How the RNG can completely destroy a game

[The Last Conformist]

My impression is that both the RNG and the probability function are borked. About the RNG, we all know that it can produce ridicolous events (such as the spear beating the tank). The flip function seems to work well with normal values, but goes astray when something tends to zero (perhaps... ). A possible result is that 6 population units turn into 6 harry potters and vanish 20 military units...

Would people please stop blaming the pRNG without stopping to think whether a bad pRNG could create the problem they perceive?

The combat system is so constructed that the Spearman-beats-Tank scenario would occur with very close to the observed frequency if the pRNG was perfect. This ought to suggest to you that the blame for any problems with combat outcomes you perceive rests with the combat system, not the pRNG.

As regards culture flips, the chief problem is stated easily enough; the reduction of the flip chance for each military unit situated in the city is too small. This is no fault of the pRNG either.

 

[Willem]

So what's the point of developing it culturally? Lots of shield that can be spent on military would be wasted.

The point is that part of the equation for culture flipping involves the number of squares that are covered by another civ's culture. If you had made an effort to develop the culture in that city, you probably could have pushed back the other civ's boundaries and ended up having at least some of the squares of the city you were trying to capture in your borders. As it stood though, it was almost completely surrounded by enemy culture.

AI's have always large armies, and the persians are tossing at me all their military might. Walls=50% defense bonus under size 7. Musketmen defense=6. Rifleman defense=9. Do I really need to expain you what this mean?

The issue in this thread is culture flipping. You captured a city that you knew had a good potential for it, so that should be your priority. Once you get some culture of your own happening , then you should think about defence.

First, when there are resistors you cannot rushbuild. Second, you're sadly mistaken.

If you were to starve the city down to one, you not only get rid of the resistors, but you also mininize the risk of flipping. According to the screen shot, you made no attempt to do so; the city is still growing in population, not shrinking.

The first thing to be built is usually a harbor (to have access to luxuries, have some people happy and shorten resistance periods).

Again, starve the city down and happiness is not an issue. At least not until it starts growing again. With your nationals, not the AIs, who would give you a much better chance of holding on to it.

Temples? Yes, but only if i need to grab some land and/or connect cities.

And also to help prevent flipping. If you have no culture of your own, sure as hell it's going to flip. Build a Temple and in 4 turns you have some culture reducing the odds tremendously.

Doesn't sound obvious to you that a number must be rolled out from a random generator function? OOPS... this is exactly the RNG

There's a lot of variables that determine flipping long before the RNG kicks in. It's really a minor element in the equation, the final link in the chain so to speak. You can't ignore the other factors then blame it solely on the RNG.

 

[microbe]

I'm fully aware of everything you say. The goal of my post is to expose what i consider to be a serious flaw in the game. I was not asking for help.

No you are not FULLY aware, otherwise you would not have ranted about it. At least, not seriously consider it a "bug".

Or were you ranting just for the sake of ranting? I could have ranted too when I lost an army in the battle, probably with better language.

 

[Grille]

No you are not FULLY aware, otherwise you would not have ranted about it. At least, not seriously consider it a "bug".

Or were you ranting just for the sake of ranting? I could have ranted too when I lost an army in the battle, probably with better language.

Heh.
Me thinks that, if it was possible, I'd rant about an unseccussful round of Russian Roulette, too...

 

[tR1cKy]

Someone told me to calm down a little... ok.
A few points need to be discussed further:

The issue in this thread is culture flipping. You captured a city that you knew had a good potential for it, so that should be your priority. Once you get some culture of your own happening, then you should think about defence.

It takes way too much to build sufficient culture on a newly conquered town, even if you have thousands of quids to spend on rushbuild. And while you're busy cranking up temples and libraries, the AI knocks at your door and catch you with pants down.

If you were to starve the city down to one, you not only get rid of the resistors, but you also mininize the risk of flipping. According to the screen shot, you made no attempt to do so; the city is still growing in population, not shrinking.

Absolutely right, but i had no time to do anything.
First turn: all resistors. I activated the governor with "manage citizen moods" set to "yes". This is necessary to avoid civil disorder, which raise the flip chance.
Second turn (screenshot): resistance ended, governor turned off. Don't remember if i put the town to starve or not, but...
Third turn: flip.

No you are not FULLY aware, otherwise you would not have ranted about it. At least, not seriously consider it a "bug".

Amazing news: only clueless people rant. Fully aware people "complain", perhaps. And you call "bug" a bug only if you're not aware. Otherwise, you would have called it a "microbe"
Yeah i ranted also for the sake of ranting. I like it. You should have ranted too when you lost your army. It's great for dispelling the stress. And if you can do with a better language than mine i'm glad for you.

 

[IbnSina]

[rant]
I would like to go out on a limb here and indicate sympathy for tr1cky. I hate the RNG's guts. I don't hate it because it is random - we need and want that - I hate it because it has all the streakiness that uniform random variables are known for. I think the RNG should work on a normal distribution, or some other distribution with more central tendencies. Otherwise, again and again we find absurd streakiness leading to the destruction of our SOD by an archer and a spear, or cities with alleged flip probability 0.14% flipping and taking our stack of 21 knights with them (as happened to me a couple of days ago). I know this happens (after a fashion) in so-called RL; as someone pointed out, Iraq might flip into one or more Islamic republics. However, it will not take 150000 occupying troops with it when it does.

Some people might enjoy that, but I am aware of the rules and take all the precautions that others have discussed, yet I still get burned. When I do, it spoils the game for me. If there is one thing Firaxis should fix, it is the distribution from which the RNG draws. The normal distribution has a lot to do with the real world. It has a overwhelming quantity of theoretical and experiential verification behind it - the real world demonstrably follows normal distributions. Why can't the game mimic RL and use a NORMAL distribution?!
[/rant]

If provoked, I can arrange for a matching rave later...

 

[Neomega]

I hate culture flipping. I love the RNG.

Don't blame the RNG.

I like culture flipping, I just think it is rediculous you lose your units.

More fun would be an uprising/rebellion/resistance/insurgency (depending on era), where you would see naturally occuring units sabotage and harass your units.

 

[Woody]

IMHO you biggest mistake was taking one of the cities closest to the capitol as your invasion's base of operations. Not only do you maximize the chance of flipping due to distance from the capitol it is very likely that the city was one of the first built so it had a large amount of culture working against you. If you had to land there you should have brought a settle with you and razed the city and built your own city.

I would have invaded at some distant point from the capitol even though my resupply lines might be longer. I would take a city much farther from the capitol that would have probably been samller and not as well defended and be less influnced by culture from the heart of the enemies empire. Also this would maybe not let the enemy counter attack with as many troops at once since they would be staggered in the time its takes to reach a more distant part of his empire. Its much easier to defend over multiple turns than the all at once counter attack you get when you land near the core of the enemys empire.

The only other option I might have tried is just taking the capitol itself instead of the city next to it.

 

[bkwrm79]

It is stupid that military force plays so little role. With a certain amount of military force (say 2 units per pop point) the city should be secure. And units should have a chance of escaping the city if it flips (obviously, another RNG role).

 

[bingen]

Ibn, the pRNG has been tested and shown to be fair. Streaks occur in random sequences more often than people expect.

 

[gunnerxtr]

That is a big time bummer. A very costly flip IMO with 2 armies and 1 leader in there..... i feel your pain.

 

[morchuflex]

I completely agree with IbnSina. My opinion exactly. Even though I'm aware that streaks can happen in real life - but that's not the point here.
And I support Tr1cky, even though he was sometimes a bit rough...

A good way to make the culture flip more plausible would be the sudden uprising of insurgent units (much like civ2's guerillas) that would attack your troops, and the city would flip if your troops were to be defeated, instead of simply converting in one turn.
These insurgents should be weaker than regulars but show up in numbers depending on the city population.
And once again, when the military forces outnumber the population, there should be no chance of flipping WSE.
Don't get me wrong: I love the culture flip concept. But its implementation simply sucks.

 

[WackenOpenAir]

The fact that battles are dicided not by 1 rng number but by at least 3-5 depending on regular-elite, causes the battles te be pretty distributed much like a normal distribution.

For city flips goes the same, if you have conquered 50 cities, you could make a normal distribution on the amount of cities that will flip vs the amount of culture and thus chance you have for flips. It will be pretty correct.

Normal distribution is the result of many random chances.
Civ uses many random chances.
So a normal distrubition there is in civ rng already.

If your game depends on 1 random chance you are doing things wrong.

Still, i do agree that units should be more able to prevent flips.

 

[The Last Conformist]

A flat pRNG is simple, elegant and good at creating normal-distributed results were we want them (eg, combat). A normal-distributed is not.

It would be much simpler to design a better culture flip system than redesign all stochastic elements of the game around a new pRNG.

 

[Longasc]

I agree with TLC.

The whole culture flip system SUCKS BADLY.

Things like the incident mentioned by tr1cky should not happen. Regardless what we blame for it, RNG or the culture flip system, it sucks.

As obviously the flip-system is flawed, and this time not the often blamed RNG, we must live with it.

As there will probably not be anymore patches for Civ3, I suggest the following workarounds: Depending on playstyle, you can probably live with "culture flips off". I also suggest turning off cultural linked starting positions, to avoid always having american civ neighbours in any game.

We found a good solution to the barbarian bug, and by completely eliminating culture flipping we can deal with this problem, too.

Still found no good workaround for the sub bug.

 

[Doc Tsiolkovski]

I fully understand anyone getting upset when a far-flung AI city with no substantial culture you used as a healing point during invasion flips back and takes your army with it.
I fully understand anyone getting upset when a reg Spear in the open defeats a healthy Tank.

But tRicKy, what you're complaining about is something completely different. Period. It's more like a Vet Rifle in a city defeating a Tank - something with a probability of about 10-20%, in other words, it'll happen every couple of turns.

And, why are you refusing to unerstand Willem's points about culture? He's not talking about fighting Persia's overall culture, or exceeding their resident culture. It's about bringing down the odds significantly. If a second city would have pushed back Persepolis' cultural overlap, flip risk in that captured city would have gone down. A rushed (read: Disband units) cultural building would have greatly influenced the resident culture ratio. Joining native Workers, or jumping your Palace closeby as well.
When you don't use those, it is simply a wrong move to station your units inside that city. That's not a flawed RNG, that's a flawed tactic.
On Deity, never capture an AI core city, unless it holds a wonder, or that Civ is about to be extinct very soon.

 

[WackenOpenAir]

On Deity, never capture an AI core city, unless it holds a wonder, or that Civ is about to be extinct very soon.

I always capture. Also on Sid btw. I just don't place units in the city after the first turn. If the first turn, 2 resisters are quelled, the city can then be starved down with no units in it. If it flips, i recapture. Recapturing also reduces the size of the city. When all cities are size 1, chances of them flipping are small enough to live with. Important wonder holding citizens can then be reduced to a 0 chance with garrisons. for others i will just keep like 1 unit around per 3 cities to recapture. until the civ is eliminated.

 

[Doc Tsiolkovski]

That, certainly, is also a valid strategy.
I only referred to the case you want a city in that location at once -then raze and replace.
Btw, some players (Lkendter, for example) believe for the flip risk calculation it matters if there are foreign citizens in a city. So, a city that was owned by Civ A formerly has a higher risk to flip to Civ B than a genuine city of yours would have.

 

[morchuflex]

BTW, does capturing or razing the enemy capital affect the chances of culture flips for other captured cities from the same civ?
If this is the case, it would be worthwhile to aim directly towards the capital, take or destroy it, and only after that capture the other cities.

 

[WackenOpenAir]

Not directly in itself.
It will however affect the distance from the city to the capital.
It will slow down the other civ's cultural growth, helping you in the future.