Someone is cheating here?

[TheNiceOne]

Originally posted by Greyhawk1
OK, the thing is. mo matter how many times you tell a civ to remove a unit from your waters the civ says "ok" then totally ignores you.

But this is how the function work, and it works similarily for AI and human players: You can ask the first turn, but when you get to say "leave or declare war" is not a result of the number of times you ask. It is a result of how many units you have inside enemy borders, and how close they are to an enemy city.

You stop just once in their waters and they leap on you like hounds.

Just like you can if you want to. They don't get to throw out your ship for several turns unless you stop it close to an enemy city though.

If I had to go to war every time a civ completely ignores my water borders then I would be at war from the time Galleys appear to the very end of the game.

the same cannot be said for land units. Occasionally they try to send a settler or a worker through but 9 times out of 10 they dont cross your borders.

This is why I think its a bug.

Try setting up a scenario where an enemy ship is trapped inside your borders and your ship is trapped inside theirs. I'm quite sure you'll find that you get to throw out the enemy ship on exactly the same turn as your gets thrown out by the AI.

Before I did such tests with land units, almost everyone (including myself) believed that the AI cheated there, but the tests made clear that the AI and the human player followed the same rule.

 

[Galcador]

Originally posted by Greyhawk1
OK, the thing is. mo matter how many times you tell a civ to remove a unit from your waters the civ says "ok" then totally ignores you.

You stop just once in their waters and they leap on you like hounds.

If I had to go to war every time a civ completely ignores my water borders then I would be at war from the time Galleys appear to the very end of the game.

the same cannot be said for land units. Occasionally they try to send a settler or a worker through but 9 times out of 10 they dont cross your borders.

This is why I think its a bug.

The rules governing ejection from territory (land or sea) are the same for the human as for the AI. The length of time a unit/force is allowed in enemy territory before being ejected depends on the size and strength of the force and proximity to cities, I think. My empty galleys are able to spend many turns in enemy territory - sure, they ask you to get out when you first wander in, but that's no problem, just ignore it. However, in a recent game I wandered into Korean waters with 15 galleys full of knights, and I wasn't allowed to stay for even 1 turn. If they'd done the same to me I would also have been able to eject them immediately.

Rules apply equally. No bug/cheat here.

 

[Tatran]

Well since there's no proof of the AI cheating
the feeling can't be taken away.

This happened in my current game.
In 1545 Persia had 56 turns to go
to build the Universal Suffrage.
In 1550 (one turn later) Antioch
had the US as a city improvement.
Ok they used a Great Leader,but
then it should be 2 turns to have
this wonder.In 1550 the production
box should be full (1 turn to finish)
and in 1555 the production box
should be empty because the
wonder has been built.

 

[TheNiceOne]

Originally posted by Tatran
Well since there's no proof of the AI cheating
the feeling can't be taken away.

Tatran: Just upload the save games for 1545 and 1500 and I'll look into it.

 

[Tatran]

I uploaded the files.I played Japan.
http://www.civfanatics.net/uploads5/Tokugawa....sav.
(Sorry I'm having some trouble with upload pop-up
screen).Maybe someone can tell me why the autosaves
are 10 times bigger?

This reminds of the old civ2 game where something
familiar happened to me a couple of times.
One of the largest civs researched a tech including
a wonder and at the end of the AI's turn the same
civ got that wonder.

 

[anarres]

You can compress the autosaves with a zip program. They are bigger because they are not compressed, unlike normal saves.

 

[Moonsinger]

Originally posted by Tatran
This happened in my current game.
In 1545 Persia had 56 turns to go
to build the Universal Suffrage.
In 1550 (one turn later) Antioch
had the US as a city improvement.
Ok they used a Great Leader,but
then it should be 2 turns to have
this wonder.In 1550 the production
box should be full (1 turn to finish)
and in 1555 the production box
should be empty because the
wonder has been built.

It's very simple, my dear. Please remember that human player always go first (before any of the AIs). For example, you may end your 1545 AD turn, but because you are going first, it was actually just the beginning of the AIs turns. Therefore, in 1545 AD, the production box should be full and the Universal Suffrage will be finished in 1550 AD and the production box should be empty in 1550 AD, not in 1555 AD.

Basically, in a single player game, the human player would always go first. Therefore, if you investigate on the AI city, please keep in mind that you are looking at the timeline of your previous turn. I hope that would answer your question.

 

[Tatran]

Very clear explanation Moonsinger.
So let's say when the production box
is full (one turn to go) as for human player
as for AIs in 1545,at the beginning of the
next turn,everyone has an empty production
box or better everyone has completed his/her
production.After that first human player can make
his moves in this turn (1550) and later the AIs.

 

[Moonsinger]

Tatran,

Here is a test situtation that you may want to check out just to be sure. Basically, let's set up a test game on a tiny map where the human player starting postion is right next to the AIs (the only way we can meet the AIs at 4000 BC). Of course, everyone has Writing as the starting tech and 500 gold to spend.

Since the human player go first, we will immediately found our first city at the starting location. Because we already have Writing as the starting tech, we should be able to build the embassy right away, but we can't build any embassy at 4000 BC becuase the AIs haven't moved yet.

By the next turn (3950 BC), assuming the AIs also found their first cities at 4000 BC too, we should be able to build embassies and investigate their cities. At this time, our production box won't be empty, but the AI production box will be empty. Why? Because we go first and when we investigate their cities in 3950 BC, we are looking at the end of the AI turns which is at the end of their 4000 BC. Basically, it is 3950 BC for us but not for the AIs.

Please note that I'm speculating at this point. When I have time, I will do this test just to be sure. If anyone have done this test, please post your findings. Thanks!

 

[Arathorn]

I don't think that's right, Moonsinger. As I understand it, the flow of the game is...

MOVEMENT
PRODUCTION (with many sub-phases)

with the human first in each main phase. So, what that means is...

4000 BC -- You move and found a city. End MOVEMENT for human.

AI moves and founds a city. End MOVEMENT for AI.

(I think, technically, 4000 BC ends here and it's 3950 BC now, but that's not how I think of it....)

Human city produces X food, commerce, shields. End PRODUCTION for human.

AI city produces X food, commerce, shields. End PRODUCTION for AI.

Human MOVEMENT again...you found the embassy and investigate the city...and the AI city will have production in it, because it's had its production time.

I feel reasonably certain of this turn order.

Arathorn

 

[aeldrik]

I think it is Human mov., AI Production, AI Mov., Human Prod., and then all over again...

 

[Moonsinger]

Arathorn,

Let's assuming that you are right (again, we need to do the test just to be sure, but for now, let's assume that you are right) and let's say that we investigate the AI city at 3950 BC and find that its production box isn't empty. Since the AIs haven't yet started its turn, we have no way of knowing that it won't fill up the production box during its movement turn (too bad no pop rush or cash rush at 4000 BC; may be we should set it up so that everyone are in the Middle Age and in Republic at 4000 BC). Therefore, in Tatran problem, the AI could have used the leader to rush the Universal Suffrage during its movement turn in 1545 AD which we don't know that until we start our turn in 1550 AD. In either case, we have successfully explained Tatran situation. The rest, we need to do the test that I just mention to find out for sure. Sorry, I'm still at work and can't start up Civ; if you guys have time, please do the test. Thanks!

PS: I haven't yet seen the AIs using the great leader for rushing Wonders before, but in theory, they could.

 

[betazed]

Originally posted by Moonsinger
I haven't yet seen the AIs using the great leader for rushing Wonders before, but in theory, they could.

I am almost certain I have seen this. Here is a scenario that I saw. I sell ToG to a civ. They are at war, there are no other wonder builds going on, so no cascade possible, and next turn they builds Newton's.

AFAIK, we never saw this is civ 3. However, with PTW this behavior occurs too regularly and prompts me to think that leader rushing has been added to the PTW AI (I could be wrong about this).

EDIT: I also think the AI never stores a leader. If there is a wonder that it can rush then it might. Otherwise it will build an army.

 

[Moonsinger]

Originally posted by betazed
AFAIK, we never saw this is civ 3. However, with PTW this behavior occurs too regularly and prompts me to think that leader rushing has been added to the PTW AI (I could be wrong about this).

EDIT: I also think the AI never stores a leader. If there is a wonder that it can rush then it might. Otherwise it will build an army.

Bingo! I think you are right about that. The AI in PTW seems a lot smarter than in Civ3.

 

[Arathorn]

The reason I'm convinced it's Human move, AI move, human production, AI production and repeat and not Human move, AI production, AI move, Human production and repeat is the following....

The AI captures a city during its move and immediately rushes a defender. That defender will appear in the city before you have a chance to take it back. This explains one of the "AI is cheating" cries that we hear occasionally. The AI moves in with one offensive unit and "magically" has two there (the attacker who captured it and the rushed defender) to defend the next turn (human movement turn, that is).... It's all explained by my ordering. If AI production were before its movement, then this wouldn't be possible.

Actual unit healing and re-setting of movement counters and such appears to be after production but before move, too, if anybody cares....

And my recollection of vanilla Civ3 was that AIs would use leaders to rush projects occasionally, but they'd rush whatever was handiest, which was almost never a wonder. I *think* (e.g. no evidence) that the intelligence of the rushing was just improved, but I could be wrong.

Arathorn

 

[aeldrik]

still leaves the option Human move, AI Move, AI Prod, Human Prod open. while playing, usually you can move your units right after going through prod of your cities, there just isn't time for AI prod between Human prod and Human mov.

 

[Qitai]

It is Human move, AI move, human production, AI production. If both you and the AI completes a wonder on the same turn (regardless if it is handbuild or GL rushed), you will get it first.

Also, I have seen AI rushing Wonder many times. I am still in vanilla Civ3 1.29. So, no doubt that AI is rushing wonder. In fact, my impression is that they do it 50% of the time when they get GL.

 

[Arathorn]

If it went "Human move, AI Move, AI Prod, Human Prod" as you suggest, the human would lose "ties" in wonder races. Since it's well-established that the human will win ties in wonder races (simply have you and the AI civ set to finish the wonder both "in 1 turn" and you will win), that rules out your proposal.

There is plenty of time for the AI to produce. Production and ordering is very fast. It only takes time for you because of the graphics change and such. There's no need to calculate routes in the middle of production, so there's plenty of time for the AI's production to complete and for new orders to be given between the human production phase and the human movement phase.

Aeldrik, why won't you accept that I'm right?

Arathorn

 

[aeldrik]

ok, Arathorn, you are right....

 

[Moonsinger]

Here is the little test game that I suggested previously:

http://www.civfanatics.net/uploads5/Moonsinger_test1.SAV

8 civs on a tiny map right next to each others (PTW v1.21f) and everyone started out in the Middle Age with Republic and 1000 gold in treasury.

Basically, my findings as follows:

At 4000 BC, I founded my capital at the starting location and set to build a warrior next then end my turn. Next, the Germans and Indians also founded their capital at the staring location at 4000 BC too. Both their capital were only 2 square away from mine.

At 3950 BC, my production box is now has about 3 shields. I immediately built the embassy in Berlin and saw that their production queue is 100% empty. Of course, before I ended my turn, I also paid gold to investigate all the AI cities again, and found out that they still have no shield in their production queues. Basically, the game movement is as follows:

Human player goes first. Then come human production, human action plus movement plus whatever. Then come the AI production plus the AI movement/action plus whatever. Basically, Civ3 is a turn base game and everyone complete their turn 100% before the next player begin their turns. My save game is included as above, please check it out for yourself.