RNG combat. Can anything be done about it?

1SDAN said:
What? I hate winning on low odds FAR more than losing on low odds. It feels cheap, like I'm cheating, equal to worldbuildering my unit to full health and my opponent to near zero.

I like to feel like I deserve to win or deserve to lose. This isn't about making the game easier, this is about making the game more rewarding.

Gambling is not rewarding, it's surprising. One requires skill, the other luck.
Click to expand...

You seem to operate on some kind of Through the Looking-Glass logic :crazyeye:. When you risk something against 10% odds, which in real life represents desperate attempt, how can the victory be cheap? Cheap is when you kill with 100% vs nothing.

And speaking about rewarding -- there is nothing rewarding about online reality with no "surprises", the way you like to call it. Game already pretty predictable as is, the unusual things happening in game make me feel rewarded. Like when we changed known locations of goody huts with randomized locations. Or when you open the game and see that Babylon somehow defeated Persia -- perhaps one lucky 10% battle tipped the scale of the war. Real history evolves in the direction of probable outcomes -- but nothing is 100% guaranteed except for the death and the tax slider.

And if we go down the path of equating 10% of odds to 0%, what about Great Artist born with 10% odds when you are hoping for Scientist with 90% odds? Should we doctor those occurrences as well? Our real life is so interesting precisely because these low probability events are still possible. Max Planck was extremely gifted when it came to music: He took singing lessons and played the piano, organ, and cello, and composed songs and operas. However, instead of music, he chose to study physics.
Munich physics professor Philipp von Jolly advised him against going into physics, saying, "in this field, almost everything is already discovered, and all that remains is to fill a few holes." Planck replied that he did not wish to discover new things, only to understand the known fundamentals of the field. What were the odds of him founding Quantum Physics?

In the war of 1812, 24 native Americans attacked 200 US troops at Brownstown. The US Troops, seeing the small number of attacking troops, broke and ran. The Indians achieved the military victory against the odds, even though they were outnumbered 8 to 1. Or, another example -- 100 years war Battle of Agincourt saw 3000 English Bowman be victorious over an overwhelming superior French army of 30,000. The most famous victory of Cesar was battle of Alesia -- at one point the Romans were outnumbered by the Gauls by four to one. List goes on and on. Some low-odds victories have been pivotal to the conflict they were part of, or provided inspiration for the other forces.A battle against odds is not a gambling -- it is a heroic or desperate attempt which can pay off sometimes.

Finally I would discourage anyone from thinking that 10% are diminishingly negligible odds. Our genome is only 2% different from Chimpanzees. Look at the difference it makes :cooool:
 
Last edited:
Obviously the discussion here is one that has played out a million times in the context of other strategy games and the answer is simple: how much randomness you like in your strategy games is a matter of taste. Some don't like it at all and others believe it is a fundamental feature that needs to be accounted for as part of ones strategy.

Civ 4 obviously goes by the latter philosophy. If you know that your unit only has a 90% chance of winning, then you're confident but not certain that you will win, and you need to build that into your strategy. Of course there are some cases where you simply can't do this and the RNG details everything (Frances conquest UHV in older revisions of RFCE was infamous for this), but Civ 4 is designed around randomness and you just need to account for it.

However, for combat specifically, there's so many non RNG related things you can do before the final dice roll to stack the deck in your favor that adding any further guards against RNG would eviscerate the balance. Yes, it's frustrating to lose a 90% battle, but combat and war as a whole is already so skewed in the players flavor that RNG helps to act as a sort of handicap.
 
Tigranes said:
Or, another example -- 100 years war Battle of Agincourt saw 3000 English Bowman be victorious over an overwhelming superior French army of 30,000. The most famous victory of Cesar was battle of Alesia -- at one point the Romans were outnumbered by the Gauls by four to one.
Click to expand...

Longbow on forested hill has very good odds even against many units, especially when taking first strikes into account.

One highly promoted Legion in a fort also has a fighting chance against four Axemen.

Orbii said:
Civ 4 obviously goes by the latter philosophy. If you know that your unit only has a 90% chance of winning, then you're confident but not certain that you will win, and you need to build that into your strategy. Of course there are some cases where you simply can't do this and the RNG details everything (Frances conquest UHV in older revisions of RFCE was infamous for this), but Civ 4 is designed around randomness and you just need to account for it.

However, for combat specifically, there's so many non RNG related things you can do before the final dice roll to stack the deck in your favor that adding any further guards against RNG would eviscerate the balance. Yes, it's frustrating to lose a 90% battle, but combat and war as a whole is already so skewed in the players flavor that RNG helps to act as a sort of handicap.
Click to expand...

In the base game you can account for RNG because you are not expected to steamroll two civs with just your starting warrior and have all the time in the world to prepare a suitable attack force for a domination or conquest victory against opponents who start out at more or less the same level as yourself while never having to worry about flips or instability chipping away at your gains. It's a different story for the likes of Persia, Rome, Greece and the Mongols in DoC. With them you NEED to conquer something immediately with your starting stack if you go for the UHV or you don't stand a chance. You can't just rely on whatever cities you flip or found in the beginning to build up an army big enough to overcome the RNG.
 
I do agree that it's a matter of taste but it's a misconception to treat the question as an isolated aspect. like you can't take Civ4 as is and just slap on deterministic combat or extra rules that prevent "sure" victories from resulting in defeat (although I take issue with the whole idea that 90% is essentially certain, it just shows that you shouldn't show probabilities to people because most people just can't handle them), or other ideas like logistics etc. Game design is always a holistic problem and that's especially true for combat rules for a game with a military aspect.
 
Leoreth said:
like you can't take Civ4 as is
Click to expand...

But that's not what we are doing. Here we are taking an already highly modified version of Civ4, so it stands to reason that some of those modifications would require further changes to make other aspects of the game work better as a whole.
 
For one thing, I only dislike RNG in combat, Great People are roughly equal enough that I don't mind getting a 10% Artist.

Secondly, I like playing Civ 4 as a game of planning. I like to build up as good of a strategy as I can through multiple games, while trying to adapt to the slight changes in the behavior of the many AIs.

Finally, I'd like to point out that it's probably not even my dislike of RNG that makes me oppose Civ 4, but my dislike of unforseeable 1 hit KOs. In every Kingdom Hearts game I've played, I've always hated the attacks that kill me in one hit. What I don't hate however are the attacks that take me down to 1 HP or the attacks where the instant kill is given enough time and obviousness that I can react and try to mitigate/avoid it.

I understand that these are percents, and that a high percent is not a certainty. I understand all of the historical precedents. Despite all of this, capturing a city at a 90% chance of losing feels wrong because I did not deserve it. I understand completely that 90% chance of a loss means there will be 10% chance of a win, I don't feel bad because I see it as a sure loss, I feel bad because I see it as completely unrepresentative of how bad I really am.

When I'm a bad player, I want to be told I'm a bad player, not be lied to and called a good played because I got lucky on a dice roll. It feels patronizing.

Also, can you please not say I think 0% is equal to 20%? I don't. The issue at hand is that I don't like RNG, not that I'm an idiot.
 
1SDAN said:
For one thing, I only dislike RNG in combat, Great People are roughly equal enough that I don't mind getting a 10% Artist.

Secondly, I like playing Civ 4 as a game of planning. I like to build up as good of a strategy as I can through multiple games, while trying to adapt to the slight changes in the behavior of the many AIs.

Finally, I'd like to point out that it's probably not even my dislike of RNG that makes me oppose Civ 4, but my dislike of unforseeable 1 hit KOs. In every Kingdom Hearts game I've played, I've always hated the attacks that kill me in one hit. What I don't hate however are the attacks that take me down to 1 HP or the attacks where the instant kill is given enough time and obviousness that I can react and try to mitigate/avoid it.

I understand that these are percents, and that a high percent is not a certainty. I understand all of the historical precedents. Despite all of this, capturing a city at a 90% chance of losing feels wrong because I did not deserve it. I understand completely that 90% chance of a loss means there will be 10% chance of a win, I don't feel bad because I see it as a sure loss, I feel bad because I see it as completely unrepresentative of how bad I really am.

When I'm a bad player, I want to be told I'm a bad player, not be lied to and called a good played because I got lucky on a dice roll. It feels patronizing.

Also, can you please not say I think 0% is equal to 20%? I don't. The issue at hand is that I don't like RNG, not that I'm an idiot.
Click to expand...

I don't quite get what the alternative system is. Also, you're not getting lucky on "a dice roll". Instead, you're getting lucky on quite a few. There are several rounds of combat, wherein the units have a chance to damage each other, which is why the winner often gets damaged, too. That there are so many rounds of combat is why the odds seem to be in the 90s or 0s so often. A much weaker unit can win a round or two, but it's so much less likely for them to win all of them. When you get <10% odds of success, the units are actually fairly close in strength, because of how many dice rolls go into the outcome.
 
Jarlaxe Baenre said:
I don't quite get what the alternative system is. Also, you're not getting lucky on "a dice roll". Instead, you're getting lucky on quite a few. There are several rounds of combat, wherein the units have a chance to damage each other, which is why the winner often gets damaged, too. That there are so many rounds of combat is why the odds seem to be in the 90s or 0s so often. A much weaker unit can win a round or two, but it's so much less likely for them to win all of them. When you get <10% odds of success, the units are actually fairly close in strength, because of how many dice rolls go into the outcome.
Click to expand...
If I were to implement a simple alternate system, it'd just be that units would go to something like 0.1 health instead of dying if their health is > said something like 0.1.

It gives you a chance to salvage the situation.
 
1SDAN said:
If I were to implement a simple alternate system, it'd just be that units would go to something like 0.1 health instead of dying if their health is > said something like 0.1.

It gives you a chance to salvage the situation.
Click to expand...

So, you just never lose units unless you send almost-dead ones off into battle? Or is this just a thing that benefits attackers? I don't see what that has to do with the use of randomness in civ 4's combat system.
 
Under a system where units just go to 0.1 instead of dying immediately, does it make any difference? In a turn based game, if Unit A attacks Unit B on its turn and then loses but survives, surely Unit B just kills Unit A on its turn immediately after anyway? Or, assuming that Unit B doesn't attack, Unit A still needs to run away in a weakened state and try to avoid combat before inevitably being killed. Surviving combat a la Civ 5/6 just slows down the ultimate resolution, or at least that's how it feels to me.

In general, it feels like the way to deal with the fact that some of your units might die is simply to send more units out/have more units in defence. If it's a particular problem of units for civs that have to make something happen in a short space of time, a simpler answer might be to just add a few more starting units so they are less vulnerable to losing.

Without wanting to make a moral statement or start another flame war (I didn't bring enough units to this fight!), sometimes we get things we don't deserve and sometimes we deserve things we don't get, but the universe carries on regardless.

One of the great things I love about Civ4, and DoC, is that no game is exactly the same and every time you play you can try something different. Knowing the outcome of each fight would turn this game into something else, I feel.
 
Imp. Knoedel said:
But that's not what we are doing. Here we are taking an already highly modified version of Civ4, so it stands to reason that some of those modifications would require further changes to make other aspects of the game work better as a whole.
Click to expand...
No you're not, you're just vaguely gesturing to the fact that more needs to be done without ever being specific about what, or actually thinking through the consequences of the changes you want to make.
 
If the system bothers people so much I once again reiterate; work around it! World Builder exists for a reason! Cheat codes exist for a reason! It's not cheap; it's part of the game.

#endcheatshaming2k18
 
cmakk1012 said:
I’m a micromanager too, changing attack order and so on in reloads as opposed to straight changing odds, though.
Click to expand...
Just a quick remark, that's also what I do: change attack orders to strategically work around bad odds. For example, I learn in the original attempt that my lancer attack (64% odds) fails against 2 entrenched crossbows. However, I can reload and use the horse archer (10% odds, 60% withdrawal): horse archer attacks and withdraws. I try to send in the lancer again and his attack is still not a success. So, second reload, I first attack with the horse archer, secondly pillage a mine (the gold amount is randomized and uses up the bad roll), and on the third roll, the lancer (still the same 64% odds, against the second crossbow) finally succeeds. Then, I can next target the injured crossbow from the horse archer's attack...

I know that there is an option to just let the RNG roll differently each time I attempt the same battle. That would mean that statistically I would have to reload 10 times for a good chance of winning a 10% battle... And that is SO not what I'm doing, because that would mean to brute-force an undeserved lucky streak.

Jarlaxe Baenre said:
So, you just never lose units unless you send almost-dead ones off into battle? Or is this just a thing that benefits attackers? I don't see what that has to do with the use of randomness in civ 4's combat system.
Click to expand...

That is not what 1SDAN proposes, I think. You should still lose units when you attack against bad odds. In fact, 1SDAN told us he wants his units not winning if the odds are bad

Leoreth said:
you're just vaguely gesturing to the fact that more needs to be done without ever being specific about what, or actually thinking through the consequences of the changes you want to make.
Click to expand...
Absolutely true: all complaints so far have been rather vague, even the proposal of "please include Pyrrhic victory". So here is one specific suggestion that is not too complicated imho:

In all battles, when attacking units lose a battle when they had odds of [100% minus "two times their level"], they automatically withdraw instead of being lost.
  • Units with 2 EXP automatically withdraw if they lose at odds of 98% or better (Level 1 * 2)
  • Units with 10 EXP will withdraw instead of losing at odds of 94% or better (Level 3 * 2)
  • Units with 65 XP will withdraw instead of losing at 84% odds or better (Level 8 * 2)
  • Units with 170 XP will withdraw instead of losing at 74% odds or better (Level 13 * 2), well this is already a ridiculous high level but maybe the cap should be somewhere at 75% or 80%.
Keep in mind that a withdrawal leaves the unit with a health of nearly zero. To prevent "riskless" hit-and-run tactics with mounted/naval units, a withdrawal (under this system) should remove all remaining movement points, as well. So, the highly promoted attacking unit will afterwards stand exposed and helpless next to the enemy (unless protected through other units). I think that a movement-promotion (Flanking III or Mobility?) could allow to keep those movement points, but that option is not cheaply available.
Also note, that this proposed system would rarely affect battles with lower odds. If you rush your highly-promoted warlord into a battle of only 60% victory odds, you should be aware that he might die, unless you have him on level 20 (401 XP).

Also, in all battles, when defending units lose a battle when they had odds of [100% minus "two times their level"], they automatically force their attacker to withdraw instead of being lost themselves.
  • Attackers will withdraw even if they won a battle at odds of less than 2%, IF the defender has 2 EXP (Level 1 * 2)
  • Attackers will withdraw even if they won a battle at odds of less than 12%, IF the defender has at least 37 EXP (Level 6 * 2)
  • and so on...
In this case, the withdrawing attackers will retain the health they would have had if they had won and destroyed the defender; while the defenders would be set to a health of only 0.1. Unless there is another defending unit, and the attacker has movable units left, the defender will still be toast, but this is less the "undeserved victory" that 1SDAN bemoaned.
City defenders have quicker healing rates and so this defense benefit of not losing the highly promoted defender is substantial. On the other hand, dedicated defenders are rarely having more than 10 EXP as far as I know.

As far as I'm concerned, this neatly solves the issue of losing valuable units on a whim of the RNG; while still keeping the RNG as the deciding factor of all battles.
 
I am not just talking about a specific implementation of a withdrawal / incremental damage combat system but rather how to account for the ripple effects that such a change would have on completely different aspects of the game.
 
TsarAndreas said:
Under a system where units just go to 0.1 instead of dying immediately, does it make any difference? In a turn based game, if Unit A attacks Unit B on its turn and then loses but survives, surely Unit B just kills Unit A on its turn immediately after anyway? Or, assuming that Unit B doesn't attack, Unit A still needs to run away in a weakened state and try to avoid combat before inevitably being killed. Surviving combat a la Civ 5/6 just slows down the ultimate resolution, or at least that's how it feels to me.

In general, it feels like the way to deal with the fact that some of your units might die is simply to send more units out/have more units in defence. If it's a particular problem of units for civs that have to make something happen in a short space of time, a simpler answer might be to just add a few more starting units so they are less vulnerable to losing.

Without wanting to make a moral statement or start another flame war (I didn't bring enough units to this fight!), sometimes we get things we don't deserve and sometimes we deserve things we don't get, but the universe carries on regardless.

One of the great things I love about Civ4, and DoC, is that no game is exactly the same and every time you play you can try something different. Knowing the outcome of each fight would turn this game into something else, I feel.
Click to expand...
Such a system means that your units won't immediately die if there's another unit that can protect them. There's definitely balance issues with the suggestion, but I feel like it will give players and AI a chance to salvage bad RNG.
 
Well, what if it was just completely random? Not anything to do with whether they had >90% odds, just a 10% that any losing unit could survive at minimum health?
 
Leoreth said:
I am not just talking about a specific implementation of a withdrawal / incremental damage combat system but rather how to account for the ripple effects that such a change would have on completely different aspects of the game.
Click to expand...

Sure. A system that leaves surviving units in a majority of combat encounters will mean that all combat situations (war campaigns, sieges) are drawn out, because after rendering 10 enemy units to near-zero health, you need another round or another 10 fresh units, to wipe the survivors out. Also, if there are lots of withdrawals, everyone gains XP more slowly while mostly having weakened units that need to stop and heal all the time.

The resulting longer wars would require more units which has immediate consequences on war preparation, war unhappiness, stability and the AI willingness to make peace, probably to soon. And we can completely forget about all conquest-related UHV, as well as many other UHVs that benefit from the current quicker wars.

I THINK that my idea would not lead to that, because only few combat encounters (favorable odds AND extreme outcomes against these odds) are falling in that system. Yes, there will be a few more withdrawals and thus more survivors in battles, but not in each and every battle: I'd assume that at most 5% of all battles might be affected. It would be most noticeable with high-promoted units, and as far as I can determine, it would not require to program different behaviour into the AI.
This last point is the MOST important one in any modding decision. Because the AI is just dumb. Teaching it new tricks needs incredible effort; as well as deep insight in the entire code.
 
Last edited:
cmakk1012 said:
Well, what if it was just completely random? Not anything to do with whether they had >90% odds, just a 10% that any losing unit could survive at minimum health?
Click to expand...
My idea was that any unit, regardless of odds, would survive death.
 
You must log in or register to reply here.
Back
Top Bottom