A small trick to keep AI happy

Qitai

.
Joined
Jan 22, 2002
Messages
1,177
Location
SG.MY.TW.US
I notice many players have problem with keeping AI happy. So, I would like to highlight a small trick that helps slightly to keep AI happy. I had Gracious and Polite AI in GOTM20 even though I keep breaking MAs.

The trick is that whenever you sell something and the AIs couldn't afford it. Instead of just settle for less (i.e. just sell it for whatever they have), why not just give them gold to make up the difference.

As an example, say you can sell Gunpowder at 1400gc or the equivalent of it. But AI can only offer 20gpt. So, instead of just accepting that. First give AI 800 gold then follow by a 20gpt+800gc for gunpowder trade. I can get the AI relation gift bonus (worth 10) virtually to last to infinity by doing this since the AI never have enough gpt to pay me the "true market value" of my techs.

The other major factors would be not to raze any cities you capture. To know more about AI attitude, read Bamspeedy's post on AI attitude.
 
Very cunning Qitai...
So that's one of the tricks you used to get those early domination wins in GOTM19 and 20......
I'll remember this...
 
Originally posted by Qitai
The trick is that whenever you sell something and the AIs couldn't afford it. Instead of just settle for less (i.e. just sell it for whatever they have), why not just give them gold to make up the difference.

As an example, say you can sell Gunpowder at 1400gc or the equivalent of it. But AI can only offer 20gpt.

Great idea. But how do you know how much they are willing to pay, if they don't have the gold?
 
Rough estimates I am aware of.

(1) Monoploy tech - 150% - 200% Research cost.
(2) Non-monoploy Tech - 60% - 100% research cost depending on how many people has the tech
(3) If you just sold the tech to another AI, the "true market value" is just slightly diluted if the first wasn't sold at monopoly price or halfed and slightly diluted if it is.

To get the research cost, I have written down all the tech multiplier on the tech card that comes along with civ3. And a simple estimate is just Tech multiplier*Map Size factor*Difficulty.

Normally, to be on the safe side, I give until slightly less (~80%). And instead of giving the full 800 gold in one go, I give 100 gold by 100 gold (or some figures which is safe). And ask what would they exchange for my tech. The limit is near when you find they are not giving everything they have (just make sure the gc they have is in multiplies of 10). Usually they give 10%-15% more of what they offer. This process may take abit of time if you really want to avoid giving too much, but it is quite worthwhile if you want AI to be happy and 800 gold can last you for quite some time according to Bamspeedy's research.


As an additional note, if they are offering much less than the above estimates, you know they are probably researching that particular tech you are selling.

Also, I overgive sometimes, but usually over by not more than 30 gold. These can usually be easily recovered by selling maps every turn.
 
I already use this little trick every time. I only gift very small amounts (100-200g) but I'm selling cheap early techs anyway.

Of course if they are totally bankrupt you can just gift them the tech but often they have 25g (razed a barb village) and I would not sell them Mathematics for that. So I gift them 100 and sell Maths for 125.
 
In the here and now we give millions to billions to other governments to help their military, infrastructure, agriculture, etc. And do you think we expect for them to look to Amercan firms to supply the jets, power plants, chemicals, etc.? You bet! At the same time this certainly leads to relations that are better than having ignored them.

Big deals from other countries now come at the price of contracting services back to them. This is another way of providing goods and technology for cash.

Great idea. Have had CivIII for some time but just started playing it pretty hard now that my boys have taken a shine to it. Great opportunity for teaching.
 
This is a great idea!

Too often in my games I notice the ai never has money to buy my techs. This makes being a middleman impossible for me. Now this trick comes along and changes everything! :D
 
Geat idea. Any other tricks to help us novices get above Reagent Level.
 
on an axillary note i tried my first emporer pacifist game today (despotism army designed only to control population and protect against barbarians--no army after republic). Theres actually some value in perfecting these strats. They are obviously helpful when your trying a more balanced beat em down approach.
 
Burroughs Economics!

"1: Never give anything away for nothing.
2: Never give more than you have to give (Always catch the buyer hungry and always make him wait).
3: Always take everything back if you possibly can."


--William S. Burroughs, NAKED LUNCH

Apply these to the AI, you shall never go wrong.


Later!

--The Clown to the Left
 
Originally posted by Qitai
Also, I overgive sometimes, but usually over by not more than 30 gold. These can usually be easily recovered by selling maps every turn.

i always had the impression you never trade maps so the AI doesn't expand past your territory

GREAT tip by the way, although i rarely trade tach for anything other than a MPP, this will come in very handy.........:nya:
 
Originally posted by ybbor
i always had the impression you never trade maps so the AI doesn't expand past your territory
Nope! They'll walk right past you, maps or no maps. They know the map already, so you might as well get some money out of them when they are prepared to pay.
 
oh no, not another exploit :(

of course this is a great idea if you want to play on levels you usually can't reach. but this is a missuse of the AI programming and therefor an exploit.

same as suicide galleys, 40 turn polytheism gambits, etc.
 
It's ridiculous to call suicide galleys or 40 turn research an "exploit". These features were deliberately put into the game by the designers. They could have chosen to have galleys sink at sea 100% of the time, or to make the maximum research time be 100 turns. They didn't because they deliberately wanted to allow these strategies. You might disagree with that, but using these methods is just doing what the designers intended.
 
Top Bottom