A kinder, gentler point about the combat system

[supremebob]

My veterants riflemans are being shreded by axemans. How is that realistic even with upgrades?

Robert

[Tiberius]

It's not realistic. Get used to it or mod your game (I am experimenting with the latter lately).

Have you checked the odds when attacking the axeman? Sometimes they have insane bonuses from city and terrain and whatnot. If you were attacking a city, bring in some catapults and it will be much easier.

What bothers me is when I see a samurai pillaging my city, I attack him with my full health, promoted grenadier (2 or 3 promotions, I don't remember exactly), the odds are 13 to 7, and then I lose. Now that bothers me. It happened to me twice in my last game within a few turns.

Don't ever attack with wounded units. They are much weaker than they seem. The only situations when I attack with them is when the enemy is even more damaged than my unit.

[supremebob]

Thank you Tiberius for your reply,

I was tired last night so I tried to use as few words as possible . However it was not so much a complaint as to a reaction of stupor: in past civ games when you attaked or defended with units that were far superior technologicaly, the odds were much more on your side. I like the new combat system in civ 4 due to it being less predictable (miriad of new factors to take in consideration) and overall more fun. At least its a step in the right direction but still, it sould not take four grenadiers to take a city with two archers defending it...

Robert

[Hamdinger]

Just for fun, I wrote a Perl script to test my proposed change from earlier in the thread, i.e - to not modify a wounded unit's strength for the purposes of hitting; the amount of damage remains modified by health.

The first test is with an attacking unit that has twice the base strength of the defending unit (assume no modifiers for this combat). Each row of the table below represents the attacker with different health values at the start of combat (the defender is always a full 100 health). The first table represents the probabilities of winning in the current system (A is the attacker's probility of winnnig, D the defender's):

HP 100    A  99.12%   D   0.88%
HP  95    A  97.61%   D   2.39%
HP  90    A  94.10%   D   5.90%
HP  85    A  93.01%   D   6.99%
HP  80    A  91.68%   D   8.32%
HP  75    A  73.34%   D  26.66%
HP  70    A  69.77%   D  30.23%
HP  65    A  51.26%   D  48.74%
HP  60    A  46.64%   D  53.36%
HP  55    A  26.75%   D  73.25%
HP  50    A  22.66%   D  77.34%
HP  45    A  11.27%   D  88.73%
HP  40    A   3.34%   D  96.66%
HP  35    A   1.03%   D  98.97%
HP  30    A   0.56%   D  99.44%
HP  25    A   0.02%   D  99.98%
HP  20    A   0.00%   D 100.00%
HP  15    A   0.00%   D 100.00%
HP  10    A   0.00%   D 100.00%
HP   5    A   0.00%   D 100.00%

The second table represents the probabilities of winning using my proposed system:

HP 100    A  99.12%   D   0.88%
HP  95    A  98.03%   D   1.97%
HP  90    A  95.76%   D   4.24%
HP  85    A  95.76%   D   4.24%
HP  80    A  95.76%   D   4.24%
HP  75    A  85.52%   D  14.48%
HP  70    A  85.52%   D  14.48%
HP  65    A  74.14%   D  25.86%
HP  60    A  74.14%   D  25.86%
HP  55    A  57.06%   D  42.94%
HP  50    A  57.06%   D  42.94%
HP  45    A  46.82%   D  53.18%
HP  40    A  26.34%   D  73.66%
HP  35    A  19.51%   D  80.49%
HP  30    A  19.51%   D  80.49%
HP  25    A   3.90%   D  96.10%
HP  20    A   2.60%   D  97.40%
HP  15    A   1.73%   D  98.27%
HP  10    A   1.73%   D  98.27%
HP   5    A   0.51%   D  99.49%

So for example, a 20 base strength unit that is 50% wounded to 10 strength will have a 57% chance to beat a 100% healthy 10 strength unit in my proposed system, as opposed to a 22% chance in the current system.

[Hamdinger]

And here are some more tables. This time the attacker and defender have the same base strength. First the table representing the current system:

HP 100    A  50.00%   D  50.00%
HP  95    A  34.58%   D  65.42%
HP  90    A  31.41%   D  68.59%
HP  85    A  28.20%   D  71.80%
HP  80    A  15.74%   D  84.26%
HP  75    A  13.47%   D  86.53%
HP  70    A   6.40%   D  93.60%
HP  65    A   2.29%   D  97.71%
HP  60    A   1.70%   D  98.30%
HP  55    A   0.53%   D  99.47%
HP  50    A   0.10%   D  99.90%
HP  45    A   0.06%   D  99.94%
HP  40    A   0.01%   D  99.99%
HP  35    A   0.00%   D 100.00%
HP  30    A   0.00%   D 100.00%
HP  25    A   0.00%   D 100.00%
HP  20    A   0.00%   D 100.00%
HP  15    A   0.00%   D 100.00%
HP  10    A   0.00%   D 100.00%
HP   5    A   0.00%   D 100.00%

And the table representing my system:

HP 100    A  50.00%   D  50.00%
HP  95    A  37.70%   D  62.30%
HP  90    A  37.70%   D  62.30%
HP  85    A  37.70%   D  62.30%
HP  80    A  25.39%   D  74.61%
HP  75    A  25.39%   D  74.61%
HP  70    A  17.19%   D  82.81%
HP  65    A   8.98%   D  91.02%
HP  60    A   8.98%   D  91.02%
HP  55    A   5.47%   D  94.53%
HP  50    A   1.95%   D  98.05%
HP  45    A   1.95%   D  98.05%
HP  40    A   1.07%   D  98.93%
HP  35    A   1.07%   D  98.93%
HP  30    A   0.10%   D  99.90%
HP  25    A   0.10%   D  99.90%
HP  20    A   0.10%   D  99.90%
HP  15    A   0.02%   D  99.98%
HP  10    A   0.01%   D  99.99%
HP   5    A   0.00%   D 100.00%

The point at which it becomes hopeless for the attacker is a little better in the second chart, but not hugely different. So a wounded unit is still at a tremendous disadvantage to an equivalent unit that's healthy.

I did notice one potential problem with my system. The probabilities are not as smooth across different health values - it's a bit of a step function. I don't know if this really matters, but it bothers me from an aethetic point of view.

[Eratosthenes]

Yeesh.

A unit at half health has lost a lot of its men, equipment, supplies, whatever. It will not be as effective as a fully equipped unit.

It might even be just as effective as a less advanced group with full health.

This allows for less advanced troops to win via attrition, as has happened before in history (Zulus vs British,

Somehow it seems those complaining about it want their advanced tanks to never lose against lesser units...evah! Even if they've just beaten off an entire armada of enemies, and have 1% power left.

A full squad of cavalry would most likely be able to find a way to defeat a badly damaged squad of tanks, though likely with substantial losses. Heck, throw enough bodies at something, and you're bound to eventually defeat it!

[Hamdinger]

Of course a half health unit would not be effective as a fully equipped unit. The question is how much less effective should it be. And it's not a matter of realism, or having advanced units never lose to primitive units, or anything like that (for me anyway, I can't speak for others). It has to do with balance and intuition. Some of the current probabilities are not intuitive to me, based on the way the information is presented in the game. I suspect that's true for others as well.

I don't think combat is broken. I'm enjoying the game just fine with combat the way it is. I'm merely wondering if it could be even better. If nobody likes my idea, that's fine. It was just a suggestion.

[Arkalius]

I did a bit of testing... in case anyone cares about this.

A unit at 64hp has even odds to beat a full-health unit that has half it's base strength. What does this mean? A 6.4/10 knight has a 50/50 chance of beating a 5/5 swordsman (ignoring all bonuses). This kind of gives you an idea of how much of a penalty having low health can give you.

[Eratosthenes]

Units don't have "health" in Civ4, apparently...they only have Power. They start every fight at 100 health (think of it as 100%), and after the fight, have their Power reduced to whatever % they have left.

So a Power 20 tank gets in a fight, and is reduced to 60 "health", he'd then have 12/20 Power.

In his next fight, he might be reduced a further 50%, and be at 6/20 Power.

Both fights he'd start at 100%.

How much each "hit" in combat does in health/percentage is based on the ratio of the attacker and defender's Power. I`m sure someone has posted the formula.

So an attacker that is at 10/10 vs a defender at 10/20 should have 50-50 odds...assuming I've not gotten something terribly, terribly wrong.

[Tiberius]

You've gotten something terribly, teribbly wrong

Take a look here:

http://www.apolyton.net/forums/showthread.php?s=&threadid=140622

While it's true that the two units in your example will have the same chances to win one round, the 10/20 unit will need to win twice as many rounds to win the battle, because he is at 50% health, while the 10/10 unit is at 100% health.

Hamdinger, I like what you've done. IMHO it looks good.
The fact that a half health, twice as strong unit has only 1/4 chances to win a battle seems wrong to me. With your system the chances are slightly over 50%, which seems right.

I was thinking on an intermediary solution, to make it more realistic, because well, it is logical that a wounded unit is not as strong as a full health unit.

Right now a wounded unit loses 10% strength for every 10% health losses. What would the situation look like for a 5% strength loss / 10% health loss ? So a unit at 50% health would still have 75% of its initial_strength.

Hamdinger, how does this sound to you? Could you use your script to generate some values?

[Xelvonar]

I'm not sure, but I think he's got it right. Based n my reading of that link, a unit that normally has a strength of 20, but that enters a combat with strength 10 due to previous combats is treated as a strength 10 unit in every respect.

So I would say str 10 vs. str 10 is 50-50, no matter what these units full strengths are.

[swat-spas2]

"Right now a wounded unit loses 10% strength for every 10% health losses. What would the situation look like for a 5% strength loss / 10% health loss ? So a unit at 50% health would still have 75% of its initial_strength."


This would seem to be a reasonable compromise that should keep the aesthetic of a smooth probability curve. Take the strength and its reduced as normal, and the attacking power is reduced at a 50% ratio to damage. The fighting power of a unit is certainly reduced by the loss of men, but the potential power is generally due to the tactical employment of the remaining men. As such, while a damaged unit will have less flexibility (fewer reserves, less immediate hammer blow), it should still have an advantage over a lesser equipped unit. Throw enough knights at a tank, maybe eventually they would run out of ammo or fuel.. but you'd get slaughtered in the mean time and it should take alot of blows to get to that point.

[Tiberius]

Originally posted by Xelvonar
I'm not sure, but I think he's got it right. Based n my reading of that link, a unit that normally has a strength of 20, but that enters a combat with strength 10 due to previous combats is treated as a strength 10 unit in every respect.

So I would say str 10 vs. str 10 is 50-50, no matter what these units full strengths are.

He is treated as a unit with a strength of 10, but he has only 50% of his hitpoints left. So he has equal chances to win a round of battle, but he has to win 5 rounds to win while the other unit only 3 rounds (assuming it has a strength of 10/10). Therefore he has much less chances to win the whole battle.

Originally posted by DeepO
Something important I noticed, which wasn't mentioned before: units will start at 100 hp at full health. A unit that has been injured to 80% starts at 80 hp, not the full hundred.

This means that injured units don't need that many hits, which increases the chances for the full health units.

An example: suppose a archer, wounded to 2 gets bonusses up to 4. Suppose it is attacked by a horse archer of full health, srength 6.

That archer starts at 67 hp, whereas the HA starts at 100 hp. The chance of winning each round is still 60% for the HA, and it will do 24 hp damage. Which means that after only 3 hits, the archer is dead, instead of the 5 hits needed like reported above.

Which means that slaughtering wounded units gets an additional boost!

DeepO

Originally posted by DeepO

- If you are attacking with a wounded unit, your chances will go down (even if the modified strength is higher). Always put a full health unit on top of wounded ones.

- the inverse is of course also true: a MA, wounded to strength 9 (modified) can be easy defeated by a musket of strength 9. The MA has to hit the musket 5 times, while the musket only needs to hit the tank once. The chance of that occuring is 50% each round.

DeepO

[Tiberius]

More clearly than this cannot be explained:

Originally posted by BgT
Ok... so basically...emmm the summary of this post is that those odds that the game gives us (right click over the unit) are useless, right?
I've had one case where I actually loaded the same savegame 4 times (with usual 30 minute load time ) and the outcome was completely the same. So it seems the seed of the randomizer is some game data and is always the same? This specific case was:
damaged tank vs. wounded maceman
I think maceman had 5.6 strength and tank around... hm 9-11. Well, the numbers after right click were showing 10.4 vs 8.5 for the tank. In all 4 cases the maceman won. Can someone explain to me how is this possible? If it's because the wounded unit stuff - why are those numbers there anyway?! 10.4-8.5 I mean. If that would be the odds, than the probability of maceman winning all 4 fights is 0.04 or 4%.

Originally posted by DeepO

The numbers are not completely wrong, but you don't take into account how wounded a unit is.

This is an interesting example, so let me try to work it out:

You had a tank, normally at strength 28, at let's say 9 after damage. This means his hp are 9/28 * 100 = 32.

His strength, modified, is the 10.4 you see. So T = 10.4

The maceman was near full health at 5.6. his hp were 5.6/8 * 100 = 70.

His strength, modified is 8.5. So M = 8.5.

Now look at the numbers: the tank deals a little bit more damage, and has a little bit better chance of winning a round. It also is less than half the hp, so in a way only half the strength.

The maceman has little less odds of winning a round, and does a little less damage if it would win, but it only has to eat up 32 hp.

The difference between modified strengths M and T is not large at all: it's only ~10%. And from the looks of it, your tank needs to hit 4 times, before getting hit twice. I'd say the chance of him winning are less than 15%. Which means that the maceman is 6 times more likely to win this battle...

DeepO

[Hamdinger]

Tiberius, your wish is my command.

Just to be clear on what I'm doing, the effective strength of the unit will be reduced half as much due to damage. So for example, a 10 base strength unit with 50 HP will have a 7.5 strength for the purpose of determining if it hits in a round of combat, and it will also have a 7.5 strength for the purpose damage. The effective strength would have been 5 in the current system.

I think this is what Tiberius is suggesting. So first the table for an attacker with twice the base strength:

HP 100    A  99.12%   D   0.88%
HP  95    A  97.83%   D   2.17%
HP  90    A  97.61%   D   2.39%
HP  85    A  97.36%   D   2.64%
HP  80    A  94.10%   D   5.90%
HP  75    A  87.74%   D  12.26%
HP  70    A  86.88%   D  13.12%
HP  65    A  85.94%   D  14.06%
HP  60    A  73.94%   D  26.06%
HP  55    A  61.22%   D  38.78%
HP  50    A  59.41%   D  40.59%
HP  45    A  40.23%   D  59.77%
HP  40    A  38.42%   D  61.58%
HP  35    A  36.56%   D  63.44%
HP  30    A  18.31%   D  81.69%
HP  25    A  17.05%   D  82.95%
HP  20    A  15.80%   D  84.20%
HP  15    A   4.38%   D  95.62%
HP  10    A   3.94%   D  96.06%
HP   5    A   3.53%   D  96.47%

Now the table for an attacker with equivalent base strength:

HP 100    A  50.00%   D  50.00%
HP  95    A  36.15%   D  63.85%
HP  90    A  34.58%   D  65.42%
HP  85    A  33.00%   D  67.00%
HP  80    A  20.49%   D  79.51%
HP  75    A  19.28%   D  80.72%
HP  70    A  18.08%   D  81.92%
HP  65    A   9.04%   D  90.96%
HP  60    A   8.34%   D  91.66%
HP  55    A   7.65%   D  92.35%
HP  50    A   6.99%   D  93.01%
HP  45    A   2.47%   D  97.53%
HP  40    A   1.03%   D  98.97%
HP  35    A   0.89%   D  99.11%
HP  30    A   0.77%   D  99.23%
HP  25    A   0.12%   D  99.88%
HP  20    A   0.10%   D  99.90%
HP  15    A   0.09%   D  99.91%
HP  10    A   0.03%   D  99.97%
HP   5    A   0.02%   D  99.98%

I could even make Excel charts for the curves, but I have no idea how to post that.