The random generator HAS to be broken

[DrFell]

Anyone here played Colonization? I seem to remember it had similar problems with it's combat system - you attacked a city, won once then unit after unit would win against the defenders, or unit after unit would lose if you lost the first combat. Although I don't yet have civ3 I'd like to know how much of a problem it is - because it got to the point that in Colonization I would go and move some other units if my attacker lost the first battle in an attack- because this would 'reset' the random number generator. Never had any problem like this in civ2...

[RedWhiteArcher]

2DrFell
Yes I remember it. Sometimes it was ridiculus. I managed to loose a veteran dragon to a colonist.
Yes the random combat is a bit suspicious but not ridiculus, you just notice it too many times.
And fine work on having the result beign done before the combat, denies you save reload cheat.

[Excelsior]

Originally posted by DrFell
Anyone here played Colonization? I seem to remember it had similar problems with it's combat system - you attacked a city, won once then unit after unit would win against the defenders, or unit after unit would lose if you lost the first combat. Although I don't yet have civ3 I'd like to know how much of a problem it is - because it got to the point that in Colonization I would go and move some other units if my attacker lost the first battle in an attack- because this would 'reset' the random number generator. Never had any problem like this in civ2...

I played colonization, but I don't remember that particular problem. Then again, I was only 8 or 9 when I played it...

[Mister Hand]

Anyway, my last two games I've noticed something funny: in the same turn that a resource would run out, another one would appear.

This by design.

As for the random number generator, I agree that it's a little suspect. Of course, there is no such thing as "random" in computer gaming. It's just a long list of numbers that the computer grabs from. I think the seeds (basically, a pointer on the list where the computer will start grabbing from) might be poorly conceived. Another problem is that the numbers in Civ3 are so small that the exact same results many times in a row is more common.

[Venger]

Originally posted by LaRusso
can we award venger a 'whiner of the year' award, please? he produces 2,37 whining threads a day

You know, you never bring ANYTHING to a thread, besides your sissy remarks. You're just a pu$$y. Try posting a constructive topical argument sometime, I don't recall ever seeing one from you. Actually, just ignore posts from me anymore, I'd rather save everyone the hassle of reading your senseless drivel.

Venger

[Venger]

Good replies guys.

Do you guys think that maybe there is some sort of built in "tide of battle" modifier or somesuch that basically modifies the number based on subsequent victories, so that one Civ doesn't win 10 battles in a row? Or some other wierd modifier that doesn't make itself apparent? That's the only way I can make sense of what happens during battle...

Venger

[Evil_Eric_4]

Anybody here ever play Europa Universalis?

talk about a broken random roll modifier

I havent noticed as big a problem in civ3 but Ill take your word for it since I know it can happen..

[Kenjura]

Originally posted by Excelsior

I am not convinced that the random number generator is broken, but I admit it does seem suspicious. More often than not, it seems, units get on streaks. One unit will get 6 hits, then the other will get 4, alternating back and forth. The streaks, though, seem to last longer the greater the unit's combat advantage. Thus, greater hitpoints reduce the effects of the streaks, making combat more balanced.

You're definately on to something there. That does have a real-world parallel...when a unit is losing, they tend to keep losing. A single unit can't lose much before it is destroyed.

The existence of streaks makes a "fast" unit that much better. The unit will withdraw when getting whomped, and then can retreat to a city (or safe place with battlefield medicine) and heal fairly quickly. I actually used this strategy in a tough war....a constant stream of attackers going back and forth, gradually weakening the enemy without ever actually losing a unit.

Streaks are probably calculated beforehand based on some combined probability. As annoying as they may be, they represent reality, and frankly, depending on the difficulty (I suspect), they benefit you as much as the AI.


Instead of simply increasing hit points across the board, which flattens the curve, perhaps in-game strategies would help. Always take advantage of terrain. Use fast units. Perhaps (as much as you may not want to) use armies to actually form armies. Though that may not work anyway...with some of the ridiculous streaks. But wouldn't you like it if *your* warrior fended off an enemy army of cavalry? ^_^

[Xerxes314]

You guys just whine and whine and whine with your anecdotal evidence until your typing fingers fall off.

WHERE IS THE EVIDENCE?

Statistics will resolve this issue once and for all. Why don't any of you have the balls to get some statistics to back up your stupid ideas?

Xerxes

[Zanzin]

Originally posted by Xerxes314
You guys just whine and whine and whine with your anecdotal evidence until your typing fingers fall off.

WHERE IS THE EVIDENCE?

Statistics will resolve this issue once and for all. Why don't any of you have the balls to get some statistics to back up your stupid ideas?

Xerxes

Maybe you're blind, maybe you didn't read the thread, but a fair bit of evidence from field testing has been posted above. Common results are starting to become conclusive as to explaining exactly how the combat in civ3 works.

Thankyou to everyone who has posted in this thread to shed some light on the "different" combat system we're seeing in civ3. I won't say "weird" or "stupid" combat system, because I'm begining to think it was thought out a lot better than the combat system in civ2!

[GodSpawn]

The sample size of the tests that have been presented so far is much too small to attach any statistical significance to the results.

We need a LOT more tests before we can draw any definite conclusions. And we'd need to do a proper mathematical analysis before it means anything.

(Sorry, I haven't got the time to do that myself...)

[SuiteSisterMary]

WHERE IS THE EVIDENCE?

Check my thread about an army with two vet longbowmen and a elite one dying in an attack to a single conscript riflemen. It's even got a savegame.

[GodSpawn]

A bigger sample size needed, one that includes all the times that nothing unusual happened. It is meaningless to extract a bunch of unusual results, and then apply a statistical analysis to just those results!

An analogy may help. Imagine if, out of a group of 1,000 people, you listed the heights of only those people who were taller than six feet. And then imagine that you said "Hey look at this - people are a lot taller than they should be!". That would be pretty stupid, wouldn't it?

[barefootbadass]

Originally posted by Zanzin


Maybe you're blind, maybe you didn't read the thread, but a fair bit of evidence from field testing has been posted above. Common results are starting to become conclusive as to explaining exactly how the combat in civ3 works.

These results are FAR rarer than people make them out to be. People just notice them more because they piss them off more, or because if it benefits them its just odd looking. I have only had ONE instance where an very unlikely streak occured. And it IS only anecdotal evidence. Record the results of a several thousand battles(actually probably several million) under the same conditions and with the same types of units and maybe then you will have some credibility with people that approaching this issue with more than 'common sense' probability.

The "fair bit of evidence" posted is only anecdotal, and just a drop in the bucket of battles and games people are having. To be "fair" these accounts cannot be considered a good cross section of people's experience having streaks, as it is biased by having people who have been lucky/unlucky enough to have these freakish results occur. And, like Godspawn said, you need to have a much larger sample size as well.

[SuiteSisterMary]

An analogy may help. Imagine if, out of a group of 1,000 people, you listed the heights of only those people who were taller than six feet. And then imagine that you said "Hey look at this - people are a lot taller than they should be!". That would be pretty stupid, wouldn't it?

Not if the definition of 'human height' was 'will be six feet tall 0.00014 percent of the time.' At that point, having even two to point out in a thousand would be, as they say, statistically significant.

99.7 (iirc) within three standard deviations and all that.