AI did something smart--accident or not?

Originally posted by ArielUK
You can also use the AI's 'superior knowledge' to your own advantage.

In a recent game as the Babylonians, I had just finished fighting a war against the Koreans to the north when the Mongols to the east decided to get aggressive and launch an invasion.

Luckily, the fools decided to invade with a force of medieval infantry and archers, through a patch of forested terrain. I had two cities just the other side of the forest, and by adjusting the size of the two garrisons I kept the mongol army heading first for one and then the other (marching through the forest as they went) as they tried to attack the one that their 'intelligence reports' said was the most lightly defended.

In the meantime, my horsemen, ancient cavalry (gotta love that statue of Zeus) and Babylonian Bowmen were able to launch attacks against the Mongol hordes from either city (once they'd arrived from the Korean front), or from the road that ran through the plains to the west, and whittle them down to the point where they gave up, sued for peace, and went home...

...which was nice.

:hmm: I never thought about doing that, interesting I think I might try it
 
Originally posted by jidosh


Hey, it don't bother me if other people want to cheat themselves. The AI has certain advantages to make up for what they lack in the strategic planning field. Since they can't plan for the future (unfortunately) they have to know everything instantaneously. IMHO. :cooool:

Yaknow, you're the one that questioned cheating in the first place back in post #5 on this thread, and now you're trying to say it don't bother you. So why'd you call foul in the first place?

I still think it boils down to a matter of personal preferences: some players don't mind reloading, others do. The reason I question whether 'cheat' is the right word is because it doesn't fit the definitions I'm aware of. I looked at a dictionary that had 11 definitions for 'cheat.' 10 of them refer to the use of deception or trickery to gain an advantage, but since you and the computer both know what you're doing no one's being deceived or tricked. The 11th definition comes close: it refers to cheating on a test or exam by knowing the answers ahead of time. Reloading is like that because you already know some of the answers, and any game of skill can be viewed as a test; we test our wits against the computer in this game all the time. It gets back to why you play the game, is it a game or is it a test? Well, it's both, I guess. But we play games for fun and we take tests to measure our abilities. Do we play for fun or to test? Even if you say 'test' it's a fun test and more game than test. So def. 11 isn't quite true either. Perhaps I'm splitting hairs here, but I just can't call it cheating. It's more of a learning and playing technique based on personal preferences. Maybe 'honor' would be a better word to use. Is it dishonorable to reload? Probably.
 
Being able to outsmart a computer is nothing to be congratulated for. That's like saying you can beat the chess board. It's really the game designers you're playing (and cheating) against. They give the AI these exploits because it can't compete with humans otherwise. It's meant to even the odds, not stack them in the AI's favor -- and even without resorting to reloading, the AI is very beatable.

I'm not saying you shouldn't reload if that's the way you want to play, but I stand before you today a Reformed Reloader. I stopped when I realized that, as long as I was reloading, there was no way I could possibly lose this game. And dang if that don't sound like cheating to me.

(caveat: I allow myself exceptions for missed clicks, typos, and sleep-induced insanity. But no longer for simple stupidity or lack of precognition.)
 
thing is they SHOULDN'T give the AI all those cheats they should just simply IMPROVE the AI.

to simply give them unfair advantages instead of useful tactics is not a good thing at all

if they were to implament a program that allowed the AI to learn from the human player than that would really provide a real FAIR challenge unlike the huge exploits and unfair advantage that they get now because the developers can't provide a "smart" AI

personally I don't reload I perfer to be shocked and enraged :D by the AI sneak attacks

oh how many times have I crippled an AI civ out of rage and frustration and leaving them alive with one city that I surround with units and starved to size 1 because they RoP raped me or because they sneak attacked me :lol:
 
if they were to implament a program that allowed the AI to learn from the human player than that would really provide a real FAIR challenge unlike the huge exploits and unfair advantage that they get now because the developers can't provide a "smart" AI

Even if they could, I don't think you'd want them to, unless you were willing to pay far more than $50 for such an advanced product!

Remember, Deep Blue only stands a chance against Kasparov because it has an extensive database of every tournament game in recorded history to compare the current game position to... the chess board is much simpler than the civ map by several orders of magnitude, and it can always see all of its opponent's pieces! The civ AI is on your comparatively puny box with more pieces to deal with and no database of historical games! I think you need to recalibrate your expectations! :)
 
Originally posted by SiggyLlama


Even if they could, I don't think you'd want them to, unless you were willing to pay far more than $50 for such an advanced product!

Remember, Deep Blue only stands a chance against Kasparov because it has an extensive database of every tournament game in recorded history to compare the current game position to... the chess board is much simpler than the civ map by several orders of magnitude, and it can always see all of its opponent's pieces! The civ AI is on your comparatively puny box with more pieces to deal with and no database of historical games! I think you need to recalibrate your expectations! :)

WHAT!!! NO NEVER :D

I will ALWAYS have the HIGHEST expections for CIV's AI

I will NEVER be satisfied until I have the SMARTEST possible AI that can ever be created :D

:hmm: no "real" civfanatic should EVER lower there expectations you must not be a "real" civfanatic :D
 
I am a civfanatic, yet I'm also a pragmatist. ;) (Not to mention a computer science student, so I'm inclined to spec out inherent limitations.) I'm a civfanatic because, in spite of the AI's unfortunate limitations, the game is still so darn fun!
 
yes it is siggy, yes it is :D

that why we're all here :)
 
Originally posted by Revolutionary
... allowed the AI to learn from the human player...

That's something programmers have been trying to achieve for decades.

BTW, if anyone's interested in reading more threads on the 'reload/cheat' topic here are some I found in the general discussion forum:

"Do you reload if you forgot to do something" - May 2, 2003.

"Does the computer cheat?" - June 15, 2003.

"Does the AI cheat on all levels?" - August 17, 2003.

"Is this cheating?" - December 3, 2003.

(Can anyone tell me how to create quick-links to these?)

Opinions vary considerably on this topic. One trend I've noticed is that people tend to reload less and less the more they play and improve. Personally, I find reloading an excellent way to learn and improve my play. I'm not yet a "reformed reloader" like some, but I'm getting better.
 
i agree with siggy there. we shouldnt expect too much from AI that it can outsmart us using human traits. if the computer thought about each move as much as we do, each turn would last a very long time and probably melt the CPU :eek:
anyway, what people want improved in civ3 will either go into a patch or civ4, so as long as firaxis know what we want, we WILL get the perfect game, sooner or later :D
 
One trend I've noticed is that people tend to reload less and less the more they play and improve. Personally, I find reloading an excellent way to learn and improve my play.

Agree completely. As you get more comfortable with the game what you begin to realize is that every adverse result or 'ooops" in the game won't in the long run matter. I use reloads to experiment with the game--"what would happen if...?" I'll play for a few turns and then may go back and see what happens if I didn't do that. It helps show me what choices are actually important and helps me better understand how the game 'thinks' -- which isn't much like me or any human opponent.

Frankly in SP its just a matter of what makes the game fun for the player. The only real danger to making game mods and the reloading thing is the chance that the player gets dependent on them. As I start I normally have some kind of personal goal in mind and sometimes I'll set myself an 'ironman' restriction -- no saves except at quit. In that case reloading would be cheating because it breaks the rules I'd set for myself.
 
I try to only re-load if a change I want to make is either relatively trivial, or absolutely vital.

For instance, I'm moving units and don't spot a river-crossing, so press the direction key twice, but it's the next unit that moves off somewhere completely unsuitable (into enemy territory, for instance). Time to re-load and re-do, seeing as it was an accident rather than an error of judgement, but could have a major impact on the rest of the game.

Or, at the other end of the scale, in a recent game I was beaten to a wonder by the AI, and was still researching the next wonder-producing tech, so I couldn't switch to another wonder.

Faced with the prospect of losing 300+ shields, I went back as far as my auto-saves would allow (admittedly only 5 turns or so), and tried everything I could to finish the tech on time to beat the deadline - switched to Democracy, ran a massive tax deficit, put as many of my citizens into 'scientist mode' as I could - all to no avail. So in the end, I had to suck it up, switch production to the largest building available, and take the shields hit.

But re-loading allowed me to try out a few options and see what was actually possible, so I learned something in the process.
 
Back
Top Bottom