)

(And again, per-round to hit = victory odds wouldn't even be true in vanilla - IIRC the math doesn't work the same for, say, an effective 1 strength unit with 2 HP vs an effective 2 strength unit with 1 HP, a known bug with the Alpha Centauri combat odds calculator.)

(EDIT : to illustrate further : say the 1 eff.strength unit gets lucky and both of these units get in the first combat round in a "draw" where each damages the other by 1 HP : it might be a "draw", but the 2 str 1 HP unit is now dead, and the 1 str 2 HP unit is half-alive and won the combat !)

The game itself clearly says "Survival odds" (and also "Defender Retreat odds") even on the first odds screen.

 

[Thunderbrd]

Those survival and victory odds are still an attempt to compile an overall battle result. They are just based on an assumption that the most likely scenarios keep taking place round after round.

Odds still, by the round, determine both the base likelihood to hit and the base amount of damage when a unit does strike. That base is still a ratio but both att and def gets their own tohit roll in a round in c2c unlike vanilla where each rnd only one roll determines which will hit.

 

[BlueTemplar]

No, since they display a whole histogram for the expected final HP, this means that the whole tree of possibilities is searched, not just the most likely events ?

P.S.: That blue R seems to be the initial ratio that you are talking about ?

 

[Thunderbrd]

No, since they display a whole histogram for the expected final HP, this means that the whole tree of possibilities is searched, not just the most likely events ?
View attachment 640468
P.S.: That blue R seems to be the initial ratio that you are talking about ?
View attachment 640469

That is how it worked before the combat engine was adjusted to recalculate strength after each round and therefore the new odds after each - those projections are a little oversimplified for the full combat mod engine and relied on old assumptions that could be made in most cases. They remain as a way to guide the AI and player, but the truth is, if you get an unusual result according to those projections, the projections would shift, mid-battle, each round, whereas these odds by the round were designed before that change. They are more technically accurate now on the vanilla combat engine but are still a 'close enough' guideline for the more in depth engine. They'd remain true in a sense of 'if average results keep taking place'.