Understanding Power: An open discussion of immortal strategy

[Jorunkun]

Over a million people play Civilisation IV. No more than a few hundred can win on immortal. If you are one of them, or if you think you have what it takes to try, this thread is for you.

It’s titled “understanding power” because immortal level play is all about power: If you don’t learn to seize it fast, the AI surely will. Our objective shall be to develop our understanding of general strategy to the point where we can win at least half of all immortal level games under randomised standard settings. (1)

This means we are not interested in step-by-step guides for beating immortal with one particular civ or other specialised strategies founded on narrowly defined premises. We are trying to master the game at large.

With your participation, I hope to make this thread home to a qualified, ongoing discussion among equals. I will do my best to structure our progress and post daily to keep things going, but anyone is invited to ask relevant questions and share insights.

You do not have to be crack at immortal to qualify to post – in fact, I have yet to win an immortal game myself and have far more questions than answers. The only thing I ask is that all contributions be based on a resonable amount of high-level (emperor and above) play experience.

Sound like you? Then welcome aboard.

J.

(1) Randomised standard settings means: Standard gameplay. All victory types enabled. Standard size shuffle maps, random climate and sea-level. All random civs. Immortal difficulty. No restarting, no reloading. We will be using vanilla civ and epic speed as the basic frame of reference. Please indicate if you are using different speeds or Warlords.

 

[Jorunkun]

*** Last update: 29 August, V1***

Summary
Under immortal settings, the player suffers slightly increased handicaps to research and maintenance costs as compared to emperor. However, the AI benefits from significant boni to growth, production and military upkeep. Expect the AI to spread and grow noticeably faster and its armies to be larger than on emperor.

While the AI is not explicitly programmed to try different tactics or be more aggressive in war, the bonuses to its military production and upkeep translate into bigger differences in army size which in turn increase the likelihood of the AI going to war.

Also, immortal barbarians begin to appear in 3550 BC, 5 turns earlier than on emperor, and about one third of goodie huts will spawn barbarians.

Details: Under immortal settings …

AI freebies

… the AI gets hunting, archery and agriculture as free starting techs, in addition to civ-specific techs.
… the AI starts with three defenders (archers), two workers (one more than emperor), and one scout.

Player bonuses and handicaps

… health (+1) and happiness (+3) bonuses, civic upkeep and inflation cost are the same as on emperor and deity.
… the player needs to produce 125% of beakers required to research a tech (emperor 120%).
… the player pays increased city upkeep and unit support cost (details see here).

AI bonuses

… the AI needs only 80% of food to grow a city pop point (emperor: 85%)
… the AI needs only 80% of hammers for units and buildings (emperor: 85%)
… the AI suffers only 30% of inflation cost (emperor: 40%)
… the AI suffers only 30% of the war weariness penalty (emperor: 40%)
… the AI pays only 10% of unit upgrade cost (emperor: 15%)
… the AI gets a 75% bonus on its workers’ work rate (emperor: 50%)
… the AI gets the same bonuses vs animals (+70%) and barbs (+40%) as on emperor

Huts and barbs

… animals spawn rate is 1 per 25 unowned tiles (emperor: 30)
… about one third of goodie huts will spawn barbs – it is more likely to spawn barbarians than to get gold
… barbarians start appearing after 15 turns in 3550 BC (emperor: 20) and found cities after 20 turns (emperor: 25)
… barbs spawn rate is 1 per 30 unowned tiles (emperor: 35)
… barb city spawn rate is 1 per 90 unowned tiles (emperor: 100) and barbs are more likely to found cities than on emperor
… barb cities will usually have 4 defenders vs 3 on emperor

Source: http://forums.civfanatics.com/showthread.php?t=158130
CIV4HandicapInfo.xml in spreadsheet format, by NeverMind.


Rough timeline for AI behaviour
Based on observations, epic speed.

Early religions
3640/3670 (turn 12/13) - early Buddhism (only if civ with mysticism is among AI opponents)
3340 (turn 20) - average Buddhism, +/- 2 turns
3100 (turn 30) - Hinduism average, +/- 3 turns
2200 (turn 55) - Judaism, +/- 5 turns

2nd AI city
3190 to 2890 (turns 25 - 35) - AIs found their 2nd cities. With settler first and some luck, you may beat the later ones to it.
2890 to 2590 (turns 35 - 45) - AIs found their 3rd cities. This is when player can get a first settler out with warrior/settler start.

Early Wonders
1060 - Stonehenge, +/- 5 turns
900s - Oracle

Corrections and further contributions are welcome and will be credited.

 

[Jorunkun]

General advice for playing immortal
Version 1

This is a summary of general, high-level advice for playing immortal games under the settings outlined in the opening post of this thread. It is directed at players who have mastered emperor and want to test their skills on immortal.

The list is a work-in-progress, the result of the ongoing discussion in this thread. If you disagree with any of the points, have something to add or a question to ask, jump right in and help improve the next version.


1. Leverage your strengths, exploit AI weaknesses
Unlike the lower levels, immortal games do not allow for much choice in playing styles and strategy. Play the cards you are dealt: Leverage any advantage the lay of the land and your traits give you. As long as you are not a major power, anticipate and respond to the strategic direction the AIs are taking.

2. Focus and efficiency
Immortal AI boni to research, production and upkeep ensure that you start at a disadvantage. As long as your empire is smaller than theirs, the gap will widen with every turn. Play efficiently, make extensive use of chopping and whipping to accelerate expansion and conquest.

3. Expand while you can
Scout and expand quickly to secure space for at least three cities. If you start close to an AI, on a peninsula or with the opportunity to seal off a backfill area, consider settler first or warrior/settler starts. Accelerate your first settlers with chopping and rushing. Extensive investment in improving your capital before founding a second city only pays off on island starts.

4. Attack early and deceisively
Early military rushes are the high road to power. Beeline for copper and/or horses and hit the AI while it is still weak. If you have a military advantage, press it as fast and hard as possible. Beware of chokes, worker steals and super early rushes though, as the AI starts with three archers. Later in the game, drive and exploit inter-AI wars.

5. Deferential diplomacy
Do everything in your power to stay on good terms with the AI, especially your near neighbours and in the early game. Open borders, trades or (temporary) gifts of unused ressources, giving in to requests for gold or tech and converting to the majority religion of near neighbours all help to reduce the risk of surprise attacks, the most frequent causes of losing a game. Pamper your friends and create joint enemies.

6. Research for trade, not for need
You start out behind in tech and research and can expect to stay there for at least the first half of the game (renaissance period). Research and lightbulb to tradeable tech deeper in the tree that the AI doesn’t prioritise, such as alphabet, literature, philosophy and astronomy. Don’t be shy of trading the alphabet early; unlike on emperor you won’t have a monopoly on it for long.

7. Don’t count on grabbing early religions and wonders
Assume you won't get any of the early religions unless you start with mysticism and are prepared to stick it out until Judaism if necessary. The Oracle tends to get built around the same time as Stonehenge (around 1000BC) and chances at getting the Pyramids are very remote. However, wonders which require an existing city improvement such as the the Colossus, Great Library, and Hanging Gardens are possible if you have the appropriate resource.

Note: Will make changes and add more points about the middle and late game as the discussion progresses.


Demogames

http://forums.civfanatics.com/showthread.php?t=183894
GOTM 10 pre-game discussion for Quin’s chinese on a low-water snakey archipelago. Contains some good posts on general immortal strategy.

http://forums.civfanatics.com/showthread.php?t=178607
The first public immortal sg on this forum, afaik. Starring Mutineer, uberfish, ThERat and Qwack as Saladin on a fractal map. Pretty impressive, as they win even without ever using cottages.

http://forums.civfanatics.com/showthread.php?t=147428
Epic Immortality SG. Shillen, ThERat, Arathorn, Kylerean, and Islandia take Victoria to a diplomatic win.

http://forums.civfanatics.com/showthread.php?t=171130
Godotnuts guide to a deity cultural win. The stars have to align pretty well in terms of maptype and civ used for this to work, but if they do, this is how to go about it.



Suggested reading

http://forums.civfanatics.com/showthread.php?t=141475
Unit upkeep explained, in great detail, by Roland Johansen and colony.

http://forums.civfanatics.com/showthread.php?t=159109
Zombie69s guide to micromanagement. A wealth of information, essential for efficient play.

Important: Need a link to a concise, current summary of whipping, including effects granaries, forges and civics. Please mail me or post recommendations.

Corrections and further contributions are welcome and will be credited.

 

[Jorunkun]

As a starting point, I suggest we look at the key differences between immortal and emperor level games, and focus on tactics to get off to a good start in the early game.

The majority of us is probably in this transition stage now (or will remember it from not too long ago), so please tell us:

- What struck you as the biggest differences in playing immortal, compared to emperor?
- What are your top issues, and how do you propose to address them?
- What questions are foremost on your mind?

Here’s my top three, feel free to comment and add your own.

1. The AI spreads faster and more aggressively
Founding a second city early is always a priority in civ, but on immortal, it’s all you can do to survive until the classical age. Except on archipelago / island starts, I usually count myself lucky to found more than three cities before I get crammed in.
Question: On average, how many turns pass until the AIs plonk down their second and third cities?

Also, I noticed the AI settling towards you more aggressively, often going out of its way to cut you off.
Questions: Is it feasible to take out these AI settlers, say with an early UU that allows for this?

What this tells me is that settler first and choking - which I consider unnecessary risks in most emperor games - become pretty solid starting options on immortal. It also tells me you need to get your strategic ressources visible and covered as fast as possible, in order for an early rush.
Questions: Any comments on settler first? What about chokes – are they doable, or even necessary?

2. You need to leverage any advantage you have, especially militarily
My experience is that in emperor, you get to pick your chances for war and can often afford to pass up a risky move in favor of better odds later. On immortal, it seems that you need to push any advantage, especially in early military rushes. In fact, attacking a neighbour slightly ahead in military strength might be your best shot in the early game, as he will turn into a neighbour far ahead in military strength in no time.
Any comments, experiences, reports of games won on Immortal without an early (pre LB) rush?
Also, I noticed aggressive AIs tend to run up much larger armies and routinely guard core cities with eight or more units. Learning: Stacks need to be much bigger than on emperor. Comments?

3. You will fall behind on tech, but can often catch up periodically
On emperor, I have learned to make getting to and maintaining a tech lead my top overall strategic priority for most of the game. On immortal, that same relentless focus will keep me slightly behind the AI … if I’m lucky.
Question. Any games where you took and kept the lead on immortal? How did you do it?

What I learned though is to not worry so much. You can often lightbulb to a tech the AIs are missing and trade yourself out of the hole at least for a while. Also, you will have to learn waging war against superior units.
Question: What are key techs for trade with the AI – different from emperor?

As a last, general question, I’d be curious the hear if you think our stated goal of winning at least half of is even achievable. I do manage to win far more than half of my games on emperor now; something I would have thought impossible when I first moved up, so I guess I am expecting the learning curve to extend on that tangent. Realistic?

Right, your turn now.

J.

 

[MrCynical]

Well I've beaten Immortal a couple of times under the above conditions, but I've also lost plenty of games as well. I'll pitch in what I've found out though.

The biggest difference I found in the transition was that I really had no choice but to be aggressive. At Emperor I could colonize and culture enough of an empire to catch up in the later game without warfare, whereas at Immortal I couldn't get close. Barring unusual terrain I couln't found more than two or three cities (and that's with chop rushing like mad) before I was completely hemmed in by the AI.

As a result I assume I'll be going to war early, hopefully with mostly axemen if I have copper. If I've somehow avoided war up to that point, I'll go for my weakest neighbour as soon as I get catapults. I'll also look to have at least one reliable ally nearby to assist, even even I have to give then some serious cash or resources to keep them on my side.

Question: On average, how many turns pass until the AIs plonk down their second and third cities?

Not many at all. The third AI city is usually in place before my second. I suppose I could dredge up some of my old replays and get a more specific answer.

Questions: Is it feasible to take out these AI settlers, say with an early UU that allows for this?

In the case of the first few, definitely not even with a UU. They'll be escorted by at least one archer and you'll be wasting a lot of effort trying to take them out with warriors. Once you've got axemen it might be worth trying, especially as you're likely to end up at war with at leas tone neighbour very rapidly anyway.

Questions: Any comments on settler first? What about chokes – are they doable, or even necessary?

Just to clarify, you're not suggesting building a settler first are you? That really doesn't seem worthwhile. I work on the assumption that I'll have to hack out land by conquest, not settling, so I won't cripple my capital trying to out settle the AI. Chokes are rather map dependent, but can be well worth it.

Any comments, experiences, reports of games won on Immortal without an early (pre LB) rush?
Also, I noticed aggressive AIs tend to run up much larger armies and routinely guard core cities with eight or more units. Learning: Stacks need to be much bigger than on emperor. Comments?

I haven't won except on an archipelago map at immortal without going to war pre longbowmen, though the wars were usually not started by me. While it's true you often get stacks of 6 or 7 units in major AI cities (and 15+ in the capital is not uncommon), the border cities can often be very weakly defended. Also, except for the capital, the AI treats these stacks as a mobile attacking force unless it's losing, so a tactic I've found to be effective is to start a war between the AI and a third party, they pull all these stacks out to one border, and I have several cities with just two or three defenders to deal with. They then can't get these forces back anywhere near as fast, or they all turtle up in the capital (which is then a job for vast numbers of catapults).

Question. Any games where you took and kept the lead on immortal? How did you do it?

Not prior to about the late Renaissance. From the start of the industrial age onwards AI research often stagnates due to their much poorer use of cottages and specialists, allowing me to go from a few techs behind to a solid lead quite rapidly. Before that though the only way I end up in tech lead is to get a monopoly tech by careful use of a GP and trade it around. I can then sometimes hang on to my next research tech as a lead rather than having to trade it away.

Question: What are key techs for trade with the AI – different from emperor?

Not all that much difference really. Alphabet is still very important due to the AI's reluctance to research it. Literature is also neglected by them (except for one or two wonder-orientated leaders). Theology or Music are easy to get with GPs, and odds are no more than one other AI will have them as they don't bother with them once the religion/great artist has gone. The AI is still reluctant to trade them amongst themselves though. Astronomy is the only other obvious tech that I always seem to beat the AI to, though by direct research rather than a GP.

As a last, general question, I’d be curious the hear if you think our stated goal of winning at least half of is even achievable. I do manage to win far more than half of my games on emperor now; something I would have thought impossible when I first moved up, so I guess I am expecting the learning curve to extend on that tangent. Realistic?

I think it's possible, and I can also reach this at Emperor level. I've only got about 2 out of 10 at Immortal so far though.

 

[yavoon]

honestly I only play immortal w/ financial. though I'm sure better ppl can beat it w/ other civs(and maybe me, haven't tried in a while).

is the map continents? continents immortal is brutal.

 

[PMabey]

firstly i think its great your using random settings. i hate strategies using specific leaders etc.. i am playing on emperor atm and winning more than i lose. I think the best leader would be one with a warring trait and financial

 

[uberfish]

- What struck you as the biggest differences in playing immortal, compared to emperor?

Faster AI tech rate, faster AI early expansion, more defensive troops in both AI and barb cities.

Also, I noticed the AI settling towards you more aggressively, often going out of its way to cut you off.

Warrior first, then settler if someone is close, becomes a better opening the higher level you go for this reason. I don't like naked settler-first, you can't get any scouting done that way as your initial warrior has to hang around the starting area.

Questions: Is it feasible to take out these AI settlers, say with an early UU that allows for this?

Only quechua are fast enough.

Any comments, experiences, reports of games won on Immortal without an early (pre LB) rush?

Yeah, you can go with mace/cat or supported treb stacks if you had a start more suited to early building. If you can start a war between your intended target and another AI so that it expends its mobile stacks, it makes invasions a lot easier.

You will fall behind on tech, but can often catch up periodically

You start behind in tech. Vertical research and tech trading is usually your friend until you can get a strong economy up and running, usually taking out 1 AI and developing their lands well having proper city specialization is enough to take the lead though you might not pull ahead till Industrial age.

Question: What are key techs for trade with the AI – different from emperor?

Alpha will get you IW+mathematics reliably. In Medieval it's hard to do better than philosophy off a GS lightbulb, it's the largest beaker value tech you can bulb directly and is also pretty useless to the AI.

As a last, general question, I’d be curious the hear if you think our stated goal of winning at least half of is even achievable.

Yes, I think 80%+ of starts should be doable, but I'm not quite up to that level yet.

 

[yavoon]

Faster AI tech rate, faster AI early expansion, more defensive troops in both AI and barb cities.



Warrior first, then settler if someone is close, becomes a better opening the higher level you go for this reason. I don't like naked settler-first, you can't get any scouting done that way as your initial warrior has to hang around the starting area.



Only quechua are fast enough.



Yeah, you can go with mace/cat or supported treb stacks if you had a start more suited to early building. If you can start a war between your intended target and another AI so that it expends its mobile stacks, it makes invasions a lot easier.



You start behind in tech. Vertical research and tech trading is usually your friend until you can get a strong economy up and running, usually taking out 1 AI and developing their lands well having proper city specialization is enough to take the lead though you might not pull ahead till Industrial age.



Alpha will get you IW+mathematics reliably. In Medieval it's hard to do better than philosophy off a GS lightbulb, it's the largest beaker value tech you can bulb directly and is also pretty useless to the AI.



Yes, I think 80%+ of starts should be doable, but I'm not quite up to that level yet.

for what purpose do u need a warrior near ur capital early in the game? I'm not sure I'd ever try to become a builder at immortal, u need to start lickity split on conquering territory, maybe if it was marathon I'd feel more comfortable building, but I dont play marathon.

and I dont think more defensive troops is correct. I wasn't sure so I checked my last immortal game. the AI has 1-3 defensive troops in all its cities(I used worldbuilder), the only exceptions are ostensibly the cities where the AI stations its army and I think a few cities w/ 4. barb cities on immortal have 3 archers, and on emperor its either 3 or 2. either way immortal AI's can certainly make more troops. but I dont think the AI is actually programmed to garrison more troops. sometimes u will run into a situation where the AI decides to build a lot of defensive troops, I ran into genghis once on emperor where he had 12 archers in his capital because he lacked any resource.

 

[douglas_21]

I'm on monarch now and have found that taking cities is very difficult. The AI even suicided a Barrage II catapult into my city raiding army/stack. However, when I spread out my units the AI picks them off individually. If I have a pikeman by itself it is crushed by a maceman or crossbow. If I leave my knights alone, pikemen attack them. When I seperate my city attackers into multiple stacks, I usually need multiple medic/stack defense units that don't fight. And doing that for 2-4 stacks gets very expensive production-wise. I can't reinforce my attackers, because the reinforcements need escorts, and that takes time. If my opponent has horse archers/knights and engineering, any faraway mounted unit can rush over and destroy my siege unit reinforcements. However, I started the game with a builder approach, going for space race then got declared on due to my low power rating. Is it even possible to do space race on immortal and emperor? From monarch and up should I take the warmonger approach from the start? How to do this with a random civ?

 

[Eqqman]

Alpha will get you IW+mathematics reliably. In Medieval it's hard to do better than philosophy off a GS lightbulb, it's the largest beaker value tech you can bulb directly and is also pretty useless to the AI.

I'm not sure if I agree with this. Leaders like Mansa Musa and Asoka will love to get their hands on Philosophy to help beat you to Liberalism.

 

[Mutineer]

- What struck you as the biggest differences in playing immortal, compared to emperor?

Main difference is shift of focus in learning. You shoould had develop your own strategies on Emperor level and early. Learn to handle your own empire, you gambits, GP use, tech trading strategies. In early levels you can do everything yourself and be in good enogth position to leverage your advatntages when you need them.
Starting from Immortal you need to concentrate on EXPLOTING AI.
you allready know how to manage your own empire, but how you can leverage that againt AI?
You can not overproduce, overgrow, overresearch AI, not untill late in game when you become resonably big. So, you need to learn to predict future, what AI will or could do and how you can exploite to your advantage.

 

[uberfish]

for what purpose do u need a warrior near ur capital early in the game?

To prevent your settler being eaten by a lion.

 

[uberfish]

I'm not sure if I agree with this. Leaders like Mansa Musa and Asoka will love to get their hands on Philosophy to help beat you to Liberalism.

If you want Liberalism first, you can usually get it anyway. Especially if you burn another scientist to part-research education. The computer will not do a focused beeline down the middle of the tech tree like human players do.

 

[Jet]

So on successful Immortal games, when do you find yourself choosing to use Great Scientists for Academies, rather than to lightbulb trade or race techs (or catch-up techs)?

 

[MrCynical]

The first great scientist will always go on an academy for my best science city. After that I look and see if I can lightbulb a monopoly, or near monopoly tech for trading purposes. Unless I can get an immediately tradable tech though I'll go for academies or superspecialists as applicable. Lightbulbing a tech that you can't trade, or is extremely undesirable to do so I don't find as effective unless it gives some immediate advantage such as an advanced military unit.

Any comments, experiences, reports of games won on Immortal without an early (pre LB) rush?

Update on my last post on this subject. I've just completed a game at Immortal level where I managed to avoid going to war myself until the start of the industrial age. The game almost meets the criteria in the OP, except that it was a small map rather than a standard size (my computer isn't very good and standard maps are sluggish in the late game), and I have Warlords, not Vanilla. That still only leaves me on about three out of 11 games at Immortal though, and that game got extremely marginal.

 

[Jorunkun]

The how-to-use-the GS example is actually a pretty good illustration of the difference between emperor and immortal level play, at least as I see it at the moment.

On emperor, I would do just as MrCynical sais and invest into an academy. In a spreadsheet-sim, this would offer the best pay back in the long run (provided the city you put it in is geared for commerce).

In my immortal games however, I have spent GPs on tech more often than not. Short term gain - provided you can trade the tech you bulbed - seems to trump long-term advantage.

That's because if you don't leverage your upside now, chances are that in the long run, you'll be dead.

 

[Jorunkun]

Firstoff, thanks for all the replies. Seems we’re in for an informed discussion indeed, thanks to the many immortal luminaries putting in their expertise. Again, if you have something to add or ask, don’t be shy. Btw, posts don’t have to be as long-winded and, ahem, wordy as mine tend to be … so suit yourself.

Update
In the last 36 hours, I played a dozen or so immortal starts and did some research to put together hard facts and observations for the “What Immortal Means” post (#2 in this thread). Please let me know what you think and by all means add to it.

My main takeaway: The AI is strengthened in production and research, the player is penalised much like on emperor. As a result, immortal doesn’t feel very different during the first moves; just a few more animals and earlier barbs, depending on the map. But when you begin meeting other civs you notice that, as all agree, the AI is spreading more aggressively and amassing an army earlier.

First insight: Leverage your strengths early
This doesn’t sound like a big revelation, but the implications are quite profound, especially when moving up from emperor. Personally, emperor has taught me to play tight and efficiently, but I was ususally able to mold the game to my preferred strategies. Getting early bronze meant you CAN rush, but with good land around you, you can also afford to play it safe, invest in more cities and improvements and hit with even more strength later.

As MrCynical, Uberfish and others say, you can’t do this on immortal. often, you are lucky to even get a shot at an early rush, and to pass it up is more of a risk than to take it. So leverage what advantage you have early. Playing it safe by investing in growth on immortal isn’t really playing it safe at all, as the AI will almost always outproduce you if you don’t hit early. Building seems to be more of a strategy of last resort, if you are lacking key ressources or facing a much more powerful neighbour.

Question: Do you agree, or is this just the exaggerated view of terrified builder-at-heart?

Potential vs actualised power
To help frame this: My abstracted view of the core dynamic in civ is one of potential vs actualised power. The game is a constant ebb and flow of parties relying on one or the other. Investing in potential power means putting your ressources in things that will yield greater returns later – building forges to boost production, libraries for research, granaries to increase growth etc. Actualising power means raising an army for war, lightbulbing a tech with a gp for trade or sending out settlers to claim land. The beauty of civ is that doing more of one means doing less of the other. Like if you build forges, your army doesn’t grow - your actual power diminishes relative to other parties. Raise an army and your infrastructure development and economy stall; you fall behind in potential power.

Understanding power
“Understanding power” for me means being able to read and respond situatively to this game dynamic in order to increase your power. It’s knowing when to invest vs when to actualise your power, depending on the situation and the game’s likely progression. It also means, tactically, knowing how to do this efficiently, achieving more with less (i.e. micromanagement and rushing, tech-strategy etc.). But my primary focus at this time is getting the knack for strategic pacing, especially as far as the start is concerned.

Which is why Mutineers post resonated very strongly with me. He sais the lower levels are about learning strong plays and executing them well - but on immortal, you need to know what the situation will call for, what plays are necessary. I take this to mean that the game’s course from turn one dictates what you need to do, not the other way around. You have to respond to the lay of the land and anticipate the strategic direction the AI is taking, not just play according to your personal preferences and habits.

Does this sound right, other opinions on this?

The immortal playbook
So to organise the results of this thread I had this idea of producing two “playbooks” for immortal level strategy. The first is a sort of laundry list of general dos and don’ts, incorporating the kind of high-level advice we’ve been gathering so far. I will post this on the #3 spot in this thread, and we can discuss and iterate on a few versions to drive the discussion and make the result more solid. I will put out a first draft later today.

The second would be a collection of plays or patterns that describe responses to typical specific game situations, characterised by the kind of map, lay and distribution of land, and taking into account your strengths and weaknesses versus those of the AI. An easy to frame example would be a playbook for island starts, with a high level outline of your options depending on who you are and what you find. Other situations to cover might be the hemmed-in peninsula start or variants of playing continents (finding these REALLY tough btw). Ideally, we would distill and proof-of-concept these in actual demo games, but let’s see how we fare with this thread first.

Feedback, questions, suggestions welcome.

 

[Jorunkun]

A couple of points from my side:

In response to my earlier question, the AI founds its cities in the following timeframe:

3190 to 2890 (turns 25 - 35) - AIs found their 2nd cities. With settler first and some luck, you may beat the later ones to it.
2890 to 2590 (turns 35 - 45) - AIs found their 3rd cities. This is when player can get a first settler out with a warrior/settler or worker-chop/settler start.

I guess the two main options in response to a near neighbour are to either race to get a settler out (i.e. to secure a way out of / block a way into an inaccessable piece of land) or to beeline BW and/or AH, plonk down a city on the ressource and prepare to fight your way through him/her in an early rush.

@MrCynical, I was indeed referring to settler as the very first build. Velocyrix has posted about the upside of this extensively over on Apolyton, and while this is a bit of a risk, consider that the AI doesn’t choke like a human would in MP, and that early barbs stay away from cultural boundaries, so you are quite safe. Share Uberfish’s concerns about the “naked” settler though, especially since you can’t be sure your starting warrior will survive until the build is complete.

As for all the good input on tech, we should get into more detail on this with v1 of the laundry list.

New questions:

What do you consider to be the hardest map types in terms of the strategic challenges they present for the whole game?
Continents sure is tough, but I found archipelagos quite manageable.

What role do wonders play in your immortal strategies?
If I build them at all, it’s usually more of an opportunity grab than a result of conscious planning, but reading the sg-threads I feel like I’m undervaluing them. Also, what’s up with the slingshot on immortal?

Cultural victory – when is this even possible?
Would hope that Godotnut can find the time to chime in here.

How do you most effectively shape AI diplomacy?
Probably too general a question, but would be interested in early open borders, use of religion, importance of giving in to blackmail and using tech to drive inter-AI wars.

And: What do you look for when “reading the game”, as Mutineer suggests?
Lots of things, obviously: AIs access to strategic ressources, room to expand, diplomatic tensions. Which of these are most pertinent, which combinations most strongly suggest what kind of response?

 

[MrCynical]

What do you consider to be the hardest map types in terms of the strategic challenges they present for the whole game?
Continents sure is tough, but I found archipelagos quite manageable.

Archipelagos are substantially easier than the rest of the maps. You can generally still make contact with AI's by sea, but it's nowhere near as hard to land grab or defend yourself. I don't find a lot of difference between continents and pangaea.

What role do wonders play in your immortal strategies?

I tend to assume I won't get any of the earlier wonders, but some of them are practical to get. Stonehenge is still viable, though it'll leave you even further behind in the land grab. Still, it might help to culture some of the nearby AI cities. The Oracle tends to get built around the same time as Stonehenge in my experience, so I think it's less of an option. It's also down a line of techs which is largely useless unless you found one of the first three religions, which requires far too much luck at this level to rely on. Even with stone, I don't think you can rely on the Pyramids at this level given how expensive they are.

The wonders which require an existing city improvement tend to get left behind for quite a while, probably just due to poor AI planning of said improvements. As a result the Colossus, Great Library, and Hanging gardens are all serious possibilities, especially if you have the appropriate resource. (I managed to build the great library without marble in my last immortal game, and the steady stream of great scientists was a huge help).

Cultural victory – when is this even possible?

Not sure about this one. I've never even come close to it.

How do you most effectively shape AI diplomacy?

This I find is the key to Immortal games. As far as open borders go I'd suggest that unless you have a chokepoint to protect you should open them up to all except one or two civs. It helps your trade routes and diplomatic relations. The only reason not to open with everyone is to avoid too many trading with worst enemy penalties.

As far as religion goes I usually assume I won't found one of my own. Adopting a state religion is generally only an option if all (or preferably all but one) of my immediate neighbours share the same one. That way they can act as reliable allies and buffer zones against the enemies you create. If you have several civs with different religions bordering you stick with no state religions. No allies or enemies is better than more enemies than allies.

Giving into blackmail is a frequent necessity at this level. I'll generally only refuse if it's a tech I need for trading purposes, or the civ s not near enough to my borders to present a threat (or I'm planning on going to war imminently anyway).

Driving inter-AI wars by bribery is a very powerful tactic. Wars slow the tech pace as resources are diveting to unit upkeep, purchase, and supporting captured cities. The more wars (that you aren't participating in) the slower the rest of the world's tech pace relative to you. It's important to keep an eye out for any civ willing to declare on another without your participation, and take advantage of it. When trading a tech around several civs, it's almost invariably better to give them the tech to start a war than not give it to them at all if you've already traded it to other civs. Take a look at which civs are pulling ahead on tech or land area and try to keep them permanently at war with at least two other civs. Sooner or later they'll start to fall apart, with a relatively small expense to you. If you have a neighbour that becoming annoyed and aggressive, but you can't beat in direct war, start a war between them and someone else. Yes, this will annoy them further, but the AI is reluctant to enter multiple wars so they are less likely to actually attack. In any case, if the worst comes to the worst you'll have some assistance.

Of course in all this you've got to watch that the civs are a reasonably even match in power. Starting a war in which one civ gets flattened in a few turns only boosts the other further. You want a long slow war where both sides get pillaged, the border cities swap hand innumerable times, but neither makes much progress. Their tech pace then grinds to a halt, leaving you free to pull ahead.

And: What do you look for when “reading the game”, as Mutineer suggests?

There are a few obvious things. Do any of the nearby AI lack copper or iron, assuming I have one of them? Prior to longbowmen there are few better targets. Are any of my neighbours losing badly in a war? Odds are both sides will ask me to join in, so pick up some extra diplomacy points by joining in at the winner's request and bag a bit of their territory. Then seriously considering stabbing your new "ally" in the back once the losing civ is destroyed. The AI is inept at defending newly conquered territory, and has a bad habit of using cavalry for the purpose. While their newly captured terrain is in revolt many of their units in cities will stay damaged and lacking defensive bonuses. Again, very easy pickings, and you'll never get a better opportunity.

Most importantly, is anyone pulling away at the top of the score graph? Does this civ have the most powerful army? If there's a civ that's well ahead in land and population but at most second on troops it's always worth starting a war to bring this potential superpower down before they get ahead on tech. If they aren't very close to my civ I may declare on them myself as it's easier to bribe other civs to join an existing war than to start a third party war.

A final thing I personally look out for is; is Mansa Musa present? His personality is sufficiently distinct form the rest of the AI that even if he's a long way down in land area he'll be ahead on tech for much of the game. I look immediately for potential enemies for him. Even his advanced tech can't save him when he's been at war with at least two other civs for much of the game.