Combat System Explained

[vmxa1]

Yes being on a hill or mountain gets a boost and I use it as often as I can. I would not be happy to see my riflemen in a city lose to swordsmen, but I do not call it a major problem as they are not that far apart. 3 attack 4 HP is within the parameters against a riflemen. It is not like the barb horsemen that just beat my vet knight that was fortified. This is on Monarch.

[Hurry]

Only the terrain of the defending unit is counted during combat, so it doesn´t matter where your units are.

And PLEASE, if you (Gen.Dragolen for example) want us to take your arguments seriously, POST A SAVEGAME!!! You have it in the autosave folder.

[Nmintz]

random number generator

Originally posted by Yohan
I've discovered something odd about the combat model.

I tried this a few times and it seems to be consistent. I tried attacking a regular spearman with a veteran swordsman and lost. I reloaded the game, tried the same attack, and lost again. What was strange was the sequence of hit point losses was EXACTLY IDENTICAL. I thought this was a strange coincidence, so I tried it a few more times, and the hit point loss sequence was identical every time. Curious, I tried attacking with a different veteran swordsman instead, and the results were once again identical.

But then I tried attacking with an elite swordsman and I won.

I have this strange suspicion that prior to each turn a table of combat results is precalculated. So depending on the attack/defense ratings of the units involved in combat and the number of hitpoints, the game looks up the results in a precalculated table to determine who wins. I have no other plausible explanation for this, and it strikes me as a very strange way to resolve combat.

I have indeed noticed this same result. I've even taken advantage of it when I'm in a cheatin mood (when your playing 15 civ games at emporer, it is awfully tempting at times).


I think the program is making use of a random number table like you will find in the back of most statistics textbooks. At the beginning of each turn it picks a line to start on and goes from there. That would explain a great many things (such as why cultural defections, which are based on probability), seem to be inevitable even if you start the turn over again.

makes sense to me.

Nathan

[Theseus]

Nathan,

If you are patched up to 1.29f, you may recall seeing a checkbox for "keep random seed," or something like that, when starting a game.

The game does indeed drive several probabilities from a PRNG, and thus reloading (unless you un-check that option) does not change any probability calculation results.

[vmxa1]

Yup, the rnd num is locked in at the start,of the turn unless you turn it off at the start of the game. The reload will be exactly the same. If you have a bad first roll and reload and do the same sequence of events they come out the same. Change the sequence and the first roll will be used else where, so the battle will be with another permutation (a form of cheating).

[Blue Moose]

The problem could be solved with hitpoints, as people say...just double or triple them. The lower the hitpoint levels involved, the closer to Civ1-type combat it is. Civ2 had hitpoints as multiples of 10, which is the main reason why it was so much better.

Anyhow, someone asked "where should you draw the line?" The answer to this is simple.

Units that are one tech behind should still be able to adequately defend and attack (though are a noticebly reduced effectiveness). Units 2 or more tech progressions back should get slaughtered nearly all the time. So warriors should get killed off easily by bowman and knights (and exceptions should be so rare you should be almost guaranteed not to see them). Cavalry should slaughter any ancient age unit. Riflemen should be at the end of their lifespan defending against tanks, and so forth.

While there would be some unrealism with the above, it would still be better than what there is now, and it would handle the complaints about the most modern units rolling over less advanced ones too easily. If you are far enough back that you are defending against tanks with musketmen (or modern tanks with riflemen)...you should get run over.

[vmxa1]

Amen brother. HP or FP or something could be used to make battles of obsolete units lose nearly always. A Fire Power factor that has each ages unit say 10 for earliest and 10 more each age. Now Warriors beat Calv once in 50,000 fights or something crazy. An extra HP each age may do the same, not sure. Many will scream that a warrior/spearmen should win those battles once in a while, but it should be like a Blue Moon, very few in your lifetime.

[Catt]

IMHO, there are 2 big problems with altering the HPs or other game features in an effort to make the more advanced units significantly stronger than older units:

(1) the AI civs are not adept enough at managing their treasuries to deal with upgrades. If they have the money, the AI will upgrade. But I've seen no evidence that the AI will adjust its science slider downwards for a turn or two in order to horde the gold necessary to make the mass upgrade that will come with the next tech (as a human will often do).

(2) It seems to me that the AI also has a problem upgrading units in cities without barracks. Rather than move older units to cities with a barracks for the upgrade, it seems that the AI leaves the older units until they happen to have a barracks handy. In a game with shifting borders, this often means that the AI borders will be guarded by weaker, older units.

Both of these aspects of the AI's gameplay processes mean that radically increasing the power disparity between old and new units tends to unbalance the game towards the human, making it easier. Maybe more in tune with the real world, but, in my experience and IMHO, less of an engaging game.

Catt

[vmxa1]

Sorry Catt, that is a very weak counter arguement. All that you said is true, but the change will not materially impact the game. How many times in a given game do these types of battles occur and the older units win? Most swear they have never seen them. If that is so they are not making enough impact to matter. I think they are not making an impact, except in the perception of the game, so correcting it will not hurt the civs much. The short comings of the AI's tactics can be addressed (or not) independently. It will have its biggest impact on the naval units, they are a joke in combat. even the nay sayers admit to seeing galleys sink battleships.
So I do not se it as a big deal, but it could improve peoples view of the game. Making the AI smarter is another thing altogerther.

[Proteus_MST]

Dunno,
I´ve seen some very extreme Forms of Mass Upgrades.

One time, I´ve observed my old enemy Katherine upgrade all of her Infanterie-Units to Mechanized Infanterie (more than 100 Units) in one turn.

On other times I encounter old Musket Men in the same city with more modern Riflemen.

But I can´t say for sure if the AI is just too incompetent to adjust the science slider, or if it just has other Priorities and just spends the money with hurrying Cavalry instead of upgrading old Musketmen.

In the first case observed I was at Peace with Russia and Kathy had only one war running, with another far distant Civ.

I often encounter the second case (i.e. obsolete and modern Units in the same City) if I am at war with an AI-Civ and just have a big Offensive running, where I strike deep into enemy territory.

[vmxa1]

I did not get into that aspect, but for my two cents, I do not think failure to upgrade is the AI's biggest problem. A bigger issue is the willingness to send troops into battle piece meal at times.It has other serious flaws as well. Like I said, correcting the combat a bit will not really crush the AI. It will only make it less irksome.

[Blue Moose]

Aye, the AI is very wierd in that if you take a city and are having a huge invasion, it just locks all its units up in cities (for the most part), instead of moving bunch to the front lines and working to destroy your attack force. It seems to then build defensive units almost exclusively as well. The AI doesn't know that offensive is better than defense in Civ III.

[vmxa1]

Among its many problems. I do not mind it being less than optimal. If it was super smart, I could not over come its extra units and such.

[Blue Moose]

Of course, if it was super-smart, it wouldn't need extra units.

[Catt]

Originally posted by Blue Moose
Units that are one tech behind should still be able to adequately defend and attack (though are a noticebly reduced effectiveness). Units 2 or more tech progressions back should get slaughtered nearly all the time. So warriors should get killed off easily by bowman and knights (and exceptions should be so rare you should be almost guaranteed not to see them). Cavalry should slaughter any ancient age unit. Riflemen should be at the end of their lifespan defending against tanks, and so forth.

Originally posted by Catt
. . . . mean that radically increasing the power disparity between old and new units tends to unbalance the game towards the human, making it easier. Maybe more in tune with the real world, but, in my experience and IMHO, less of an engaging game.

Originally posted by vmxa1
Sorry Catt, that is a very weak counter arguement. All that you said is true, but the change will not materially impact the game. How many times in a given game do these types of battles occur and the older units win? Most swear they have never seen them. If that is so they are not making enough impact to matter. I think they are not making an impact, except in the perception of the game, so correcting it will not hurt the civs much.

You could be right, but I certainly see a lot of battles between units that differ by two tech progressions - and the smart human player would rush like hell with the first significant tech lead. Get to chivalry (knights) before neighbor has large numbers of pikemen (or is without iron)? Bye-bye neighbor. Get to cavalry and a neighbor has no saltpeter or hasn't upgraded all pikes? Bye-bye neighbor. And so on.

I haven't played with modded HPs under 1.29f, so it might be diffferent, but I suspect it also has a proclivity to be worse from a gameplay perspective. It's been my experience that the changes to the tech progression in 1.29f have allowed some civs to essentially fall out of the tech race - they get behind and there is no way for them to catch up. Modded HPs in such a circumstance would exacerbate the already nearly overpowering strength of a tech lead -- it essentially would lock out a civ that doesn't stake a claim very early.

Don't get me wrong - I'm not saying don't mod the game to one's liking; I'm just saying that modding carrires risks - one of the principal risks being that gameplay balance is messed up and the game loses some of its challenge / fun / strategic options. Not true for all players and not true for all mods, but I eventually stopped modding and went back to the straight game because I felt I couldn't make the game more interesting through my mods. Others have had more success.

Originally posted by vmxa1
Making the AI smarter is another thing altogerther.

Yes, indeed. This is the real challenge.

Catt