Combat Cheats by the AI

[Rhino Ryan]

Originally posted by Strider

If we judged everyone's experience on there post count then we would have a VERY messed up site here. DaveMcW is one of the best Civ3 players I know, yet has a post count of 500 or less.

I understand that and all I say is that people with 1000's of Post counts are more experienced and know this question has popped up 100's of times (as they said in this thread). I am not saying that they are better but I am saying that they are knowledgable and have probably seen DaveMcW in other threads, spoken to him, and learned from him.

Also I appologize for my earlier post to Arretium. It was a little nasty and I am terribly sorry and i hope all can forgive me

"it is okay"

 

[carlosMM]

posted by Bamspeedy:
Build a larger stack to prepare for those bad streaks


what I meant (and that is rather obvious i think from my post): you need to play counterintuitive, with a LOT more unit than one would initially think. Thus, the way it works favors the AIs over an average human.


play a while and you learn to deal with it



regentMan: sorry, was a bit harsh there, I jsut hate it when people do not want to think on how I meant what I said and post ahead......

 

[Moulton]

In my current game I was losing legionaires to swordsmen and wariors. I thought, 'too many'. So I started noting the resuils of every battle.. vet status, defense bonus, etc. I guess the game knew I was taking notes, because it started letting me win some.
Actually what happens to me, being fro military background, and trained in properly planning war to minimize troop loss, :: I hate it when I lose a single troop unit. In reality, you cannot expect to not lose some, but I still hate it.
My other favorite game was Jagged Alliance. My win goal in that game was victory without a single mercenary lost--if I lost one person, I did not consider it a good win. The game playse totally different, and was on a tactical level... but the same desires exist. Every loss hurts.

 

[anarres]

Where is the thread starter with their list of 600 battles?

Maybe the test didn't prove their point after all...

 

[Eklektikos]

Facts are irrelevant. Statistical analysis is irrelevant. The AI is a lousy cheatin' scumbag, it's self-evident!

 

[Catt]

Originally posted by Hygro


I would LOVE to see the results!

Do you mean regarding radar tower tests? If so, I think the thread was titled "Questions on Radar Towers" and morphed into "Questions and Answers on Radar Towers." Probably posted late last year or in the first few months this year. They were pretty basic questions I wanted answered such as actual radius of coverage (i.e., effective when attacking from Tower+2 tiles into a tile at Tower+3?), whether they affected bombard, naval, and air units, terrain improvements, etc.

The search function is disabled, otherwise I would provide a link.

 

[RegentMan]

Originally posted by carlosMM
regentMan: sorry, was a bit harsh there, I jsut hate it when people do not want to think on how I meant what I said and post ahead......

It's ok. I just can't stand it when someone, faced with overwhelming evidence, still argues against it. I'm sorry for my last post, if you got to read before Padma edited it. Sorry about that post too, however I was mad about the previous post. Not my usual self, but hopefully we can just move on.


The AI doesn't cheat...

 

[anarres]

RegentMan, I guess you two will move on anyway, but:

- carlos said the combat system favours the AI

- you (and most others) interpreted this as meaning the RNG gives different results for the AI and human, i.e. the AI cheats.

I know both of you know how to play the game, I guess you *both* took the other as not knowing what the other was on about, and you were both right in a limited sense.

As an aside, calculate the chance of a flip for a newly captured city. You often get values around 5% for a flip. That is very unlucky if it flips in one or two turns, but even when quelled the chances can still 'steady out' at 70 or 80% of the initial flip chance (4% say). Losing one of those cities is really unlucky in any given turn, but when you see that you've kept it for 50 turns and not had a flip at 4% per turn, you are one lucky SoB....

 

[eqphreak]

Oh man Eklektikos nice sig! Mr Smith goes to Washington LMAO

Moderator Action: Yes, Eklektikos has a nice sig and an intereting avatar. And this post furthers the discussion about combat cheats how? --Padma
Please read the forum rules: http://forums.civfanatics.com/showthread.php?t=422889

 

[carlosMM]

RegentMan:




anarres sums it up well: I do think the RNG is random, but I also think it is not the best way an RNG could be done for Civ....

I would actually prefer a gaussian spread of results.

 

[Stapel]

Originally posted by arretium
Stapel Wrote:

"No offense, but your story is quite laughable. Check bamspeedy's link and judge again. Some members on this forum are completely lunatic and have analyzed each and every possibility of AI cheating."

Well No Offense Stapel, but your arguement is laughable. You critique my argument but you fail to point out its flaws, you provide zero evidence to support your point, and simply engage in an ad homenem attack. Perhaps being a civ addict has reduced your ability to engage in conversation with others?

Where's YOUR evidence?

What a dumb post! I did post evidence! Can't you read? I advised you to check the link provided by Bamspeedy and asked you to judge again! I think that save will convince you.

If that is not enought evidence, well, let me know.

 

[anarres]

I would prefer a normalised spread of RNG results too, although the total unpredictability of it makes it exciting.

 

[Mark Young]

Originally posted by carlosMM
RegentMan:


I would actually prefer a gaussian spread of results.

Could you explain what this is?

 

[Frisian Warrior]

Avoiding danger is no safer in the long run than outright
exposure. The fearful are caught as often as the bold.
--Helen Keller
If you don't know who she is, try your favourite search engine. Thought that might add to the discussion.

The human perception and biased memory of a battle favours the AI. No question about that.

 

[nalves]

I think that AI does not sheet. But it should a little it would be more of a chalenge .

 

[carlosMM]

Originally posted by Sir Eric


Could you explain what this is?

sure!

Gaussian means low and high ends (i.e. in this case very improbable results) are rarer, while middle ground results are more common. Gaussian distributions are also called bell-shaped: most in the middle, little on the edges....

 

[anarres]

I would prefer a normalised spread of RNG results too, although the total unpredictability of it makes it exciting.

A 'normalised spread' also refers to this distribution of numbers:

 

[anarres]

So, I found a better graph:

Ignore the axis of that graph, but you can see that you get many more results (greater height) around the modal ('average' in this sense) value.

 

[nalves]

can't be more clear

thanks Anarres

 

[carlosMM]

tx, anarres