100 turns...

[Derelict]

Lo' all, hope turkey day was all good for those inclined and exposed to such things.

Anyhow, the purpose of this post was a little "test" of sorts that I did. Basically I ran 100 turns of a game and every 10 turns I went into the editor to observe what the AI had done during each round. I did this for 10 rounds, or a total of 100 turns (hence the title of this post, eh?).

I thought it might be informative and hopefully a bit educational to see how the AI goes about its business. 100 turns effectively takes the game from 4,000 B.C. to 375 B.C., or through the entire ancient and roughly half or more of the classical period.

Three AI in particular were selected for the test; Frederick, Ghandi and Genghis.

Frederick was selected because I wanted to get an idea what a more culture minded civ that didn't have ready access to religious techs (via mysticism) would pursue.

Ghandi was selected for both his religious start (i.e., mysticism) and because of his industriousness. I thought it would be instructive to see how he focused his energies between the two from the start.

Genghis was selected because he was an aggressive AI and I wanted to see what sort of build up a military minded Ai would take early game. As well, I thought the expansive trait would be interesting to observe from an AI in so far as cities went.

I took Washington for no specific reason. Every choice made was a recommended one (for city placement, tech, building, etc.) and I made no specific intent to interfere or otherwise avoid contact with the other AI along the way.

The map was a small pangea world with typical settings all around. I chose pangea for the fact that there was one less AI for the world than normal and it would engage the AI with one another in a fashion more typical to a continents world with the standard number of players, or a terra world period. It also made it a little easier to go about this test.

Here are the results of the test, I've avoided drawing any conclusions so far and the data is sorted by leader (in accordance with date) to get a better idea of their progression singularly. I did not repeat the techs they had at each time frame, and just show any new techs they discovered (with the exception of the starting time period, naturally)

========================================

4000 to 3600 B.C. (ten turns)

Frederick -- 1 scout, 1 warrior and is building a worker. City size is 2. Techs: Agriculture, Hunting, Mining
Ghandi -- 2 warriors with another being built. City size is 2. Techs: Techs: Fishing, Mining, Mysticism
Genghis -- 2 scouts, 1 warrior with another warrior being built. City size is 2. Techs: Wheel, Hunting, Archery

3600 to 3200 B.C. (twenty turns)

Frederick -- 1 warrior, 1 worker, 1 scout. Constructing a warrior. City size is 2. Techs: Wheel
Ghandi -- 3 warrios, building a worker. City size is 2. Techs: Meditation, Polytheism (Hinduism & Buddhism)
Genghis -- 2 scouts, 2 warriors, building a worker. City size is 2. Techs: Mysticism

3200 to 2800 B.C. (thirty turns)

Frederick -- 2 warriors, 1 scout and a worker; building a settler. City size is 2. Techs: Archery, Animal Husbandry
Ghandi -- 3 warriors, 1 worker; building a warrior. City size is 3. Techs: Agriculture
Genghis -- 2 warriors, 2 scouts and worker; building an archer. City size is 3. Techs: Agriculture

2800 to 2400 B.C. (fourty turns)

Frederick -- 1 archer, 2 warriors, 1 scout and 1 worker; building a settler. City size 3. Techs: Fishing
Ghandi -- 4 warriors, 1 worker; building a settler. City size 3. Techs: Hunting, Archery
Genghis -- 2 scouts, 2 warriors, 1 archer and 1 worker; building a settler. City size 3. Tech: Polytheism, Masonry, Animal Husbandry

2400 to 2000 B.C. (fifty turns)

Frederick -- 2 archers, 2 warriors, 1 worker; building work boat and archer. City sizes 4 & 1. Techs: Bronze working. Civic: Slavery
Ghandi -- 1 archer, 4 warriors, 1 worker; building archer and settler. City sizes 3 & 1. Techs: Masonry
Genghis -- 2 archers, 2 warriors, 1 scout (bear killed one); building walls and archer. City sizes 4 & 1. Techs: None

2000 to 1600 B.C. (sixty turns)

Frederick -- 3 archers, 2 warriors, 1 worker (scout killed by barbs); building archer & settler. City sizes 4 & 2. Techs: Pottery, Masonry
Ghandi -- 1 archer, 4 warriors, 1 workers; building settler and archer; City sizes 3 & 2. Techs: Monotheism
Genghis -- 2 archers, 2 warriors, 1 scout (one killed by barbs?) and one worker; building settler and archer. City sizes 4 & 2. Techs: Monotheism

NOTE: First "round" of turns in which barb (non-animals) have appeared, including one city. Genghis also got Judaism. Genghis and Ghandi both go organized religion civic.

1600 to 1200 B.C. (seventy turns)

Frederick -- 2 axemen, 1 spearman, 4 archers, 1 worker; building 2 archers, 1 settler. City sizes 5, 3 & 1. Techs: None
Ghandi -- 2 archers, 3 warriors, 1 worker; building archer and barracks. City sizes 4 & 2. Techs: Priesthood, Pottery
Genghis -- 5 archers, 2 warriors, 1 worker and 1 settler; building archer and worker. City sizes 6 & 3. Techs: Fishing

1200 to 875 B.c. (eighty turns)

Frederick -- 6 archers, 2 spearmen, 2 axemen, 1 worker & 1 settler; building Pyramids, spearmen, worker. City sizes 6, 3 & 2. Techs: Iron working
Ghandi -- 2 archers, 4 warriors, 1 worker; building stonehenge, settler, archer. City sizes 5, 3 & 1. Techs: Writing
Genghis -- 8 archers, 2 workers; building Pyramids, settler and archer. City sizes 6 & 3. Techs: Bronze working (went slavery civic).

875 to 625 B.C. (ninety turns)

Frederick -- 6 archers, 2 spearman, 2 axemen, 1 worker; building Pyramids, spearman, settler and worker. City sizes 6, 3, 2, 2 & 2. Techs: Writing
Ghandi - 5 archers, 2 warriors, 1 settler 1 worker; building archer, obelisk, Stonehenge. City sizes 5, 3 & 2. Techs: Bronze working
Genghis -- 10 archers, 3 chariots, 2 workers, 1 settler; building chariot, archer Oracle. City sizes 6, 4 & 2. Techs: Writing, Pottery

NOTE: Frederick took his two axemen to the far southwest (of everyone) and captured a barbarian city, making his 5th city.

625 to 375 B.C. (one hundred turns)

Frederick -- 8 archers, 3 spearmen, 2 axemen, 2 workers; building Pyramids, archer, spearmen, axemen, settler, barracks. City sizes 6, 5, 3, 3, 2 & 1. Techs: Sailing
Ghandi -- 7 archers, 2 axemen, 2 workers; building 2 spearmen, 2 archers. City sizes 6, 4, 3 & 1. Techs: None
Genghis -- 12 archers, 4 chariots, 4 workers; building Oracle, 2 archers, barracks, J. missionary. City sizes 6, 4, 2, 1 & 1. Techs: Sailing, Priesthood

NOTE: Barbs are notably archers now vs. the warriors they were earlier. Genghis also captured the 2nd (and only other) barb city, making it his 5th city.

Final Score:

757 Frederick
634 Genghis
586 (Me)
536 Ghandi

========================================

I'll refrain from making any but the most overt of observations just yet. A couple things I would remark on:

* Genghis, by the end, had roughly 25% more military units than anyone else. He did not have access to copper, but had two horses available.

* Ghandi, while having 2 sources of copper from early on, did not go for bronze working until late (sometime in the last 20 turns), and stayed with mostly warriors even though he had the archery tech.

* Frederick went very quickly for bronze and iron, being the only one in the game with iron by the end and had bronze before even 2000 B.C. He didn't waste any time hooking it up or building respective bronze units, which he also used in a deliberate manner (via his 2 axemen) to capture a distant barb city. I was amused that he did so, and held it, since it was fairly distant and likely cost him a decent bit to hold onto maintenance wise.

* Ghandi was the only one to build any cultural buildings (an obelisk), not counting wonders of course.

* None of the AI started on wonders until eighty turns into the game. All three started on a project during the 1,200 to 875 B.C. time period. Genghis was the only one with two cities at the time, but he did have a settler enroute to a new location.

* It may be related or not, but both Frederick and Genghis started on the Pyramids with size 6 cities. Ghandi did not pursue that wonder and instead went for Stonehenge. One might think that Genghis would find Stonehenge more desirable as well considering he had a religion started and could build it quickly. I'm presuming here (which I said I'd avoid, but anyhow...) that civ's with larger cities will instinctively pursue much more engrossing projects without respect to some more direct benefit like religion.

* All three switched to slavery but I do not think any used it to pop-rush a unit/building in the course of things. I am curious to know just when the AI feels it desirable to do such a thing, though I'd bet it reserves that option for when it is invaded and needs or wants additional defensive units.

* All three had agriculture no later than 2,800 B.C. Ghandi put off hunting and archery until 2,400 B.C. in favor or religious tech, Genghis pursued it immediately and Frederick favored development techs before archery, though still had it by 2,800 B.C..

* Frederick expanded the fastest, having both his fourth city via settler up, and a fifth by capturing a barb city, before anyone else had really even started to move their settlers out for their fourth.

* Ghandi never got animal husbandry, though he never settled a location that had nearby access to any animals, including horses until his last city was established, and even then it's borders had not yet expanded enough to make the cows available. I'm curious if the AI would pursue it anyhow if it *knew* that horses were within range, even if a similar human player wouldn't be able to see them at the time.

Okies...I'd be curious to hear what some of you think from this limited bit of data. I realize that a few more tests like this would be more revealing and, for that matter, the AI may have made decisions based on specifics to this particular world, etc. Even beyond all that, I think there are probably some patterns already that can be gleaned from it that could be helpful for strategies, such as wonder building, military/rush strats, etc.

[snafuc4]

Good post! I'd be /very/ interested in a similar analysis of the Demo world (Ghandi, Alex, Monte(? I think) & Caesar) if someone can extract the map from there..

Did you note your own positions during the test? Assuming you took the first recommended item on the list & only settled/improved blue circle areas it may be an interesting comparison..

Interesting that Ghandi took fishing; was he the only one to start on the coast?

[Derelict]

Originally posted by snafuc4
Did you note your own positions during the test? Assuming you took the first recommended item on the list & only settled/improved blue circle areas it may be an interesting comparison..

No I didn't as much as I wish I had. Even still I can tell you what I ended up with for comparison, and yes I only settled blue circle/recommended locations and always selected the first unit/building when recommended. I did not, however, adjust any of my cities regarding citizen allocation (either directly or through the food/production/etc. buttons), or for specialists.

*** That does beg a question--are we under the assumption that the recommended techs and unit/build orders are by and large the same that the computer would select if it were playing that civ? I've rather assumed that to be the case, more or less...

At the end I had

Units -- 4 archers, 7 warriors, 3 workers (automated, naturally). I did not have access to copper or horses (though one city would have a border expanding into a horse resource before much longer). I did have 1 elephant resource already hooked up.

Techs -- Fishing, Sailing, Wheel, Agriculture, Mining, Pottery, Animal Husbandry // Mysticism, Polytheism, Masonry, Priesthood // Archery, Bronze working // Writing, Alphabet

Code of laws was being researched currently.

City production was Settler, Oracle, Archer and Warrior. City sizes were 8, 5, 3 & 2.

Interesting that Ghandi took fishing; was he the only one to start on the coast?

Ghandi had a start on a small inland sea. Frederick started on a coastal/ocean and even had a fish resource in range of an expanded border, but didn't go for fishing until some fifty turns in. Genghis had only inland cities until the last round leading into the 100th turn, and then he had 2 established (both of his size 1 cities).

Also, as an added note, this was run on Noble difficulty setting in case that wasn't made clear from above.

[Dis]

I may have missed it in the post, but what difficulty level was this?

edit: I see in the post just above mine

[Dis]

How did Frederick have that second archer at turn 50? Scouts don't upgrade to archers do they? (I never build them).

[Derelict]

Originally posted by Dis
How did Frederick have that second archer at turn 50? Scouts don't upgrade to archers do they? (I never build them).

In a decent sized city an archer would only take three turns or so to complete.

From the looks of it, Frederick had started a settler probably towards about fourty turns in, was still building it as time reached fifty, but it would've been very close to completion at that point. I don't imagine it was anything difficult for him to produce another archer in less than 8 turns or so (assuming the settler was within a turn or two of being completed, which probably was the case since we know he had it under way for at least 11 turns already).

======================================

A couple more observations

Another observation I didn't make earlier--it looks like Genghis switched from building the pyramid sometime before the last twenty turns. I'm not sure how long he was building it by that point but he had started it at roughly the same time Frederick had (was building it at 875 B.C.) and switched to building the Oracle instead within ten turns (by 625 B.C.).

The other thing that caught my attention was that Ghandi had only just gotten his second worker out by the last 10 turns--I thought he may go for two earlier than that given that his economy would've been that much more developed with the UU/Fast Workers. Perhaps the AI views the unit as 1.5 workers or some such already, for example, or perhaps doesn't even consider the UU beyond a regular worker at all. For example, Frederick also got his second worker out late, within the last 10 turns, while Genghis had a whopping four workers by then! In fact, he had a second worker kicking around for at least twenty turns.

With that in mind, and being that Genghis is a military-centric AI (we must presume given his traits), it would seem that earlier development of the economy would be in priority so that he could support a larger number of troops. As mentioned, he had 25% more units and was also operating under the organized civic, resulting in higher costs all around. It should be noted, in conjuction with this, that he had fewer cities until the very end (and then tied with Frederick).

Military considerations

Finally, from a military perspective, it may pay off to archer/warrior rush (or if Egyptian, Incan or Perisan...naturally with respective UU's) a civ like Ghandi since military seems to be last on the agenda. He didn't even go for bronze working until some 90 turns into the game and this, assuming the AI was aware of them even without the tech, had two sources of copper and a potential iron source if I recall. At the very least, and again assuming he follows this sort of pattern with any sense of regularity, storming a city of his with chariots or an early UU could present the likelyhood of capturing a religion he had so intently founded not so many turns earlier. Waiting too long, however, will probably result in some accumulated culture (and hence defensive bonuses). He had, after all, obtained at least two religions by the 20th turn.

Or, on the other hand, if any sizable force (even if not particularly advanced) could be drummed up, it may pay to attack a military civ with pillage in mind specifically, in order to weaken them up and make any aggressive plans they may have more costly in the near term, at least, or utterly put off until they can get their economy back in order. Naturally this holds some longer term repercussions, particularly with an aggressive AI, but if there are any thoughts of expanding in their direction (or simply invading for capture of a cities), damaging their economy early on would be a big plus while you continue build-up and development. Of course, even an early invasion of Genghis or a similiar aggressive AI would meet with substantially stiffer resistance than, say, a Ghandi.

In any case, from a purely military perspective, none of the AI had bronze working + copper (for bronze age units) until 70 turns in. It's quite possible that a player could achieve as much and even have a few bronze age units in the field by that time, if they were to focus intently on it. It still stands with some risk but a high potential reward, while also likely causing city development to suffer and potentially obviate a chance at some of the more production-demanding wonders (like the pyramids).

[Derelict]

P.S. Scouts upgrade to explorers...

[snafuc4]

Originally posted by Derelict

Ghandi had a start on a small inland sea. Frederick started on a coastal/ocean and even had a fish resource in range of an expanded border, but didn't go for fishing until some fifty turns in. Genghis had only inland cities until the last round leading into the 100th turn, and then he had 2 established (both of his size 1 cities).

OK, so the AI doesn't /appear/ to take into account city expansion. Would another AI civ having a coastal start (but without special resourses (clams, fish or whales) research fishing? We don't know.. yet

Originally posted by Derelict
Also, as an added note, this was run on Noble difficulty setting in case that wasn't made clear from above.

Umm.. that's the one where there's no advantage/disadvantage whether you're human or AI, right? (sorry, can't find it in the manual ATM)

[snafuc4]

Originally posted by Derelict
P.S. Scouts upgrade to explorers...

Yes.. & after that they're pretty much useless. How much map is there left to explore in the late game? How many huts? OK, they can manage as DEWS for a while but there are better units for that. They should be upgradeable to spies or (my preference) civ3 geurillas - they know the layout of the land intimately & are used to working on their own..

[Derelict]

Originally posted by snafuc4


Yes.. & after that they're pretty much useless. How much map is there left to explore in the late game? How many huts? OK, they can manage as DEWS for a while but there are better units for that. They should be upgradeable to spies or (my preference) civ3 geurillas - they know the layout of the land intimately & are used to working on their own..

By and large I agree, though I have (and do) use them from time to time to run around a rival civ whom I have open borders with to keep an eye on their development, cities, etc. I think I've used them a couple times on terra maps as well, but even then their use if fairly limited.

And you're right, they are at a dead end--there is no upgrade for explorers. Like you mention, guerrillas or maybe snipers or something would be a nice addition for them to upgrade into.

[cal_01]

Originally posted by Derelict


By and large I agree, though I have (and do) use them from time to time to run around a rival civ whom I have open borders with to keep an eye on their development, cities, etc. I think I've used them a couple times on terra maps as well, but even then their use if fairly limited.

And you're right, they are at a dead end--there is no upgrade for explorers. Like you mention, guerrillas or maybe snipers or something would be a nice addition for them to upgrade into.

Perhaps a mod to put back in Partisans?

Anyhow, I definitely agree with your previous post about rushing Ghandi. Most other civs, in the beginning, would try to rush me with a stack of 6-8 troops. When I hit Ghandi with Chariots and Axemen, he had nearly zero military. I ended up taking his capital and made Buddhism my own religion.

Though, I'm kind of surprised at the AI activity at this point... 6 cities by turn 100?! I can't even get that high without sacrificing my tech/finance levels.

[snafuc4]

Originally posted by Derelict

And you're right, they are at a dead end--there is no upgrade for explorers.

Is there /any/ other unit that cannot be upgraded to something in the late game? Settlers, obviously, but even workers get their chance..

Originally posted by Derelict
Like you mention, guerrillas or maybe snipers or something would be a nice addition for them to upgrade into.

Snipers - luvvitt! Speciality trait - one shot at taking out [player selects] unit in a stack Variation on fortification bonus needed here - the sniper stays in one place to gain a bonus to his 'to kill' IYSWIM

If the sniper fails, he must move instantly, as he's -25% everything (at least), & can't take another shot until out of visual contact with enemy (or enemy's allies, natch) until he's bedded down again (ie at least 1 turn with no visual contact, else he's spotted & chased)

Very expensive unit to build, start at longbowman (although some civs may be able to do this earlier). Movement: 1 (fixed); class.. archery; promotions.. well, I'll leave it for you to decide..

[cal_01]

Originally posted by snafuc4


Is there /any/ other unit that cannot be upgraded to something in the late game? Settlers, obviously, but even workers get their chance..



Snipers - luvvitt! Speciality trait - one shot at taking out [player selects] unit in a stack Variation on fortification bonus needed here - the sniper stays in one place to gain a bonus to his 'to kill' IYSWIM

If the sniper fails, he must move instantly, as he's -25% everything (at least), & can't take another shot until out of visual contact with enemy (or enemy's allies, natch) until he's bedded down again (ie at least 1 turn with no visual contact, else he's spotted & chased)

Very expensive unit to build, start at longbowman (although some civs may be able to do this earlier). Movement: 1 (fixed); class.. archery; promotions.. well, I'll leave it for you to decide..

To the best of my knowledge, melee units can't be upgraded to gunpower units. Or am I wrong here?

[Blake]

Explorers make GREAT auto-explorers, since they hardly ever die to barbs. Especially if they start at 1 or 2 promotion levels for both woodsmanII and guerillaII, then they'll hardly ever stray from those nice defensive terrains. They are actually a remarkably good unit strength-wise considering how early they can be got and that they require no resources. They are good on the all-land maps like Highlands where they is much exploring to be done.

[Dubhghlas]

Though, I'm kind of surprised at the AI activity at this point... 6 cities by turn 100?! I can't even get that high without sacrificing my tech/finance levels.

You can do six cities easily. You use a settler pump.

Take a city, say, your starting city, especially if it has lots of food available early. Build warriors or maybe you have access to an early cheap building. Wait for your city to get to size 3. Then, when it is size 3 or 4, start pumping out settlers, one after the other. They won't take long to produce, because you have so much food and hammers being made. You will find a lot of settlers can be had quickly that way.

Better, pump out only three, then go back to growing. Wait for the first two added cities to get to size 3, then each produces a couple settlers. You can do that, too, in 100 turns, I would imagine.