AI tournament: Season 3

[Major Tom]

1. JC. Comes out on top after the early wars.
2. De Gaulle. Grab lot of land in the late game.
3. Sury. Also agressive but can lose focus.
4. Gilgy. Takes some points as peacevassal.
5. Peter. Stays in the woods.
6. Wang. 3 city vassal in this wolfpack.
7. Monty. 1 city vassal, the stupid a*****e.


Thanks to Kossin for hosting this.

 

[kossin]

Starts today, few hours later.

 

[Noble Zarkon]

Starts today, few hours later.

Thanks for all your work in hosting this, looking forward to the next update.

 

[The Diocletian]

Indeed, very enjoyable to read, and to laugh @ the AI. I appreciate the work you do.

 

[kossin]

Group C Game 1 - Turn 1

Here are the starting positions for each AI. The map has undergone some heavy modifications to normalize starting areas a bit, including: adding/removing land; moving some resources around; removing some forests; extending some rivers near AI starts; pushing some jungle further away and the usual hiding of the spectator.

Spoiler :

Gilgamesh gets blessed with the obsolete start. Not the plains cow one, the good one.

Not an outstanding start for Julius, however the Ivory is a nice boon.

Suryavarman's start seems so-so. It follows the 3-resources start though.

Peter: on a peninsula and gold! The AI tendency to get Monarchy early will also be nice. Food is weak though.

De Gaulle will have some health issues, but if he can manage, he has great commerce potential as well as some production.

Montezuma gets the nice 2H city tile, decent food and quite a bit of production. Also smartly avoiding the early religion in favour of economic technology.

Finally, Wang Kon has all the Ivory reserves... well a lot of it.

A look at the land overview.

Spoiler :

The save is joined if others want to run their own RNG !

I thought I'd share some snippets the original map. Some of the reasoning and more details in the next two spoilers. First, the original overview.

Spoiler :

You can see the original spectator island. Peter's area was gigantic and he could get 15+ cities without warring. De Gaulle also had too much land whereas, Montezuma, Wang Kon, Suryavarman and Gilgamesh got stuck with the short end.

Spoiler starting areas :

Gilgamesh had a really bad start. Of course the AI can build some cheap galleys quick but I preferred the starts to be more equal.

I didn't change too much here, just removed some forest for early production. The bigger changes were in cutting land to the east.

For Wang Kon, I pushed the jungle a bit back and made a decent early spot available to his north with the pigs. Some forest was removed as well.

Here I cut a lot of trees and added a hill to compensate. Some land cut to the north.

Riverside copper is not supposed to happen in Fractal maps, the copper was moved. Deforestation.

Well, this was a very poor start with very little food and too many brown tiles. I removed hills, forests, added pigs and green tiles.

This is where the most tinkering went. I had to cut a LOT of land from Peter and also remove a lot of resources, this start was just plain not fair compared to others. He had access to a lot of great sites uncontested and could run away with no effort.


Before giving the go-ahead to the modifications, I kept running trials until 1AD to see if expansion was more or less fair. In the end I felt it wasn't great but it was a lot better than initially and acceptable... then it was all in the AI's hands.

 

[Major Tom]

Julius, what are you doing? Mysticism?

 

[soundjata]

Well balanced map you achieved here. No excuse for the dumb AIs this time.

 

[enKage]

Wang's start is the worst still, but he would vassalise early anyhow.

Gilgamesh is not great too.

It would be funny to swap Gilly's and DeGaulle's starting position, there would be so much fun for French to have both stones and marble in cap

 

[Kjotleik]

If anyone is interested... here is a free-to-use 16-game league setup (se spoiler):

Spoiler :

The numbers inside the Groups (A, B and C) are the rankings. The better the leader, the higher up the ranking.





Pros:

A) Each leader will play against a minimum of 36 other leaders (!) during the season.
B) Each leader plays four games a season (less impact from random bad/good spawns).
C) League is only 16 games in total (but you could add playoffs if you wanted to).

Cons:

A) Each leader misses out on 12 or 15 leaders during a season.
B) Each leader faces 3 or 4 leaders (same pot) four times a season.
C) Each game has (5+4+4) 13 leaders each, thus...
D) ... this set-up is best suited for large maps (and epic speed).

PS!
I do find this thread fascinating... go on, make me laugh AI's...


Yours Sincerely
Kjotleik of Norway

 

[nfw]

From the land overview pic, it seems Peter is giving everyone the finger.

 

[Major Tom]

From the land overview pic, it seems Peter is giving everyone the finger.

Peter is rather more like the flee on the cats tail.

 

[Aje]

I'm gonna make a crazy bet for Monty by Domination on this one.
Second pick is that Petey takes it by Spaceship.

Some noob thoughts/observations:
All AIs are making an archer as their first build. Are they nervous?
Wangy is screwed for sure. I bet he'll be the first to cap.
Jully is teching myst. Will he go for an early religion?
Petey could go far if he makes friends with JC and adopts his religion.
Gilly could snipe this by culture if he starts to spam wonders, but I don't think he'll play that tactic.
Gaully will farm all those pretty FPs and make us laugh. Could do well if he finds the "build cottages" button.
I like Sury starting position, but he will have neighbour troubles...
I think that Monty has a good starting position to go Monty

 

[sinimusta]

To me French seem to have a strong position. Their start has very good commerce and actual food, wet rice, which is good food compared to Russias start. Also they seem to have good land to expand to and enough of it. Gilgamesh should be a peaceful neighbour.

Where Rome decides to expand influences this a lot. Russia could have a strong position if Rome expands more to west. Then maybe if Rome gets involved into some conflict in west Russia could have a good opportunity to backstab. And vice versa Rome could decide to expand east and go for Russia.

I bet for De Gaulle first, Peter second choice. Meanwhile Monty, Julius and sury will tear each other apart after some of them vassalizes Wang. Let's see how wrong I am .

 

[CivNoobie]

1. Sury
2. JC
3. WK

Wang will stay strong, PRO will keep him in the game long enough until gunpowder/liberalism; and that's when Monty gets completely left out and becomes a non-factor. *unless he gets double teamed

 

[Nothingontv]

Sury, DG and Gil will probably do well, but a lot depends on what religions the AI's adopt.

 

[godallmighte]

can't wait for this to continue

 

[kossin]

Few more days, RL issues.

 

[Major Tom]

Hope this thread not die away.

 

[kossin]

No, I'm just having a hard time meeting RL deadlines these past 2 weeks. I'll do at least 1 update this weekend.

 

[kossin]

Group C Game 1 - Turn 50

Sorry about the long delay, life has a tendency to get in the way. I haven't given up on this thread, but I haven't given up on getting a good night of sleep either so it had to wait

~~~

Leading up to Turn 50, I took note of an event that could shape the game in the long run.

First, as some had noticed, Julius had opened up with Mysticism, leading into Buddhism:

Spoiler :

Another glance at AI stupidity. Really? So many roads. And poorly chosen. I know the 'road to' function is poorly written but this is just so ugly to watch.

Spoiler :

At the end of the round, here's the doings of each AI.

Gilgamesh - with his double wonder resource start, it's not too surprising to see him go after Wonders. Furthermore, I was pleased to see he has not one, but two settler parties on their way to found cities 5 and 6. The big problem he faces is amount of land available (and its quality). With Wand Kon in the South and De Gaulle to the East, it seems pretty boxed in already.

Spoiler :

Julius Caesar - Buddhism is nice for border pops, but with resident crazy head Hinduist Montezuma next door is it worth it? Julius favours the 10-Praetorian National Wonder which is a boon with the amount of barb cities spawning around him. 5 (6 soon) cities and growing as Imperialistic civs usually do.

Spoiler :

Suryavarman II - disappointing start doesn't get much better. Look at all those flood plains! He has no chance in hell at getting the Pyramids but maybe the fail gold will finance research while grabbing more cities. One can hope...

Spoiler :

Peter - boxed in by barbarians and not enough units for the moment. Getting the Iron is very important here as he will be in a race with the Romans to grab the great land in between them - swords are inherently much better vs. garrisoned barb archers. Usage of workers isn't very convincing here either. Unfortunately I just noticed a WB artefact in an Ocean tile instead of Coast SW of the Iron.

Spoiler :

De Gaulle - surprisingly good start for the French General. Decent land improvement but slow expansion having just settled his 4th city. With Gilgamesh desperate next door, maybe not the best plan. Still.. 2 cities westward is good. Judaism could become dominant in the north with the 2 early religions in the south not spreading very much.

Spoiler :

Montezuma - no red first but I think city builds speak for themselves. Lacks fishing and some basic economic techs

Spoiler :

Wang Kon - second city in the Jungle. That's why it's not a bad idea to give these kind of positions a small break, AI seems completely oblivious to jungle when evaluating cities. Maybe this statement is incorrect but from what I've seen, they don't consider it enough. The end result is a rather under-developed AI with the wrong peaceweight.

Spoiler :

Various demographics. A bit too messy to make anything out of them yet. Wonders have been surprisingly slow this game.

Spoiler :

The remaining land available.

Spoiler :