How Does the AI Keep Up In Science?

[Cordelayne]

I think this issue has been discussed in several threads, but I'm curious if anyone has looked at this systematically: how does the AI tend to consistently keep science production so high? It's got to be more than buildings and specialists.

I've had several games at Noble level, on standard maps, where I have systematically expanded and conquered the globe, while the AIs tend to sit in little huddles, quaking with terror knowing that sooner or later they will be next to be assimilated.

I always play with tech trading turned off.

So I constantly have situations where I have about 50 cities, have at least 30 cities maxed out on science buildings, have used Great Scientists to make many academies, have made scientist specialists, have culture set at 2 or 3--and in contrast, an average AI will have maybe 8 cities, one-quarter of my civilization score, and yet they are close to me or even superior in tech.

I don't get it. How is that possible? Does the AI get a very large science bonus at Noble level?

[GeoModder]

I think it's mainly the maintenance costs that kills your science progress until you catch up with a high number of libraries/Universities.

Remember that in the F4 technology screen you can't see the AI techs from which you don't have the prerequisites.

You can try to check the AI's research rate with your spies. I've done it once as well when a small civ had only 5 cities, and they actually managed to stay ahead in lab output throughout most of that game. Only my superior production gave me victory in that one.

[James M]

Try playing an OCC game. It's quite possible to far surpass the AI in tech with only one city at noble.

The game makes small empires of specialised cities quite possible.

[Urban Ranger]

Re: How Does the AI Keep Up In Science?

Originally posted by Cordelayne
So I constantly have situations where I have about 50 cities, have at least 30 cities maxed out on science buildings, have used Great Scientists to make many academies, have made scientist specialists, have culture set at 2 or 3--and in contrast, an average AI will have maybe 8 cities, one-quarter of my civilization score, and yet they are close to me or even superior in tech.

How much of your income is used for research? 50%? 40%?

[Cort Haus]

The AI's often have a lot of cottages, at the expense of production. As towns in the end-game they can pump a lot of research.

[Enigma_Nova]

towns

Urge to Pillage ... Rising ...

[Blake]

Another factor is Trade.

Trade does not benefit the large empire, it benefits the small empire.

Small empire of 10 cities, creates 10 outgoing trade routes.
Huge empire of 50 cities creates 50 outgoing trade routes.

If the two civs have open borders, small civ gets (up to) *50* incoming trade routes, while Huge Civ gets just *10* incoming trade routes. Trade is one of the great equalizers in Civ4. A small friendly civ could be getting 2-5x as much trade income than a huge, disliked civ.

Basically your huge civ is generating a HUGE amount of trade income for other civs .

[=DrJambo=]

Is that true Blake? If so, how do you know?

[Drachasor]

The AI also has a discount on tech research at higher levels. I think it has one on noble, but I am not sure (and I am too lazy to check the xml).

I think the AI is probably better at getting along with the AI generally too. AI tech trading can make it so a group of AIs can keep pace with one larger Civ. The higher maintainance costs of the larger Civ can help this too.

-Drachasor

[Drachasor]

Originally posted by Blake
Another factor is Trade.

Trade does not benefit the large empire, it benefits the small empire.

Small empire of 10 cities, creates 10 outgoing trade routes.
Huge empire of 50 cities creates 50 outgoing trade routes.

If the two civs have open borders, small civ gets (up to) *50* incoming trade routes, while Huge Civ gets just *10* incoming trade routes. Trade is one of the great equalizers in Civ4. A small friendly civ could be getting 2-5x as much trade income than a huge, disliked civ.

Basically your huge civ is generating a HUGE amount of trade income for other civs .

You always have a certain number of trade routes per city, depending on what techs and wonders you have (and what improvements are in a given city). The number of cities you have doesn't change this. A larger Civ will always have more trade routes than a smaller Civ. However, the larger the Civ the more likely a lot of those routes are relatively useless (only 1 or 2 gold per route or so).

-Drachasor

[Arrian]

1) The AI is a cottage spammer.
2) I'm not sure about the trade stuff Blake is talking about, but that's a possibility.
3) The AI may be using great people for tech.
4) AFAIK, the AI does get *some* small bonuses even on Noble. I think the line on that level is that it is "closest" to even... as opposed to exactly even.

Without seeing your empire it's tough to say whether you should be shooting ahead.

-are you playing a civ suited to research (financial or, to a certain extent, organized?)
-are you utilizing cottages and/or other commerce-boosting measures (great lighthouse, colossus, harbors, airports, free trade...)?
-is the sheer size of your empire hurting you, perhaps? Courthouses? FP? Versailles?

I'm not accusing you of playing poorly, I'm just pointing out that it's hard to tell based on your description alone.

In AU101, we played as Washington (Org/Fin), which is one of the best combos for science/economy. I played on Prince, although due to a quirk in how the WB scenario was set up, the AI might have lacked its free units (it still got its other bonii, though). I didn't get huge until late in the game, but by the renaissance I was ahead in tech and pulling away. The AI made a late push to challenge me for the SS launch, but I was researching future techs and they were just starting their SS's. One of the opponents was financial - Elizabeth of England.

I do generally find myself surprised at some of the AI's tech prowess... but then I usually find that it's a financial civ and their terrain improvements are mostly towns. That's great for research, but bad for production - such a civ is emminently stompable. Or at the least pillagable.

-Arrian

[ZargonX]

I've noticed in my past few games that the isolated AIs actually tend to be the science powerhouses later in the game. England especially, if it starts out with good land by itself, will catapult ahead tech-wise around the late Renaissance if not invaded at home. I was the tech leader in my last game until that time period, at which point I was ahead of the other AI by 2 or 3 techs, and England was ahead of me by 5, then 6, then 7! I was running at 90% sci, with more cities than her. But, like has been said, her cottages, plus the financial trait, plus no wars, made her a tech beast.

[Crossfire]

Once you get spies though it is pretty easy to stop the AI from completing the space race. I have never had the AI really get to me with spies. Has anyone else?

[Cordelayne]

Thanks for many fascinating comments. I'll have to try OCC. I have a hard time believing that one city could be ahead on science while 50 cities are not, but I accept what is claimed. As to some questions about my civ in this game, it's the Scots (which I invented), Industrious and Philosophical. I've kept science spending steady at 6 or 7 throughout most of the game. I favor farms over cottages, but build cottages when nothing else can be built. In this particular game, the smaller AIs who are keeping up with my science are Washington, Caesar and Alexander, none of whom are Financial.

There have been conflicting reports, but I suspect the AIs do get a flat science bonus, even at Noble.

Also, I don't really understand Blake's thoughts about trade, but that seems a promising theory.

I don't really mind that the AIs compete so well. I just always prefer to understand the game mechanics as deeply as possible.

As for last question, oh yes, I've definitely had the AIs eat me alive with invisible spies! You have to put spies plus military units on top of every key resource.

As to question about sheer size of my empire, I've never gotten into trouble on this game. I play in pulses. Max out first island. Expand some more. Max out. Expand some more. Invade. Assimilate for a while. Invade. Assimilate for a while. I have a courthouse in every city, plus Forgotten Palace, plus Versailles too I think. I usually build lighthouse first, then courthouse. In this particular game, I'm just crushing the daylights out of the AI, and it's obviously time to move up from Noble level.

[Blake]

The trade stuff is a bit weird but it is how it works.

Basically, you'll only ever find a particular foriegn city in *exactly one* of your cities trade route boxes. However, one of your cities can be trading with multiple cities of the same civ. This means that trade routes aren't paired - if they were paired, that would introduce a logical contradiction. (Spies are good for confirming this)

The amount of income you give other civs is basically proportional to your city count and size. Now dosmetic trade routes can actually be pretty good sometimes because they work entirely differently (just in case anyone isn't confused, yet), in that basically every one of your cities can trade with your biggest trade city (again, trade routes are clearly not paired!). This means if you have a few uber-huge cities you can have pretty decent domestic trade.

But in any case the lions share of trade income goes into the top 3-5 trade cities, pretty much regardless of empire size. For the rest of the cities, the trade basically covers upkeep, if you're lucky.


My own general observation is that AI's have a much great cottage focus. I like to look at where my beakers come from, usually about a third to a half comes from my Oxford city, most of the rest comes from 3-5 of my trade/cottage cities. I then have another dozen cities that contribute nearly nothing.
The AI may only have a half dozen cities, but if they are ALL optimized for beakers, they are going to at least match my research. The main benefit of big empires is extra production, the added research is quite marginal unless you really focus on cottage-spam.

Now, the big AI weakness is they play in a short-sighted way, they are good at researching spaceship techs - even sacrificing great people for them, but appalling at actually building the ship. I wouldn't be suprised if my 2 best hammer cities surpass their entire empire hammer output.