Is it just me or does the AI get an insane combat advantage?

[poprawkz]

I am sorry but calculating the odds in my head, even inclusive of a defensive terrain bonus etc... the AI has a hugely unfair win probability in most combat situations. When I watch my infantry fall to cavalry units while defending on hills or in cities I just gape in wide eyed wonder...

 

[Zouave]

For the thousandth time, Firaxis' combat values are ridiculous.

My mod's values are both historical and more playable. See that mod, or others, and make some changes.

 

[Hurricane]

The AI does NOT get any combat advantages on ANY difficulty level. The "strange" results you see is due to the fact that the relative strengths between units is relatively small. If you can´t stand losing the odd combat I suggest you tweak the values in the editor, or use a mod like Zouave suggests. Note, however, that all mods I have seen give the human player an even bigger advantage over the AI than he has now.

But that might of course be just what you want.

 

[Zouave]

Oh, yes. In Civ 2 the only AI cheat was that the AI DID get progressively more combat benefits as we went up in Difficulty level.

I prefered that over the sneaky and insidious cheating the AI does in Civ 3.

 

[Zouave]

Originally posted by Hurricane
. . .I suggest you tweak the values in the editor, or use a mod like Zouave suggests. Note, however, that all mods I have seen give the human player an even bigger advantage over the AI than he has now.

If reality and facts regarding unit strengths give the human an advantage (even despite all the other AI cheats) over the AI, then the AI is more screwed up than we thought.

 

[Lt. 'Killer' M.]

Originally posted by poprawkz
I am sorry but calculating the odds in my head, even inclusive of a defensive terrain bonus etc... the AI has a hugely unfair win probability in most combat situations. When I watch my infantry fall to cavalry units while defending on hills or in cities I just gape in wide eyed wonder...

It is just you, but it has been found that the random streaks are maybe a bit too long and thus unpractical for a game that uses such small samples to decide battles. You can simply try higher HPs, maybe go

Conscript=4
regular=6
veteran=8
elite=10

 

[poprawkz]

Sounds good but, wont that affect my bombard values? I mean it will make artillery and the like next to useless unless I double their attack power correct?

 

[Sirp]

It's entirely just you. The combat mechanics of the game work precisely as stated, although the mechanics themselves are a little controversial.

Personally, I see little problem with them. I rarely lose to a truly inferior unit, and sometimes cavalry *should* defeat infantry (world war I), sometimes warriors *should* fight off swordsmen.

And yes, true, occasionally a tank does get beaten by a spearman etc etc. Doesn't happen often, and it happens just as often in my favor as it does against me. (except of course, I tend to attack obsolete units more than the other way round).

Personally I'd make a few changes (longbowmen more expensive, better, english uu and everyone else gets crossbowmen for instance), however I do value playing the "real" game so I can compare with others and so I know the AI is working as intended.

 

[Zachriel]

Whether you like the combat system or not, it is the same odds for you and the AI. The defender gets advantages, though; so if you are on the attack most of the time (we hope ), then the odds are stacked against you somewhat.

 

[Gastric ReFlux]

Originally posted by Zouave


I prefered that over the sneaky and insidious cheating the AI does in Civ 3.

This is one my favored methods of argument. Rather than relying on documented verifiable proof, it is a lot easier to claim something is sneaky and insidious, a ghost in the machine that is beyond any juridstiction.

 

[Reboot]

I think its just that combat resolution is flawed. I've seen too many cases of AI units that can kill my defending, fortified, and sometimes enhanced-with-defense-perks units without taking damage to think otherwise. Mind you - this even happens in Chieftain.

The AI probably didn't get special advantages, since I get those occurances when I take a rare offensive at the AI cities sometimes too.

Have you seen a regular archer kill an elite fortified-in-city HOPLITE without taking damage? I have. That such a case is even possible sounds utter surreal. We know that defenders get lots of perks, but somehow, these perks can fail (so exactly how much of a random element is there?). Why does the combat system even allow this if its out to give aggressors a hard time, then? Seems pretty ironic.

 

[The Last Conformist]

I very much recommend doubling HP if you want to minimize "weird" combat outcomes. And yes, if you do it, you should also increase the rate of fire of bombardment units. Exactly how rate of fire works is AFAIK not known, but it seems to me that you should increase it by more than 100% in order to perserve the relative efficasy of bombardment units. In my double-HP mod, rate of fire 1, 2 and 3 became 3, 5 and 7.

 

[warpstorm]

FYI, Rate of Fire is how many actual attacks the bombarding unit does per strike.

 

[Zachriel]

Originally posted by Reboot

Have you seen a regular archer kill an elite fortified-in-city HOPLITE without taking damage? I have. That such a case is even possible sounds utter surreal. We know that defenders get lots of perks, but somehow, these perks can fail (so exactly how much of a random element is there?).

Why would that be surreal? Battles are a very chaotic events, so a certain amount of randomness in combat results is expected. A regular archer will beat a fortified hoplite 1/23 times. If you have a couple of hundred combats in the 6,000 year history of the world, you will have a few such events.

The combat resolution has been analyzed in quite a bit of detail in other threads. Also,try the civulator at
http://www.columbia.edu/~sdc2002/civulator.html

 

[Flavor Dave]

What gets me is not the single combat that a unit wins against the odds. It's when the same unit wins several. I call these units "magical units." It's one thing when an unfortified, regular spearman beats a vet swordsman. It's pretty unlikely, but it could happen. It's another when he beats 3 in a row.

But it works in my favor sometimes too.

 

[Lt. 'Killer' M.]

about bombard: do change it as The Last Conformist did if you double.

 

[Zachriel]

Originally posted by Flavor Dave
What gets me is not the single combat that a unit wins against the odds. It's when the same unit wins several. I call these units "magical units." It's one thing when an unfortified, regular spearman beats a vet swordsman. It's pretty unlikely, but it could happen. It's another when he beats 3 in a row.

But it works in my favor sometimes too.

Ye ol' invincible spearman tale.

Actually the odds for three veteran sword in a row to lose against an unfortified, regular spearman is only 1/100*. In several hundred combats not all that unusual. If the spearman were veteran, fortified and across a river, his odds of survival are 1/6*.

Xerxes had the same problem at Thermopylae, only the stakes were much higher.

http://www.geocities.com/the_temple_of_ares/300spartans.html

* Assuming each combat is calculated independently. If you are concerned about a specific event, you may want to post a game.

 

[Reboot]

Originally posted by Zachriel


Why would that be surreal? Battles are a very chaotic events, so a certain amount of randomness in combat results is expected. A regular archer will beat a fortified hoplite 1/23 times. If you have a couple of hundred combats in the 6,000 year history of the world, you will have a few such events.

The combat resolution has been analyzed in quite a bit of detail in other threads. Also,try the civulator at
http://www.columbia.edu/~sdc2002/civulator.html

It's not so much that the Archer beat the Spearman - accidents happen in even the most planned out battles. It's that the Archer did so against a more experienced Spearman without taking a single hit.

 

[Tweedledum]

Originally posted by poprawkz
Sounds good but, wont that affect my bombard values? I mean it will make artillery and the like next to useless unless I double their attack power correct?

Don't worry - Artillery is ALREADY next to useless

 

[Sirp]

Reboot: the spearman being more experienced only gives the spearman more hitpoints. Thus, a conscript archer is just as likely to beat a veteran spearman without taking a hit as an elite archer is. So the "anomaly" is simply that "an archer beat a spearman without taking a hit". Anyhow, it happens sometimes; it would happen in a real battle sometimes. Despite many claims, no-one has yet posted any convincing study which shows the RNG to generate weird, strange, subversive, or otherwise unexpected results. Personally, I did find Civilization I annoying - the combat was just too simplistic. But I don't have any problems with the Civ III combat system. Giving units more hitpoints just makes combat too predictable, imho.