Ai / dice rolling combat?

[maholic]

Do individual units get a dice roll each turn?

eg my 38 exp cavalry with max health, will die to any unit it attacks in this certain save i have in 1398. (i know that odds of combat are calculated once per turn but i thought it was based on odds to fight different units, not the individual units dice roll thus it ends up being universally useless that turn if it strikes the terrible dice roll?)

It attacks an elephant and only takes out half its hp, now the odds were like 90%, it tries a musketman, same deal it dies, tries a conquisitor, same deal it dies.

So say its dice roll was a 1/100 that round X str modifiers, its going to die vs anything it fights that turn above minimum str to survive the first round of combat?

whereas the next cavalry in my stack can kill any of these 3 units and win every single time that same round.

Also why does invasion Ai have to be so fixed, its always 3 destroyers 3 transports? Do you know how easy that is to counter with a wolfpack of subs with withdraw at 80%

[Yosh]

I've notived the same thing, unfortunately...

The odds will be slanted ridiculously in my favor (like, 98%), and I will still lose - this has happened enough times to the point where I suspect there may be a new bug in the RNG!

I took screenshots of my combat log, because some of them are simply ridiculous (i.e. a full strength/HP tank with City Raider III losing to an Infantry that is at 50 HP) - that was happening consistently to me!

[Hauptman]

Yes they diffinatly ruined the "tank vs spearman" feature with this patch, I constantly loose combats that are 2 to 1 strength difference. I thought the patch was supposed to make the stronger units even less likely to loose now... i tell you it gets frustrating to loose 3 full strength artillery units on one hurt infantry...

[Yosh]

Well I don't know if I'd go so far as to say that, Hauptman...

To me, it just seems as if there is a bug in the RNG, because the weaker unit will get a bunch of "hits" in a row, while the attacker gets none, so that even though the damage is being figured correctly (as far as I can tell), the stronger unit still will lose consistently enough to bother me a LOT - I've only seen this "bug" (if indeed there is a bug) when I am attacking... I'll post some screenies of my combat logs later when I get home from work

[Vimpster]

Before you start a game there is an option that you can toggle to make it so that everytime you load the game it creates a new random seed. Without that checked you will have the exact same results everytime you reload your game.

[Yosh]

Yes Vimpster... that relates to what Maholic posted; the issue that Hauptman and I were discussing isn't related to re-loading the game however.

That's a good point though; I wonder if I would see the same craziness with the RNG if I had had that option checked, and then reloaded?

[Yosh]

ok finally returned from work and put the screenshots together -- these shots were all taken during the same session of the same game -- a standard sized map, continents, monarch difficulty. I took the shots to get the combat log, then cut and pasted them into a collage for greater clarity..

To me, it seems that there is something odd with the RNG, but maybe I'm just paranoid

Let me know what you think of this!

[maholic]

nice logs

man look at those multiple hits in a row...

[Yosh]

heh it'd be nicer if I were on the other side of those hits >:O

that being said, it's the very fact that those multiple hits happened so often that makes me suspect gremlins lurking in the code somewhere

[maholic]

um, looking at those logs more, it almost seems like the wrong unit is taking the wrong damage?

why would your battleship take more damage per round???

it has more STR so it should do the greater amount per successsful hit?

and that helicopter seems to be doing way more than what your tank does?

i know helicopters get a tank kill bonus but the combat odds show you have enough promotions etc to compensate surely?

like the helicopter damage its doing is what the tank should be doing to the damage helicopter and vice versa?

maybe some screwey math with damaged units where the wrong unit is taking the damage per tick?

[Yosh]

well damage-wise, with the new patch, units that are damaged still *inflict* damage as if they were full strength...

still, you may be on to something there in that sense, as the gunship did do more damage/hit than the tank...

Maybe math AND the RNG?

I can't remember enough of my Bayesian equations/probability equations to calculate the odds of randomly getting those reslults, but my gut tells me that something is fishy

[ghen]

the random seed is created a few turns before the current round of combat. This was the same in Civ3 and maybe even Civ2. So if you reload a save and play the turn exactly the same you will get the same results each time.

[cal_01]

Yosh: I get similar results with combat on my end. It's frustrating to attack cavalry with modern tanks, and have my tanks either die or come out with less than 15 strength. It's rare that they would actually suffer no damage.

[Crossfire]

It seems even worse than that. Often when I should win easily, all combat in the first 2 rounds kills my guys but then something is recalculated and I win ALL succeeding rounds. I will try to start capturing combat logs to substantiate my point.

[Yosh]

That'd be great Crossfire; it'd be nice to get more data on this.

If anyone can think of some quantitative tests we can run to check this, that would help a lot too!

Maybe tonight, I'll set up some combats in the world builder, and gather some statistics on the combat - it seems to be most noticable/prevalent in combat between the more modern (higher-strength) units