This is embarrassing.

[jkk]

Okay. Just finished the game for which I sent screenies (I do not get large blocks of time to play, so a game takes me a while), ready to start a new one. While rereading this whole thread to refresh my memory, one question that seems to have multiple answers: is there a population level, in the very early game, at which one should switch to building settlers? If you built a new settler every time you hit pop 3, you'd never get much research done. My normal thinking would be, find a settler factory (ideally two) and let those build the bulk of them, let the capital grow a little more for research (because it is not subject to corruption). Is this incorrect?

 

[justanick]

is there a population level, in the very early game, at which one should switch to building settlers? If you built a new settler every time you hit pop 3, you'd never get much research done.

I rather disagree. Not letting cities reach size 3 by building settlers is the best way to go, unless corruption, lack of free land or specific qualities of certain towns overwrite the standard logic. The standard logic would be that founding new towns as early as possible helps to maximize growth and research, especially during anarchy. Getting enough luxuries for full growth during anarchy has some merits and luxuries create a bit of cap on the sensible city size.

There are of course those that prefer higher city sizes, which implies to build a granary early in your capital to make it a settler factory. A settler factory with granary will usually not fall below size 5.

 

[jkk]

I rather disagree. Not letting cities reach size 3 by building settlers is the best way to go, unless corruption, lack of free land or specific qualities of certain towns overwrite the standard logic. The standard logic would be that founding new towns as early as possible helps to maximize growth and research, especially during anarchy. Getting enough luxuries for full growth during anarchy has some merits and luxuries create a bit of cap on the sensible city size.

Well, I just tried that method, and I got to about 2700 BC with three cities and about 28 turns left to go on my initial research project. At 100% research. Is this to be expected?

 

[vorlon_mi]

I expected to see no temples, but you also do not recommend colosseums or cathedrals. So dissent is controlled entirely through the luxury rate, luxury distribution at marketplaces, and entertainers?

Since my earlier game is almost done, I'm just sort of going over the precepts and logic again here.

This is covered by some other posts, but I would add terrain factors again.

I like to play continents maps, and have frequently discovered that some luxuries are only found on the other land mass. Until I get ocean trade routes, I can't trade for them and can't easily conquer them. Thus, I may need to build some temples in the early going, to keep size 7-8-9 size cities productive. I usually set the lux rate at 10 or 20% in Republic. I like to choose a different tribe every game, so yes, if my civ can build temples at half price, I build them more frequently then too.

The list from justanick is great, and the implicit assumption is that you're not going for a cultural victory. In that case, one might prioritize temples and cathedrals a bit higher.

 

[Buttercup]

Yeah, I was going to post something similar to what Vorlon posted. While justanick posts a lot of good information he has a habit of talking in the definite when a lot of stuff in Civ3 depends on what type of world you create at the start, what civ you're playing with and what your end goal is.

In my case I play a lot of Archipelagos. Good luck securing all 8 Luxuries before you discover Navigation. Sometimes not even the Great Lighthouse can get you many. Further to this, if you're not exploiting the AI for everything they've got and playing a game where you prefer to leave the AI as starved as possible then you might want to choose to not trade for their Luxuries. In such cases you can only get your cities to a decent size with a lowish luxury rate by building Temples, Cathedrals and The Sistine Chapel and then possibly Colosseums and other Wonders, such as The Oracle, The Hanging Gardens and JS Bach's Cathedral.

Temples can also be a quick way to expand borders before you've learned Literature and a quick way to defend outlying cities that are close to another civ's Capital, or to get a Resource into your cultural boundary (as you can't Settle on Mountains or Wetlands). Also, they provide a massive boost to Culture wins as you can be producing 5p/t instead of 2p/t very quickly in every city in your core. Religious civs can get Temples up without barely noticing a break in any other production.

There's way too many variables to make any specific definite claim about how useless Temples are and while the odd sentence of this post might cause a nit-pik here or there you should be able to get the general idea of the post.

Similarly, how big you want towns before making a Settler is going to depend a lot on what difficulty level you're playing at as constantly dropping form 3 to 1 size with lots of cities will normally net you slower research than segregating Settlers to specific high food.production towns while leaving the rest to accumulate normal Science and Population. On the higher difficulty levels you wont be getting good Science whatever you do, so building more towns is always preferable. Etc etc etc.

Also, securing your 8th Luxury gives you a Palace Upgrade, if you take pleasure in getting Palace Upgrades

 

[tjs282]

Well, I just tried that method, and I got to about 2700 BC with three cities and about 28 turns left to go on my initial research project. At 100% research. Is this to be expected?

Why were you running 100% SCI%? If you're at 2700 BC, that means Turn 26(ish). 26+28 = 54T, so turning up SCI% to 100% wasn't going to get you the tech significantly faster than the default max.-50T-to-tech cap. In the early game, if it's going to take me more than ~40T to get a tech at SCI%=90-100%, then I would rather run 10% SCI% (or a lone Scientist) and get it in 50T, while also banking gold -- gold which I can then use as a top-up (if necessary) for buying the AICivs' starting techs, or to pay them off if they demand tribute before I'm ready to fight them. If I haven't got Alphabet as one of my starters (i.e. my Civ isn't COMM or SEA), I might even sit at SCI%=0% until I can trade for it (or pop it), and then start a 50T-run on Writing.

Have you read this article already? All the AICivs follow the same basic programming when it comes to tech-research choices, and they tend to research tier-by-tier, rather than beelining through the tree -- so their first project will nearly always be the most 'valuable' 1st-tier tech(s) that they don't have, e.g. BronzeWorking or Masonry, at least until they meet each other (or you) and can start trading. The AICivs will tend to trade all their 1st-tier techs between themselves ASAP (especially at higher levels, when they give each other discounts), and AFAIK they will also tend to ask for the tech they're currently researching.

Once they've got Bronze, you can be nearly certain (9 times out of 10) that they will (all) go for IronWorking next, regardless of whether or not they have Iron in their vicinity (despite them knowing where all the 'hidden' StratRes are). In the time it takes them to complete those two research projects (they will also be running SCI%=100%, even if it will take them 50T), hopefully you've been able to get Alph and finish a 50T run on Writing -- which you can then trade for any 1st-tier techs you don't know yet, or (quite often) a single 2nd-tier tech. At Regent level, the CoL+Philo-->Republic slingshot is almost almost feasible, and even at Emp it's still a high probability of success (though you might want to hold off a little longer on trading Writing).

Re. Temples and other Culture-buildings:

I usually play long high-tech games, so I rarely build Temples for Culture unless REL, and almost never for happiness unless I'm severely Lux-starved in the early stages (if then, I'll sell them off once I have reliable overseas Lux-trade-routes and Markets). I've only once gone for a Culture-vic (as the Babs in unpatched Vanilla, at Regent), and I failed -- but won a Space-vic instead. That said, (I've heard that) whipping Temples can be a useful tactic for a Culture-game, especially as a REL-civ: you only need 10s in the box (which you could get from a Forest-chop), and Pop2 in the town. The new Temple then quells the whip-unhappiness, and (assuming the Temple was the town's first building) it will take somewhere between 10T (if food-bonuses, and/or AGRI+freshwater) and 20T until the town gets back to Pop3, by which time your citizens will have (nearly) forgotten your cruel oppression. Also, Culture-points-per-turn double after buildings have existed for a certain period (1000 years?), and as a REL-Civ (starting with CBurial), you can build Temples from the very beginning of the game...

 

[justanick]

Well, I just tried that method, and I got to about 2700 BC with three cities and about 28 turns left to go on my initial research project. At 100% research. Is this to be expected?

Huge maps suffer from twice the research costs than small maps, so i slightly lack evidence about what is to be expected. But that it takes a while to increase research output is to be expected. All in all larger maps favour the settler before granary approach as corruption will be low while research costs will be high.

In my case I play a lot of Archipelagos. Good luck securing all 8 Luxuries before you discover Navigation.

There is no necessity for that. Before you have unis and possibly banks more or less nationwide the upkeep for temples ofsets the content face. Also before you have hospitals securing more than 6 luxuries is a waste as 6 luxuries do give you 12 happy faces when you only need up to 11 of them.

Also, Culture-points-per-turn double after buildings have existed for a certain period (1000 years?)

1000 years it is.

 

[jkk]

Huge maps suffer from twice the research costs than small maps, so i slightly lack evidence about what is to be expected. But that it takes a while to increase research output is to be expected. All in all larger maps favour the settler before granary approach as corruption will be low while research costs will be high.

Ah. The logic, then, is that one ends up with a whole lot of trade much sooner, due to more cities improved by more workers. A breakneck expansion to the potential limits of same depending on geography, followed by letting all those cities start growing up?

Why were you running 100% SCI%? If you're at 2700 BC, that means Turn 26(ish). 26+28 = 54T, so turning up SCI% to 100% wasn't going to get you the tech significantly faster than the default max.-50T-to-tech cap. In the early game, if it's going to take me more than ~40T to get a tech at SCI%=90-100%, then I would rather run 10% SCI% (or a lone Scientist) and get it in 50T, while also banking gold -- gold which I can then use as a top-up (if necessary) for buying the AICivs' starting techs, or to pay them off if they demand tribute before I'm ready to fight them. If I haven't got Alphabet as one of my starters (i.e. my Civ isn't COMM or SEA), I might even sit at SCI%=0% until I can trade for it (or pop it), and then start a 50T-run on Writing.

I was running 100% Sci because I was pedal-to-the-metal on the Philosophy slingshot toward Republic, plus early Library builds. It was the very first lesson that was pounded into me, after all. That said, I'll start experimenting with the sliders.

I also do get what people are saying: that circumstances vary a lot with map size and other concerns. In my view that's one of the great things about the game: we can each create a fresh gameplay opportunity that suits our preferences, and can vary those. So I guess I should specify that I'm a creature of habit who likes to play with max random civs (I think that totals eighteen), 70% water (so I usually get two big continents, two small, and a couple of dinko islands), average geologic age, and a huge map. I don't use any victory conditions other than Conquest--for me, the fun of building is to build up and employ a formidable fighting force, remembering and punishing past slights while planning campaigns of conquests.

 

[justanick]

Ah. The logic, then, is that one ends up with a whole lot of trade much sooner, due to more cities improved by more workers. A breakneck expansion to the potential limits of same depending on geography, followed by letting all those cities start growing up?

I would not put it in quite these word, but yes, that is it. The growing of cities happens ideally after leaving despotism. The expasion should not be to neckbreaking either, it may need to be backed up with military. The germans are fine for that. They start with both military techs and they can build cheap barracks. You can expand aggressively, then idle your expansion for very long untill you get Panzers which are then spammed out via mobilization.

I was running 100% Sci because I was pedal-to-the-metal on the Philosophy slingshot toward Republic, plus early Library builds. It was the very first lesson that was pounded into me, after all. That said, I'll start experimenting with the sliders.

Imo there is little to experiment. Writing is the one tech were 10% or lone scientist is reasonable. As you are playing on huge map there probably is no alternative to it either, unless you donnot start your research with Writing.

 

[Buttercup]

On Huge maps I often encounter a glitch with regards to Writing. If I start out researching Writing, because I start with a civ with Alphabet, then no matter how much Science I produce it will remain at 50 turns to completion (de-accumulative) no matter how much more Science I contribute. For example, I could start out with 4 Science and it'll say 50 turns. 20 turns later I'll be contributing 13 to Science and it'll still say 30 turns left.

I've no idea if anyone else has this nor why it happens in my games, but I'll usually learn either Bronze Working or Masonry first and then move to Writing which will then say something like 27 turns left (effectively making it so you might as well get Bronze Working or Masonry included for your 50 turns). That's how I work the 100% Science method anyway.

This, obviously, would be different at the very higher levels when they all cost more anyway, I'm referencing Regent/Monarch here.

I also like to get Bronze Working or Masonry anyway as I like to build Colossus and like to have the chance to build Pyramids (or have something worthwhile in the bank should I get an early SGL).

 

[Buttercup]

So I guess I should specify that I'm a creature of habit who likes to play with max random civs (I think that totals eighteen), 70% water (so I usually get two big continents, two small, and a couple of dinko islands), average geologic age, and a huge map. I don't use any victory conditions other than Conquest--for me, the fun of building is to build up and employ a formidable fighting force, remembering and punishing past slights while planning campaigns of conquests.

Ah, right. You should probably start a new thread specifying this as your criteria as I suspect as the thread grows to 20 pages many people wont be reading page 7 and you'll get a lot of stuff you're not going to be concerned with.

 

[tjs282]

On Huge maps I often encounter a glitch with regards to Writing. If I start out researching Writing, because I start with a civ with Alphabet, then no matter how much Science I produce it will remain at 50 turns to completion (de-accumulative) no matter how much more Science I contribute. For example, I could start out with 4 Science and it'll say 50 turns. 20 turns later I'll be contributing 13 to Science and it'll still say 30 turns left.

This is not a glitch: this is the way the game is programmed. I'm surprised that you don't (seem to) know this already: if a tech will take >50T to research -- even at 100% SCI% -- the turns-to-tech will always cap off at 50T (the min./max. turns-to-tech is settable in the Editor). However, a tech-cost is always a fixed number of beakers (initially based on base-cost and map-size and difficulty level*, with reductions based on how many known AICivs already have the tech).

*EDIT

Techs are vastly more expensive on Huge maps than they are on Tiny -- and later techs in an era are always more expensive (i.e. higher base-cost) than 1st-tier techs. So on Huge Maps, even the 2nd-tier Ancient Age techs may already exceed the 50T-cap, even at SCI%=100%, because regardless of map-size, you still need X number of turns to build each Settler, but you'll need to found (many) more towns before you'll be able to start raking in the commerce (and hence being able to do the necessary research) significantly faster. So if you can't do a tech significantly faster than 50T, you might as well just run SCI%=10%, and take the TAX%-gold instead.

But if you do decide to do that, then after a certain point (which will vary depending on the tech-cost, but will become progressively earlier in the 50T-run as you advance through the tree/eras), you'll have to stick to it, because you will have passed the tipping point where putting in more beakers-per turn -- even at SCI%=100% -- won't be enough to make up for the 'beaker-deficit' you accumulated over the preceding turns, so the total-time-to-tech will still be capped off at 50T.

(And just to add insult, the meagre amount of beakers accumulated -- relative to the beakers needed on a Huge map -- likely won't be enough to persuade the AI to give you any significant discount on that tech, either).

This, obviously, would be different at the very higher levels when they all cost more anyway, I'm referencing Regent/Monarch here.

Not exactly true, as I said above. Map-size is the critical factor affecting tech-costs, not difficulty. But for any given map-size, the human player actually pays the same tech-cost at all difficulties, just like they need the same number of shields to build any given unit at all difficulties. Relatively speaking, yes, at difficulties above Regent, the techs are more expensive for the human than for the AI. But what actually happens is that beakers-to-tech is lower for the AI (just like their food-to-grow and shields-to-build requirements are lower), and they give each other bigger discounts on trades, allowing them to move through the tree faster, at least once they've got g(r)o(w)ing -- which is also easier for them at higher diffs, because of the aforementioned lower food- and shield-costs.

 

[Buttercup]

Well done for explaining all that, but I know all that, you simply don't grasp what I'm saying. I don't know how I could word my post differently to make it clearer for you, perhaps someone else might understand it better and do the explaining better.

As regards your second point, I began the post by stating "on huge maps", so why you then go on about what you did is beyond me.

 

[justanick]

Not exactly true, as I said above. Map-size is the critical factor affecting tech-costs, not difficulty. But for any given map-size, the human player actually pays the same tech-cost at all difficulties, just like they need the same number of shields to build any given unit at all difficulties. Relatively speaking, yes, at difficulties above Regent, the techs are more expensive for the human than for the AI. But what actually happens is that beakers-to-tech is lower for the AI (just like their food-to-grow and shields-to-build requirements are lower), and they give each other bigger discounts on trades, allowing them to move through the tree faster, at least once they've got g(r)o(w)ing -- which is also easier for them at higher diffs, because of the aforementioned lower food- and shield-costs.

This you got seriously wrong. The beakers AI has to pay do not depend on difficulty setting, they are always as on regent.

The costs for the human player do vary as [base costs] x [map modifier] / min(10;[AI cost factor]). So at Sid you pay double the beakers compared to emperor.

http://forums.civfanatics.com/showthread.php?t=29485

 

[tjs282]

This you got seriously wrong. The beakers AI has to pay do not depend on difficulty setting, they are always as on regent.

The costs for the human player do vary as [base costs] x [map modifier] / min(10;[AI cost factor]). So at Sid you pay double the beakers compared to emperor.

http://forums.civfanatics.com/showthread.php?t=29485

OK, I stand corrected. (Although that formula is for Vanilla 1.29,) seems like it still applies to Conquests (at least, a cursory search didn't turn up anything more recent). Struck out the mistakes.

Buttercup:

A worked example for you, using your own numbers:

Spoiler :

Base cost of BronzeWorking is 3. Base-cost of Writing is 8. Map modifier for Huge is 400 (from the thread Justanick linked to). If I understood (Justanick's rewrite of) the formula right, then that means that while Writing is still unknown (to anyone that you know), for a Regent-level Huge map, it costs 3200 beakers.

Per your example, at 4 BPT, Writing would take 800T to research, so time-to-tech is capped off at 50T: but you'll only amass 200 beakers in 50T, so the deficit to actual cost = 3000 beakers. Even if (again per your example) you can increase your science-investment to a (whopping) 13 BPT over the next 20T -- let's split the difference and call it 9BPT over 20T, and that's being generous -- you've still only amassed ~180 beakers in total, so the deficit to actual cost is still ~3020 beakers, divided by 13 BPT ≈ 233T to tech! So the time-to-tech will still be capped at 50T, minus the 20T you already put in --> still 30T to tech.

By my count, you would have to octuple your BPT (i.e. from 13 to 105 BPT) just to get the remaining time-to-tech down by 1T below that 30T-remaining (29 T * 105 BPT = 3045 beakers). And that's 105 BPT average over those 29T, i.e. your BPT would have to have increased to much more than that by the 29th turn. Which would mean you're gonna need towns -- lots of towns... So how many Settlers do you think you'll be able to build in those 29T (while still under Despotism to boot)...?

Of course, meeting some AICivs who've already researched Writing would help reduce the absolute beakers required. Only problem is, at Regent, the AICivs'll be operating under the same constraints as you -- and thanks to your preference for Huge (80%-water?) Arch-maps, none of them will be able to share notes with each other (or you) for a long time yet, either. And they'll most likely all be busily researching Bronze/ Masonry/ Iron anyway...

OTOH, if you've spent time researching Bronze/ Masonry first, then by the time you start on Writing, your BPT is probably already (much) higher than 13 BPT. (I don't think you'd be able to do either of those in <50T: you'd need 1200 beakers for Bronze, 1600 for Masonry). Safe in the knowledge that unhappiness is not a major issue at Regent (and the expectation that you won't be sharing your island?), I would imagine that you've probably been building almost nothing but Settlers and Workers (plus maybe a boat or two?) and ReXing like mad. For Writing in 27T from scratch, assuming no-one knows it yet, you'd need 3200 / 27 = 119 BPT -- which is pretty high, but at SCI%=100%, you could get that from working 60 tiles, each giving 2CPT (before corruption, but also not including any CPT-bonuses for Luxes, or Colossus), e.g. ~12 fully roaded Pop4 river- or coastal-towns.

So there's no mystery, no glitch -- and you're right, I don't understand why you're claiming that there is...

 

[Buttercup]

Those stats are insane and bare no reflection whatsoever to reality. No idea what you're doing there.

Starting on Bronze Working, for example, it usually asks for about 25 turns from the start (maybe 20 on good turf). By the time I have BW I have about 13 Beakers coming in and when I switch to Writing it will be around 27 turns. Usually, and I've repeated this in practically every game I've ever played on Huge maps over the past few years, I get both Bronze Working or Masonry and Writing within 50 turns.

It's great that you have an equation to sperg at me, however, this is how my games play out. Sorry about that. I'll do some screenies for you if you like, though I don't know why you can't test it yourself before sperging rather than just sperg random stats, it would only take a half hour or so.

 

[justanick]

Base cost of BronzeWorking is 3. Base-cost of Writing is 8. Map modifier for Huge is 400 (from the thread Justanick linked to). If I understood (Justanick's rewrite of) the formula right, then that means that while Writing is still unknown (to anyone that you know), for a Regent-level Huge map, it costs 3200 beakers.

At regent AI cost factor is 10, therefore the tech costs only 320 beakers. 3200 beakers would be the costs for AI cost factor 1 at Beyond Sid. There getting below 50 turns per tech is sort of an exception. Maybe we should rather limit us to cases of the normal degree of insanity.

 

[Buttercup]

Playing as the English, Huge map, 11 opponents, Regent, Raging Barbs, Archipelago and etc etc etc:

Game A - 100% Science learning Writing off the bat:



1 Turn to Writing:



F8 screen for turn time:



I've only 'gained' 8 turns from 50 - by this method, yeah, you might as well have gone 50 turns at max cash... However... see next post (I think there's a screenie limit per post?)

 

[Buttercup]

Game B - going 100% Science for Bronze Working first:



1 turn to Bronze Working:



F8 screen for Bronze Working turn time:



So I still have 29 turns in reserve and I can now build Colossus instead of having nothing interesting to build under 50 turns of Writing (of course you have options, but more of that later).

Onto next post for conclusion to Game B.

 

[Buttercup]

Game B continued - approaching Writing:

Oh look, I've built the Colossus already, not even possible with 50 turns of Writing:



And got a Golden Age to boot:



And here comes Writing:



At turn 49:



And I'm now good for Code of Laws in 16 or Masonry in 7, with a Golden Age making The Pyramids a good possibility before I even reach Republic. I can also start producing my Settlers now to no great and dramatic loss of space, I just generally have far more options at this stage than simply a bit more gold: