Late Game Problem

[WeirdoJoker]

My last game, in which I suffered a UN defeat (it was a last gasp against an impending Spaceship loss, so it wasn't a surprise), presented me with a problem I'd not had before. (Monarch/Ottomans/Continents/70%/Small map.) Despite having plenty of funds, even spending 70-80% on research, it was taking too long to research and I was falling behind without being able to buy the techs I needed (they wouldn't sell at any price once the space race began). I was having to spend 20-30% on the LUX to keep from falling apart. I think the problem was just not enough luxuries - the map was generally resource-poor - but it still doesn't explain the problem. Had I been able to maintain my tech lead - and this is the first time I ever had it and then lost it - I likely would have won the Space Race. Is there an article (that I missed) discussing late-game research & strategies?

 

[justanick]

In the late game research is rather expensive. The costs go up to 400 base costs which at monarch and small means 8889 beakers per tech. The earlier techs are a bit cheaper. The one for modern armor has base costs of 280, which would be 6223 beakers. So for 4 turns per tech you would need 1556 to 2223 beakers per turn.


You need to increase your research output. Obviously you need the commerce bonus from republic. Your cities should have all essential buildings and preferably they are no longer mere cities but metropolises.

Once you have cavalry you need to expand. By the late middle age you should be able to have 8 luxuries via trade. Somewhen later you should change so that you have all 8 luxuries in your own territory. War waged with cavalry is the way to do that. This will increase the amount of your cities and hence the costs for importing luxuries. AI will demand higher prices from you, which is a good reason to take them with force.

Having more territory will also contribute much to a higher research output, at least in the long run. In other words: Your lacking research ability in the modern age is most likely rooted in the late middle age.

 

[WeirdoJoker]

You need to increase your research output. Obviously you need the commerce bonus from republic.

Republic itself was not a problem, as I made the slingshot & switch from Despotism right away. Trouble was that I could not increase my output without sacrificing order, and just to test I raised it to 100% and it still didn't get me to 4 turns, so I knew I had done something wrong somewhere along the line that I had not done before.

Your cities should have all essential buildings and preferably they are no longer mere cities but metropolises.

That's something I've been struggling with lately. In order to maximize research, I've been building my towns fairly close together, limiting maximum growth. I suppose if I'm going to do that I should probably convert them all to workers and abandon a few here & there?

Once you have cavalry you need to expand. By the late middle age you should be able to have 8 luxuries via trade. Somewhen later you should change so that you have all 8 luxuries in your own territory. War waged with cavalry is the way to do that. This will increase the amount of your cities and hence the costs for importing luxuries. AI will demand higher prices from you, which is a good reason to take them with force.

I couldn't get anywhere near that many luxuries at all ... it was a luxury-poor map. Resource-poor, too. It's a reason I'm considering no longer playing Tiny & Small maps and going for Standard and higher (never played Large or Huge before).

Having more territory will also contribute much to a higher research output, at least in the long run. In other words: Your lacking research ability in the modern age is most likely rooted in the late middle age.

In general, that makes sense. I always worry about making war when even on tech with the target civ ... I tend to lose battles more than win, so I'm not comfortable unless I have really big stacks going in, and I haven't figured out how to create those before the AI does.

My next game (going for the Spaceship with Byzantines/Standard/Continents/80%) will hopefully have more resources & luxes. I have generally done better at Monarch on Standard maps (3/3) than Small (0/4) or Tiny (2/4), which was never a problem at Regent or below.

 

[justanick]

That's something I've been struggling with lately. In order to maximize research, I've been building my towns fairly close together, limiting maximum growth. I suppose if I'm going to do that I should probably convert them all to workers and abandon a few here & there?

I would go for metropolises and therefore abandon lesser cities. Still, at Monarch the impact should not be that big. You can get reasonable results with cities of only size 10-12.

I couldn't get anywhere near that many luxuries at all ... it was a luxury-poor map. Resource-poor, too. It's a reason I'm considering no longer playing Tiny & Small maps and going for Standard and higher (never played Large or Huge before).

You should make sure that no less than 6 civs start on the map. Resources and luxuries should therefore be ample. You still need to take them, of course. They are supposed to be scarce enough to encourage fighting over them.

In general, that makes sense. I always worry about making war when even on tech with the target civ ... I tend to lose battles more than win, so I'm not comfortable unless I have really big stacks going in, and I haven't figured out how to create those before the AI does.

Learn to use artillery amd armies. Have 30 trebuchets(or better artillery) ready. Start a war. Let the enemy come to you. Use the artillery to reduce the hitpoints enemy units daring into your teritorry or adjacent to it. Attack and suffer very few losses. Your units gain experience. Eventually you will get a leader. Make an army. Give the army proper units. Go on the offensive and conquer.

It is of course much easier if you use military alliances. Share the burdon of war with everone else.

 

[CKS]

Is there an article (that I missed) discussing late-game research & strategies?

I don't think there is. What I think you need to do is research what the AI is not researching so that you can trade. While they may not want to give you satellites for money, they'll generally be willing to trade if you offer a tech. You may have to offer a tech plus money, or you may get money in return, depending on what tech you have available. I've generally had good luck going with the lower half of the tech tree (the laser, robotics) and getting the AI to research satellites and superconductor for me, but YMMV. There is somewhere a "What will the AI research next?" article.

I disagree with justanick about having metropolises. Researching sanitation is never worth it. If the AI research it for you, and give it to you (say, in return for you gifting them into the modern age), then maybe it is worth building some hospitals, but I've never found it worth rearranging my core and abandoning cities to grow some cities bigger. A city with Copernicus, Newton, or SETI is definitely worth growing if you have sanitation, though (or Shakespeare's Theater, which sometimes happens if I turn a 20k game into a space race).

When my research rate drops as I enter the modern age, I generally find that it is because I'm not using my corrupt towns as science farms. I'm just letting them grow and do whatever. This worked okay in the industrial age, where techs were cheaper, but fails miserably in the modern age. SETI and the Internet give my core a nice boost, but (aside from culture) the Internet does nothing to help out my towns in the boonies. So, what I'll do is go through all my corrupt towns and give them plenty of scientists. Usually, in addition to citizens making zero shields and beakers, I find a bunch of towns with civil engineers building libraries or harbors that won't do any good, and I switch these over. (I tend to build libraries for cultural expansion on the edge of my land, but once the Internet is in, I have research labs for that.) There is usually significant extra land on my borders to settle more science farms, and sometimes I take advantage of that, though I'm often too lazy to do so.

 

[justanick]

There is somewhere a "What will the AI research next?" article.

https://forums.civfanatics.com/threads/what-will-the-ai-research-next.45559/

I disagree with justanick about having metropolises. Researching sanitation is never worth it.

As i understand it there are 2 reasons for this stance.

1. Research is limited at 4 turns per tech, hospitals cannot change that.
2. If hospitals are worth it in the long run, they are still not worth its because the game ends too early for them to become a decisive advantage.

At Sid reason one becomes void. At below Sid it can be different. At Sid research is twice as expensive as at emperor and rank corruption is crippling high unless you build courthouses and police stations and use metropolises to get the best of of a limited amount of low corruption(say below 55%) cities.

When my research rate drops as I enter the modern age, I generally find that it is because I'm not using my corrupt towns as science farms.

At least in the long run its is worth to make even high corrupt cities metropolises. They need courthouses and police stations to reduce corruption below 70%, but it will give you a higher research output than else possible. Here however reason 2 is significant. Chances are that it will not pay out before researching future techs. So scientist farms are the more realistic choice.

 

[WeirdoJoker]

https://forums.civfanatics.com/threads/what-will-the-ai-research-next.45559/

I'm at that point where I've actually been trying to guess where the AI is at on this. Apparently - at least, up to Monarch - it will research everything, even the optional stuff, and that seems to be a major reason I get/maintain a tech lead. I get all kinds of weird trade offers, like maybe Democracy (which I never research) for Steam Power, or something along those lines.

 

[tjs282]

Apparently - at least, up to Monarch - it will research everything, even the optional stuff, and that seems to be a major reason I get/maintain a tech lead.

The Civ3-AI will (eventually) research every tech at all difficulty levels, if it has time to do so!

The main reason there's a difference in tech-pace between the levels basically boils down to its handicaps or bonuses (AI cost-factor), because that's what governs how quickly it can (build/capture new towns to) do or buy all that research.

And remember, (even while it might sometimes not feel like that !) the Civ3-AI is purely reactive: it doesn't plan ahead at all. It chooses its next research-project based on what prerequisite techs are currently known to it, and what the 3-4 potential projects will do in and of themselves. And all the AI-tribes on a map are using the same tech-valuation routine, so will generally follow roughly the same tech-priorities, with only minor/rare variations (based on e.g. UUs, and also AFAIK some degree of pRNG-influence).

The Civ3-AI generally 'prefers' to research techs which unlock new Wonders, buildings, units, governments and/or abilities, the more the merrier. Therefore, "empty" (or at least, emptier) techs represent the bottlenecks; so as a general rule, if you are at tech-parity with an AI-Civ, then researching the empty(est) 'next tech' as fast as you can, is more likely to give you trade-bait (if you want it) — or to open up a new monopoly-tech beeline that the AI won't get to until it's finished running through the shiniest techs on its current beeline.

That's a(nother) reason why the Industrial era is frequently the turning-point where the human draws level/ pulls ahead even in higher-level games. By ignoring optional Medieval techs (including the expensive Democracy-beeline), the human gets to the Industrial that much faster than the AI-tribes (all other factors being equal) — either while they're still finishing the Medieval optionals, or while they're still locked into the upper-branch 'Nationalism' upper-beeline.

That latter happens because Nationalism gives so many 'advantages' (new unit, new treaty-options, new governor-button), that the AI generally chooses it over Steam (new resource, new Worker-job), and always over Medicine (empty) — which will generally be the last of the 3 first-tier techs researched (exceptions generally only occur when Sci-Civs 'luck out' and get Med as their freebie, allowing them to trade for Steam and/or Nat).

And once an AI-Civ has Nat, it can then see Communism (new Gov! New building! New Specialist! Ooh, shiny!), and Fascism (new Gov!), so it will be more likely to go for those, before even 'considering' the rest of the tree.