RNG Value: Two

[Longasc]

The thing is that those 1% of the time results happen more than 1% of the time. Take the psychological effect in account, that people complain an remember better when BAD things happen to them. I still say it happens way too often and not rarely.

I am not satisfied with the way it is right now. Sure, there are more annoying things like the stopped final patch...

 

[Arathorn]

The thing is that those 1% of the time results happen more than 1% of the time.

No, they don't. People remember them more, but they happen truly randomly. People, in general, are very bad at understanding and predicting randomness. Ask any person to create a random string of 1's and 0's of about 40 binary digits in length and it will, almost without exception, be demonstrably non-random. Even those of us who work with extremely strong psuedorandom stuff as part of our job will fail that task fairly often.

If you eliminate selection bias and do a true test, the PRNG is nearly perfect. It generates 1% likelihood events 1% of the time and will have two come back-to-back exactly as theory indicates it should.

The problem is that you don't

Take the psychological effect in account

, because people simply do not truly understand randomness and streaks well. And don't realize that 1% is actually very likely. I often have thousands of combats in a single game -- something that is 1% likely should happen LOTS of times every game. I saw, today, someone complaining about back-to-back things that happen ~25% of the time, in a game where they're expecting thousands of combats. Umm...to NOT have that would be the problem, seeing it only indicates that things work like they should.

I still say it happens way too often and not rarely.

It happens exactly as often as it *should*. Once you start changing that and forcing shorter streaks or not-truly uniform random, exploits open up all over the place, and some of us would start exploiting them ruthlessly, whether we wanted to or not. Truly uniform random data is the only way that makes sense.

That said, I can see reasonable arguments for a more dramatic increase in stats over time with units. I will disagree, mostly, but that's definitely a point where I can see the validity to both sides. Changing the PRNG is just a very VERY bad idea.

Arathorn

 

[IceBeast]

I have to agree with Arathorn. To mess with Randomness is to only make it less random. I wouldnt like a game where I already know the outcome because it lacks randomness. Also, just to remind everyone, 1% of the time over 1000 battles would indicate that said events should occur roughly 10 times (assuming all the battles have this 1% possiblity of outcome). I have no doubt that if we conducted this battle (where the spearman has 1% probablitiy of survival) a million times, the spearman would have roughly 10,000 victories. In a similar tone, it is quite likely for a 1% occurance to occur 2 or 3% of the time on when the number of samples is small. And as Arathorn said, you're more like to remember the times it did occur vs the times it didnt. I agree that it can be frustrating, but messing with the random generation will only make things more frustrating. It would be better to just decrease the odds from 1% to less than this. (You should realize that 1% is by far very probable occurance. When it comes to something that can't must not occur, 1% is way too likely. No nuclear plant will be certified if it has a 1% chance of meltdown, and no medicine will be released if it has a 1% chance of causing death).

 

[Longasc]

I like more reasonable and predictable results. It must not be chess-like, but I could give tons of examples, especially galley-galley combat. Chances are 54.5% for a defender win.

Now I throw 4 galleys at one enemy galley -> the first dies, but redlines the enemy galley. Okay, three more -> all die.

This is bad luck, of course. But I dislike it - I would less dislike this if I would be the attacked galley of course, but these are the things that people hate.

We could calculate the odds for 4 galleys vs one. Assume all are regulars, 3 HP.

The chances are roughly 1% that the enemy galley survives an attack of all 4. It is reasonable to attack. The psychological problem is, all planning and all certainty of the result have proven wrong.

Now we can debate that this unwelcome outcome is realistic, or that is is near impossible but not impossible and so on...

... just wanted to say, I hate this randomness. I do not like playing cards either, but sometimes I see more systematics in playing Poker than in playing Civ3 battles.

 

[toh6wy]

I have an idea: maybe if two units are in combat, and one has a large technological advantage over the other, then the primitive unit is limited by the amount of damage it can do. It would have a minimum of being able to do at least 1 damage, and this might increase depending on just how far apart the two units are in technology, or the units' difference in experience levels. So, if you have an elite modern armor attacking a regular spearman, and the spearman manages to do its one damage, and the modern armor is still alive, then the spearman is automatically destroyed.

 

[Jon Shafer]

In principle they ARE far apart.
The rifleman is almost the soldier of the American Civil War, of the Prussian soldier of the Prussian-Austrian War (both in the 60ies of the 19th century), whilst the tank is - well, just the tank (the one of WW2).

Since that discussion is not new, I know about all the arguments for the rifle to kill the tank: "Imagine the tank battalion is sitting there with broken chains, but without any ammunition", "Imagine they have been caught in bad terrain, and without ammunition", "Imagine they [the tank personnel] are all sleeping and they don't have any ammunition" and so on....

The fact is, however, that in Civ all units are not sleeping, that there is no concept of running out of ammo and that all units have the same level of skills (except for there experience).
All this leads directly to the conclusion, that after a certain "distance" between both units (in regards of technology) there is no reason to assume that the superior unit should loose a given combat - except for the desperate hope of the player: "I didn't play well enough to be still competative - but I dislike the idea to be confronted with the consequences of my bad play".

Well, when it comes to the game they're only a few techs apart.

Maybe the tech tree needs to be tweaked then. As I said in my example, it's pretty easy to get Tanks before anyone else gets Infantry. Giving Tanks a huge advantage over Rifles wouldn't help make the game any more fun or balanced, IMO.