Who hates Civ3's combat system?

Is Civ3's combat flawed?

  • Yes!

    Votes: 58 21.8%
  • It has it's flaws but I can live with them.

    Votes: 125 47.0%
  • No! What are you talking about? It's a great system.

    Votes: 83 31.2%

  • Total voters
    266
I think some would not like the popup or the delay.

Here a suggestion:

Civ always has one unit highlighted. Simply getting with the cursor over an enemy unit could bring up a probability of victory, e.g. 33%, 66%...

This could be displayed right over the unit/in the tile that you want to attack.

Just as an example. Alpha Centauri was awful for me, I did not like the graphics, but it had a similar and very good concept.
 
CB: If I'm getting you arightly, you are concerned with probable rounding errors in the 5e-4 range? That's not an error you're gonna spot this side of a tolerably thorough statistical analysis. It's an order of magnitude smaller than the smallest combat bonus.
 
Originally posted by Commander Bello
To be honest, I am a little bit surprised, that this obvious fact constantly is neglected by interested parties.
I can only assume, that some of you desperately want to keep the current system. So, any argument against it seems to be just inexistent for some of you.
Very well then. Let's drink our milk.

As a bystander, actually, it looks to me you are the one who seems desperate to want to have it changed. The rounding issue you mention (if it exist) is neligible by any statistical standard. This means that making it perfect will have neligble impact to the combat results.

I like to have it changed slightly (only slightly, anything more I am against it). For me, I think the simplest solution is actually increase the HP. I have tried a mod game with a simple change of doubling each experience level's HP. I am glad to say that I like that mod very much since combat is alot more predictable, with extreme luck happening sparingly. And the best part is that ganging up several weaker troops to wack a tough defender becomes a really viable strategy. The only problem I find with this simple mod is that each combat round takes significantly longer.
 
First, i like the combat system. Whats the point in having one tank crushing every musketeer you encounter because you are much stronger by definition.

In world war II often russion soldiers attacked german tanks without real weapons. Armed with bottles filled with gasoline (know ans 'molotov cocktails'), they used the ruins of their cities to get close and throw the burning bottles at the tanks. They didn't won at a regular basis, maybe not often. But they achived some success, even when they are just better equipped warriors ;)

And real world combat has a lot of factors which can't be influenced by the commander in chief, the allmighty leader represented by the civ player. So your tank lost: Well, maybe the general of this unit just was a dumbass, only in this position because his father and grandfather were generals too and helped him in this position. Maybe your general is good at his job, but gets hit at the start of the battle from one lonely bullet, arrow, stone...
Maybe the enemy who knew the territory laid some trap, maybe your advanced technology doesn't work correctly because of bad weather, maybe maybe maybe...

A lot of 'maybe' no one on top of the state can influence. Thats where probabilities are. Better tech makes victory more likely, but don't garanties it. That how it works in civ and in real world.
 
The Last Conformist:

Personally, I am not (very much) concerned about it. I just did my statements to explain, that the use of *any* (pseudo)RNG suffers from certain inaccuracies.
Depending on how the engine deals with them, it might give a different result from the one you would have got by using dice.
For me, it seems just to be a matter on what level the roundings will occur. In the worst case, the rounding would occur already at the level of the modified defense value as of my example, and then I would think that it would have a certain significance. Nevertheless, my personal opinion (or just call it my hope) is that it will occur much later in the course of the various calculations.

Of course, as long as we don't know exactly about the coding a lot of the statements given in this thread are just assumptions. I am pretty aware of this, and for that emphasized *my* assumptions each and any time (at least, I tried to do so).

All I tried to do was to explain that those inaccuracies are one factor (amongst others) that may lead to experiences, which are not satisfying some of the players.
Personally, I could think of settings which I believe to lead to "better" results (at the moment of writing this, Qitai has submitted a proposal how to do this). Others may think differently about this, and I accept that.
What I do not like, is that some statements here seem to mark me as being a dumbass, just because I am pointing at possible weaknesses.
But even this happens from time to time in a discussion about such a controversial topic as the battle results.
So, please allow me to raise my cup of coffee and to cheer anybody, who contributes in a pertinent manner to the discussion. :goodjob:

Addendum:
As I have just read Qitais statement, the suggestion he makes is indeed the one and obvious solution with the given means to counter the current weaknesses, since it counters (to a certain degree) any of the assumed "streaks".
This was exactly my approach in late PTW. Personally, I found that 3/5/7/9 hitpoints (for drafted/regular/veteran/elite) proved to be a good solution. Ah.. the 9 hitpoints just because I gave some UU's the hitpoint bonus, so that there was a maximum of 10 hitpoints possible.
 
Actually, rounding errors could be done away with in CivIII combat.

The final defense value will always be a multiple of .05, so multiplying everything with 20 lands us with all integers.

Then have your RNG return an integer between one and the sum of the highest attack factor and the highest modified defense factor*. If the number is equal to or lower than the attack factor, the attacker wins that round. If the number is higher than the sum of the attack and defense factors, draw a new number. Else the defender wins. Could be optimized, but you get the picture.

* In the interest of modders, one should allow for higher numbers than are actually used in the Epic game, of course.
 
I could agree to a hit point increase, for example 2/4/6/8, but only as long as the artillery rate of fire was increased also, at a reasonable rate.

As for me being harsh and cruel, guilty as charged :(. Sorry, I can get worked up in these kind of debates. And when people get worked up, they speak more than they listen. Sorry about the harsh way I presented myself, it could have been much more effective if I were nicer.
 
By the way, I just did a test with 540 crusaders attacking 540 fortified spearmen (all regular, all on grassland).
The win ratio of the crusaders was around 76%, which seems to be pretty much as expected.
But there were 6 streaks of 4 spearmen to win in a row, so this happens almost every 90's time, and 6 streaks of 3 spearmen to win in a row.
So you have a chance of a little bit more than 2% to encounter a streak of 3 or more successful spearmen under the given circumstances.
 
Originally posted by Qitai
... For me, I think the simplest solution is actually increase the HP. I have tried a mod game with a simple change of doubling each experience level's HP. ...

Ugh, having a spearman with 8 hp is basically equal to having 2 spearmen 4 hp each except that this 8hp monster is difficult to kill and it costs same as a 4hp guy. Assuming that he is attacked by the same monster 8hp swordsman... It would be close to Age of Mythology then. Gods met on the field of the battle, wrestled from dawn to dusk for a week with short breaks and still it is unclear who is going to win. How do these guys heal? 1hp/turn in the open and full healing/turn with barracks?
 
It is not exactly the same. Since there is more combat rounds, the expected result is quite different (Some maths stuff which I will not go into). Try it. The longer battles are indeed an issue if the HP is increased that much. IIRC, healing actually is faster. I think it was 2hp per turn out in the field.
 
no, the healing is the same. if you make elites have the 8 health, then in a barracks it is 1 turn. if it has health bonuses, then it heals 5 one turn, and 5 the next. i tried.
 
Originally posted by Qitai
It is not exactly the same. Since there is more combat rounds, the expected result is quite different (Some maths stuff which I will not go into). Try it. ...

Apparently, fluctuation rate is very similar but it results in less casualities for the units. On average it is about 80-85% more dice rolls for each combat with 8hp than with 4hp each. Hence 2 results in a row of "less than 90% probability" independent dice rolls (spearman takes 1 hp off modern armor or archer taking 1 hp from mechanic infantry) will less likely contibute to combat resolution with 8 hp than with 4hp. For 3 "outbreaks" in a row it is hard to tell whether they can contribute and 4 are very unlikely to occur. Hence, in this case spearman cannot win against tank from the practical point of view. However, remote possibility still exists. And then, killing one strong defender with several weak attackers makes practical sense. It is hard to tell the analytical expression for the shape of probability distribution and hence determine variability. Distribution of combat results would be substantially more narrow with more hp.
 
well, on my current game i lost in a big battle a hell of alot tanks to cavalary - THAT really annoys me.

it were not just a few, a hell of alot - but after all balance is very important.


btw, i liked the combat system in Call to Power - you could group units - very nice :) - anyone else liked that ?
 
@ Legator:
I liked it as well :)

Ok, I just did some testing with 180 crusaders (attack 50) vs 180 spearmen (defense 20).
All units had 9 hitpoints, all units were unfortified on grassland.
Here are the results:
171 of 180 crusaders survived: 95 % (all percentages based on 180)
5 crusaders left with 1 hitpoint: 2.7% status: red
14 crusaders left with 2 hitpoints : 7.7% status: red
22 crusaders left with 3 hitpoints: 12.2% status: red
41 crusaders left redlined: 22.8 %
11 crusaders left with 4 hitpoints: 6.1% status: yellow
28 crusaders left with 5 hitpoints: 15.5% status: yellow
30 crusaders left with 6 hitpoints: 16.6% status: yellow
69 crusaders left yellowed: 38.3%
35 crusaders left with 7 hitpoints: 19.4% status: green
19 crusaders left with 8 hitpoints: 10.5% status: green
7 crusaders left with 9 hitpoints: 3.8% status green
61 crusaders left green: 33.9%

1 spearman left with 1 hitpoint
3 spearmen left with 2 hitpoints
2 spearmen left with 3 hitpoints
3 spearmen left with 4 hitpoints.
9 of 180 spearmen survived: 5%

Observations during the fight: Redlined attackers *seem* to do quite good. Quite a lot of the 1 and 2 hp crusaders managed to kill off the spearman, although he at this point still was yellow.
On the other hand, 180 tries may be to less to be significant.

Interpretation:
A battle of 180 elite crusaders vs 180 elite spearmen resulted in 95% of the attackers surviving. Accordingly, 5% of the unfortified defenders survived.
61.1% of the attackers are considered "torn down" (yellow and red). Additional 5% have been killed. 66.1% of the attackers are not able to continue fighting (without recovering).

Personally, I am quite fine with these results. Although the attackers were of double the strength as the defenders (A50 vs D22), still the "unpredicted" results occur, such adding an element of surprise. On the other hand, this element of surprise is not as high that it would spoil the fun (or the human battle plan)
Nevertheless, the attacking army is not able to "stomp" through the enemy territory, even less when considering the fact that the setup was to the maximum advontage of the attacker.

I will go on and test the same for spearmen, fortified in a size 6 city with walls.

[edit] It took the crusades 5 turns to heal (with the exception of 1, who in turn 6 still was recovering. I have no clue, why this one takes longer :confused: )
 
Originally posted by Commander Bello

[edit] It took the crusades 5 turns to heal (with the exception of 1, who in turn 6 still was recovering. I have no clue, why this one takes longer :confused: )

The healing rate increases if you increase the hps for units in the experience level tab of the editor. If you increase hps for units individually than the healing rate stays the same as for the basic unit.

And the combat in Call to Power was good, although the AI was too bad to make it interesting.
 
@ Roland:
I made use of the editor, of course. Although, it might be caused by the fact that I gave those 9 hitpoints to regulars.
I hope, there are no hidden modifiers for the combat experiences in the combat calculator.
 
To make my previous post more clear, because we are clearly not understanding one another.

I did the following test:

I gave conscripts 8 hps, regulars 12 hps, veterans 16 hps, elite 20 hps. Healing rate inside a city was 9 hps (without barracks), outside the city 4 hps.

In another test I gave the spearman unit 12 extra hps. I didn't change the hps for different experience levels (conscript 2hps, regular 3 hps, etc.). The healing rate was the same as in a normal game (inside city 2 hps, outside city 1 hp).

So the healing rate depends on the hitpoints you give to units in the experience level tab of the editor.
 
Ok, now I attacked with 180 crusaders (A=50, hp=9) vs 180 fortified defenders in a town with walls (D=37, hp=9).
As stands true for the posting above, all units are regarded to be elite (for that, the 9 hp), but are just regulars with 9 hp. This *SHOULD* make no difference, since they are no fast-moving units, and therefore there shouldn't be any difference.

Here are the results:
119 of 180 crusaders survived: 66.1 % (all percentages based on 180)
11 crusaders left with 1 hitpoint: 6.1% status: red
25 crusaders left with 2 hitpoints : 13.9% status: red
15 crusaders left with 3 hitpoints: 8.3% status: red
51 crusaders left redlined: 28.3 %
22 crusaders left with 4 hitpoints: 12.2% status: yellow
20 crusaders left with 5 hitpoints: 11.1% status: yellow
14 crusaders left with 6 hitpoints: 7.7% status: yellow
56 crusaders left yellowed: 31.1%
7 crusaders left with 7 hitpoints: 3.9% status: green
5 crusaders left with 8 hitpoints: 2.8% status: green
0 crusaders left with 9 hitpoints:
12 crusaders left green: 6.6%

15 spearmen left with 1 hitpoint: 8.3% status: red
14 spearmen left with 2 hitpoints: 7.8% status: red
9 spearmen left with 3 hitpoints: 5.0% status: red
38 spearmen left redlined: 21.1%
5 spearmen left with 4 hitpoints: 2.8% status: yellow
7 spearmen left with 5 hitpoints: 3.9% status: yellow
10 spearmen left with 6 hitpoints: 5.6% status: yellow
22 spearmen left "yellowed": 12.2%
2 spearmen left with 7 hitpoints: 1.1 % status: green
0 spearmen left with 8 hitpoints:
0 spearmen left with 9 hitpoints

Interpretation:
A battle of 180 elite crusaders vs 180 elite spearmen resulted in 66.1% of the attackers surviving. Accordingly, 33.9% of the unfortified defenders survived.
59.4% of the attackers are considered "torn down" (yellow and red). Additional 33.9% have been killed. 93.3% of the attackers are not able to continue fighting (without recovering).

Personally, I think that both tests show that increasing of the hitpoints benefits less random results, but still doesn't give the (strong) attacker any "immortality".
In the second test, I doubt that - although the city might be conquered - the attacker will be very happy with it.
 
Originally posted by Roland Johansen
To make my previous post more clear, because we are clearly not understanding one another.

I did the following test:

I gave conscripts 8 hps, regulars 12 hps, veterans 16 hps, elite 20 hps. Healing rate inside a city was 9 hps (without barracks), outside the city 4 hps.

In another test I gave the spearman unit 12 extra hps. I didn't change the hps for different experience levels (conscript 2hps, regular 3 hps, etc.). The healing rate was the same as in a normal game (inside city 2 hps, outside city 1 hp).

So the healing rate depends on the hitpoints you give to units in the experience level tab of the editor.

Ahh.. ok, I got it. Well, that seems fine for me as well, since otherwise it would take ages for a unit with more hitpoints to heal.

Well, my tests of course are just very rough. I assume that I should do them with at least the triple amount of troops involved before one can assume them to be really significant. But, they may give a first hint.
What I am interested in, is there somewhere a calculator available which supports both, much higher A/D values and much higher hitpoints?
The ones I've checked so far, don't support those figures...
 
In summary, it seems that most people like the combat system as is, including myself for that matter, though there is a vocal minority who may have been on the receiving end of one random streak too many for their liking. This is often blamed on the apparent "unpredictability" of the pRNG.

As noted in above posts you can reduce this "unpredictability" by increasing hit points. This solution is:
a) available through an easy mod using the existing combat system, and therefore not an indication of a fundamental flaw with the combat system
b) actually a form of increasing the differences in strength between the various units, rather than addressing the randomness of the combat calculations.

Personnally, I like the balance as it stands now. It appears to me to have been determined after substantial testing and feedback, and provides a reasonable simulation of the historical process of the cycles of defensive superiority followed by times of offensive superiority that have shaped warfare through the centuries.

Additionally, the "unpredictability" of CIV III combat actually is more predictable than it is given credit for. You can mathematically determine with substantial precision the required force to beat a given stack of defenders with a given probability of success. You don't actually have to do this maths to play the game well, but it is interesting to note how much better your overall odds of success improve with the addition of a few more troops.

In later ages, there are usually so many troops on the battlefield that the occassional streak is less irritating than say in the ancient age. This is also reflected in the concerns raised by a few posters under this thread. Their actual examples relate to the ancient age and involve very small numbers of troops. eg 4 swords versus two fortified spears for a must take city doesn't seem to me to be a sufficient force. The solution is simply wait until you have a sufficient force for your objective.

Strengthening the stronger unit over the weaker unit, either through hit points or base unit strength modifications, results in a more significant advantage given to the first Civ to research the relevant tech. In terms of gameplay, and reality for that matter, you should have some chance even if you are behind the latest available military tech. I think the current balance and the current combat system reflects this well.
 
Back
Top Bottom