Unit production baselines, rules of thumb

[jairgraham]

Any advice on this if you want to play conqueror? Champion city decked out with rax/range/shrines is not enough on its own, three ' pure war' cities isn't enough either. Do you pretty much have every 'war family' city decked out with 2 rax 2 range building units all the time? That doesn't seem like enough either, do you 2 rax 2 range about half your cities and dedicate them mostly to units? 2/3rds? Every city you can?

For those of you who kill other majors, how much do you prepare and how much do you commit?

[Siontific]

I can't imagine you ending up with a shortage with that type of setup. I suspect where you're hitting shortfalls would be in tactics, science, and rush-buying.

Tactics:
Ultimately if you're losing more units than you're killing consistently, then it's going to feel like you don't have enough production. Similarly, at the end of the day you're likely just going to be forced to fight someone who has more units than you do, and better productive capabilities too. So conversely, you just need to learn to kill 5 units while only losing 2. Or killing powerful units while only losing cannon fodder. Etc.

Science:
Leveraging power spikes is a big part of successful aggression in this game. It's entirely possible to consistently be the first nation on the map with an 8 strength unit by leveraging your laws. This is a powerful burst in military might. Just a handful of 8 strength units at the back of an army in a world where your enemies are still fielding strength 5 or 6 units will be devastating to deal with, and will make the above goal of killing more units than you lose all the more achievable.

Rushing:
raw training is only one part of your economy. You can rush an 8 strenght unit with just a few hundred civics, or a couple thousand gold once holy war is activated. You can, and **should** be doing this. Late game economies spiral out of control with the potential for massive stockpiles of gold, civics, or training. These are units you can build in a single turn from cities with zero military capability. It is entirely possible to win this game via war and conquest without ever building a single barracks if you don't feel like it.

Typically I make 1-3 major military hubs and build myself up toward a power spike I can leverage. Then, come the end game, I use a stockpile of low-rank units to mass upgrade into endgame units. From there, I use the strength of my economy to turn any city I own into a military hub - limited only by the scope of my economy.

[Gregorovitch]

I'm an improver level OW player currently trying The Noble so take this FWIW:

I've watched folk like FluffyBunny playing MP and noticed they build a *lot* of barracks, garrisons, ranges etc. They therefore build up a *lot* of training. So I kinda do the same. The principle I work on is that is if I can see off any funny business from either tribes or other civs today then I can sort out my economy tomorrow but the the other way round is not always true.

Consequently I regularly pop military buildings in all my cities for the training and I also regularly build more units. I do have certain cities that have bonuses to the production on certain units but I also build units across the board. I tend to build two buildings/citizens for the city then one for the army unless I'm at war or thinking of going to war, then I spam units in all cities until I feel I've got enough, then build a few more just to make sure.

I don't think you can have enough units really. The AI is ruthless but not suicidal. If you can alpha strike it it's like ooh err, I'm out of here. You need a good army to do that, numbers matter to focus down the kills. At the end you may have to invade a bunch of cities to make up some points - quickly - and that needs plenty of units immediately available as the clock runs down.

[krabdr]

I don't purposefully go to war a lot, but I strive for minimum of 2 barracks, garrison, and stronghold in each city. The movement bonus from the garrison-stronghold-barracks adjacency is very strong in the mid game. I try to get shield production up so I can make Axeman or Spearmen in 4-5 turns. I then try to have 5-6 melee units in each city. If the aggressor is stronger, I build more units, including more ranged. If my special unit isn't mounted, then I'll build a 2 ranges in each city and pump out my specials. I strive for a Citadel in at least 3 cities.

I haven't tried the Holy War rush, but it looks very interesting.

It's a game of attrition, but local superiority and solid defenses are key to that attrition.

[Siontific]

One tip that might not occur to most people immediately; leverage mills. Windmills and watermills grant 100% yields to adjacent mines, lumbermills, and quarries.

When you find iron and improve it, and construct the military complex around this city like usual, don't forget to surround your iron tile(s) with mills once you unlock them with hydraulics.

This applies to marble quarries, too. Generally speaking, it's better to surround a single iron mine with mills rather than spreading them out for perfect tessellation to boost regular mines - the reason being is that the marble quarry and the iron mine produce key production resources, so enhancing these tiles specifically make your cities more efficient as you head into the endgame.

Absent Marble or Iron, you can target the Gold/Silver/Gems if you want to enhance your gold income, but tessellation will probably be more useful here.

[IceMatrix]

Originally posted by Siontific:

One tip that might not occur to most people immediately; leverage mills. Windmills and watermills grant 100% yields to adjacent mines, lumbermills, and quarries.

When you find iron and improve it, and construct the military complex around this city like usual, don't forget to surround your iron tile(s) with mills once you unlock them with hydraulics.

This applies to marble quarries, too. Generally speaking, it's better to surround a single iron mine with mills rather than spreading them out for perfect tessellation to boost regular mines - the reason being is that the marble quarry and the iron mine produce key production resources, so enhancing these tiles specifically make your cities more efficient as you head into the endgame.

Absent Marble or Iron, you can target the Gold/Silver/Gems if you want to enhance your gold income, but tessellation will probably be more useful here.

lowkey this is something I am really bad at planning. I always think I should be doing this, and then I forget to. Excellent advice.

[krabdr]

Good point about mills, but they usually enter the late game for me. Interesting point about tesselation. I'll have to play around with that.

[Siontific]

It's possible to beeline Hydraulics relatively early, there are 3 laws long the way, and you're rewarded with two Crossbows at the end of it so it's not the worst idea.

With a wider empire, especially in combination with an ishtar bomb, you can set up couple mills in each city such that you now have 2-4 science in every city that isn't strong cultured.

Plus your resource output explodes, allowing you to relax your order economy.

[jairgraham]

Originally posted by Gregorovitch:

The AI is ruthless but not suicidal.

See I think it's the opposite. They will take relatively bad truce deals as soon as peace is an option, and truce events only adds to that. They're not ruthless at all. But they will grind their economy into dust building units and suicide them into a bad choke or just generally suffering negative attrition.

I just did my first 'Glorious' level game, and honestly the Major AIs all preformed worse than 'Noble' I think due to the ridiculous amount of tribes and barbs everywhere hindering them more than they hindered me. I actually had one guy 'weaker' than me most the game, and the #2 and #3 were only 'Stronger' most the time and rarely 'Much Stronger'.

The game was ending with a double war with the 'Stonger' and 'Much Stonger' #2 and # 3 (Greece and Egypt) and them attrition themselves to a standstill while I gathered the points I needed to have a large (15 point) lead. When the AI is 'stronger' I find I can take a city every 20-30 turns, when it's 'Much Stonger' I find I have to give ground and focus on unity kills.

It's a slough, it's not particularly fun, and I don't really think the AI behaving this way actually makes the game more challenging. Luck was not really on my side this game, but like I said the victory margin was large while on Noble it's sometimes much closer. Perhaps a few more games on this setting will change my mind.

Siontific, it's interesting to suggest ways to get more resources, but I don't usually find resources to be the issue, it's just raw quantity of units. The idea of getting Xbows early is appealing though, as they seem pretty hard to deal with as they're not really that weak against horses and they're very strong against a lot of units. More science options is always nice too.

On the other hand, I'd opine that other than very early game Gold/Silver/Gems are bad resources, as they don't produce much when resources are valued at 10-20. That gold mine is worth 2 stone per turn when stone is 20 coins. When resources cost 5 and they're worth 8 of whatever, they're okay, but just okay, and still worse than a 12 point iron mine or even an 8 point mountain quarry. Like a gold mine midgame is usually worse than an arid, low yeild farm, or a mine or quarry dropped in a random spot. It's not even worth developing until normal tiles are all done. Towns are only worthwhile for their side perks like landgrab and urban tile building.
___

I think you need like 60-80% of your city production going to units with Rax's and Ranges to 'keep up' to AI's, it's pretty crazy. And kind of, eh... I might just play on a lower level and enjoy the sandbox, because killing scads of barbs into killing scads of AI hoards is kind of tiresome.

[Siontific]

Well, different strokes and all of that. I like to be at war for most of the game because I find it interesting to have my attention divided between a number of things.

Ultimately it's worth noting that as a single player experience, the computers ultimately serve as obstacles to the player. No one else in the game is actively trying to win the game or else the player would virtually never win on even the remotely highest difficulty.

You cite things like truce, for example, and well - truce is a diplomatic tool; a diplomatic tool doesn't work against a player who's trying to win the game. So if the computer were only concerned about winning then the theme of the game effectively gets tossed out the window as every computer nation would automatically recognize the player as the largest and singular threat since the game is only happening due to the existence of a human player; which means they "win" and the game ends as soon as you, the player, loses.

Similarly, as you describe, they'd never attack each other and for the same reason; it doesn't benefit them on a strategic level since the only ACTUAL win condition that exists in reality for computer opponents in a game is for a human player to lose.

So this is a bit where the game becomes a game, and theme impacts gameplay; Nations will go to wars with each other just because "reasons" -- these reasons have mechanic explanations; power level, opinion, whether it's more or less tactically sound.

But if you're expecting the computer to go "wait a minute, if we fight this war, the human can sit by building all of the last couple of wonders and win a points victory", well; two things;

1.) You're expecting the computer to play like a human, and it's not going to. I don't mean play by the same rules as the human, cheating vs not cheating; I mean actually PLAY like a human. It's bound by rules and algorithms that we don't have -- the entire point of a peace treaty, for example, is to make a nation less likely to declare war. Except the human player can declare war whenever it wants, so this doesn't apply to human players, while it does, mechanically, make computer players less likely to declare war. (diplomacy is very powerful in this game, by the way)

2.) You should probably toggle on "ruthless A.I. mode" -- This setting will make the A.I. less beholden to the typical rules that tie it's hand when it comes to considering the human player. For example, in Ruthless mode, enemy nations are in fact more likely to just declare war on the human player solely because they're close to victroy. They are also less likely to accept and instigage truce deals. This is probably a setting you'll enjoy.

[Siontific]

Back to your concern about barracks and ranges and spending 80% of your time building units, this is just untrue.

I'm looking at a turn 116 conquest victory on Glorious Difficulty.

I settled 3 cities myself; built 4 barracks myself, 3 ranges myself, and had exactly 2 military cities. I constructed 45 units in the game: 12 warriors, 3 ballista, 9 onagers, 2 hoplites, and 19 phalangites.

I began the conquest of the Hittites on turn 59, and the conquest of Assyria on turn 90. The final city of both empires fell on turn 106.

But I do agree that playing on a lower difficulty can be a lot more fun; that's why glorious is my "have fun" setting - I can get crazy with stuff. If I feel like things have been too easy for a while, I'll jump back to The Great.

Also, a note on difficulty settings since I know some players like to reduce the development setting to none to keep things "even" with the computer; If you reduce the development setting but up the tribe strength, you run the risk of setting the computer player up for failure.

[jotwebe]

Originally posted by Siontific:

Similarly, as you describe, they'd never attack each other and for the same reason; it doesn't benefit them on a strategic level since the only ACTUAL win condition that exists in reality for computer opponents in a game is for a human player to lose.

So this is a bit where the game becomes a game, and theme impacts gameplay; Nations will go to wars with each other just because "reasons" -- these reasons have mechanic explanations; power level, opinion, whether it's more or less tactically sound.

But if you're expecting the computer to go "wait a minute, if we fight this war, the human can sit by building all of the last couple of wonders and win a points victory", well; two things;

1.) You're expecting the computer to play like a human, and it's not going to. I don't mean play by the same rules as the human, cheating vs not cheating; I mean actually PLAY like a human. It's bound by rules and algorithms that we don't have -- the entire point of a peace treaty, for example, is to make a nation less likely to declare war. Except the human player can declare war whenever it wants, so this doesn't apply to human players, while it does, mechanically, make computer players less likely to declare war. (diplomacy is very powerful in this game, by the way)

OP isn't expecting the computer to play like a human, they want the different AI factions to play as if played by humans. Your scenario where all the AI played nations collaborate to put the human down is not what this is about. They (I think) want AI Egypt to play to win, and AI Rome as well. So AI Egypt would go after AI Rome instead of Player Hatti if that was more promising.

What they are criticizing, if I understood correctly, is that AI nations burn down their economies in order to win defensive wars via mass rush buying, and that's not how humans would play.

For many, more peaceful minded players, I think that's not wrong, but that's really those players (and I count myself among them) not playing the game aggressively enough. So the AI are just playing better because of a lack of sentimentality and sunk-cost fallacy.

There's also the fact that, in a losing war, it is also very human to burn down the nations chance of winning out of spite. Long-term, over several games, that is also a good meta-strategy - make it clear that you will be no easy conquest and will burn the earth behind you will make you a less attractive prospective meal.

So overall I think OP is wrong here, but not for the reason you're going at. It's true that OP should try to set the AI to "ruthless" though, then in the game they described the combatants would probably have gone after them as they approached victory threshold.

[jairgraham]

Originally posted by Siontific:

...

No, this game was made with the C-Evo philosophy, the AI's are trying to win. That doesn't and shouldn't mean the humans can never win.

Human players in strategic and tactical games absolutely use tools like truces. Some games, like 'Diplomacy' revolve almost entirely around agreements, truces, and unstable alliances.

War should be a situationally strategically good option in a game with evenly matched players. It is in Old World, but the orders costs and AI unit spam makes war a generally worse option than other 4x games, not a better one. The Old World is weirdly peaceful outside of tribes.

On Magnificent I got an alliance with Rome, the weakest player in the game. They had over 30 Dromon on turn 90. Plus an entire army even more unity spammy. The other AI's had even more. How does dealing with that mob say anything but 'tedious' even if you could keep up to it (and no, a dozen cities can't realistically keep up to that plus build an army in a reasonable time)

I WAS on the back foot by 10 points that game and probably needed a war or something to win, but the only reason I was on the back foot was because Carthage got 8, yes 8 wonders (I had 2) and large, unobstructed tracts of land far away from me to expand with. And they got the wonders mainly due to the city and technology advantage that that level gives the AI.

Game would still be winnable for me via peaceful expansion if I set the score threshold to 'High', which seems to be what other Magnificent and Great players tend to do from what I've read so far.

"But if you're expecting the computer to go "wait a minute, if we fight this war, the human can sit by building all of the last couple of wonders and win a points victory", well; two things;"

In fact I am. And it's relatively easy to implement. The leader getting dogpilled when they're within 10 points of victory would be great, and people turning against the #2 when they're threatening to overtake #1 would be even better.

But even if that option is out. The fact that they will wreck their chances to win to totally commit to war is kind tedious.

So you say you can wipe out an enemy AI on Glorious with about 20-25 units on turn 60? I've never seen an AI with so few units, not even Babylon. 45 units ALL GAME and you're fighting an AI on turn 90? You can't be serious. In my experience a war on turn 90 means fighting 40-50 late game units minimum. Even assuming your warriors etc are all upgraded to swords at that point it sounds, unfeasible, to fight that many units efficiently. A turn 90 war to me means having two clusters of 5 mangonel/onagers at choke points guarded by a much larger mix of infantry and archers with Calvary in the flanks for cleanup and looking for gaps in the endless stream of units so I can gradually and painstakingly take boarder cities one at a time. And then I pray the naval game doesn't distract too much because they have so many ships and there aren't really tactics to deal with them in the same way you can deal with land units.

I don't know. I don't have Hittites so perhaps DLC changes more than just adding stuff?

Originally posted by jotwebe:

...

Well, yes to the first paragraph. Kind of to the second? If SOME AI's did that SOME of the time it would be interesting. But they all follow that script and it's certainly not 'playing better' to consistently take a losing strategy. Come to think of it I probably could win the game I'm in by just asking a weak power to war with the wonderspammer and staying out of it otherwise, because the AI will consistently stall everything else they're doing when at war, and they don't usually make any progress against each other.

I wonder if 'ask to declare war' is just an 'I win' button? I should experiment with it.

[Siontific]

Originally posted by jairgraham:

Originally posted by Siontific:

...

No, this game was made with the C-Evo philosophy, the AI's are trying to win.

Except it literally wasn't. The nations are not explicitly playing to win the game. They are characters in the world. Indeed they're playing the game, and they're doing it decent enough- but they aren't aiming to achieve victory. Victory is achieved purely by the natural forward momentum of culture, science, development and expansion all providing Victory points - such that a large and well developed nation is going to naturally collide with the victory point cap because it's inevitable.

Essentially everything in the game serves as an obstacle for the player to overcome and the computer nations are no different; they set the timer for the game, and this timer is sped up the more you increase in difficulty because the higher the difficulty, the higher the disparity between human and player start conditions, and the more likely the the momentum of the game will cause the computer to crash into the point total. The player's job is to win before this moment.

In any event, reading the rest of your post - you're clearly looking for what ruthless A.I. has to offer.

In fact, that option used to be named "play to win" - why do we think that's the case? Because as an option, it makes the computer seem like it's more concerned with winning; clearly demonstrating that in a regular game of Old World, it doesn't much care.

This isn't a crazy phenomenon. In a single player game, the enemy nations are your "boss fights" - it's not about a completely symmetric one to one experience where all players are playing some chess like version of the game. All that matters is that a player is challenged, and that the world feels "alive" - There have been numerous player complaints in the past, for example, that they can play an entire game of old world and none of the nations would go to war with each other.

In fact, that was eventually tweaked, and perhaps that's part of the problem creating the issue you're noticing. But here's the thing;

people are playing a videogame about dynasties and empires in history. The expectation is that these nations will end up in wars against each other, and that a sort of history is going to unfold in the game.

Let's say for example, If the absolute best and most optimal decision a computer nation could make would be to avoid wars because it's too draining on their economy, and then that's exactly what they did - you would have those same types of complaints;

"Hey, I just played a game as babylon in the middle east and I noticed Assyria, Persia, Egypt, Hittites and Greece never ONCE fought a war in 150 turns? That doesn't seem realistic"

Instead, it's much more entertaining to watch these empire clash around the player - and, sure, create opportunities for the player to take advantage of political situations to their benefit; like attacking Assyria at the tail end of a decades long war with Persia.

[Siontific]

Originally posted by jairgraham:

Originally posted by Siontific:

...

Game would still be winnable for me via peaceful expansion if I set the score threshold to 'High', which seems to be what other Magnificent and Great players tend to do from what I've read so far.

Also, again - this is just not true. Plenty of players can win this game on the great under normal conditions. You can win the game with 1 city, you can with the game with 3 cities, 6 cities, or a dozen. You can pivot into conquest or ambition from either of those states.

One of my go-to strategies is to build only 3-5 cities, consolidate a strong core, and then conquer another nation. I do this on the great difficulty, with zero settings modified, on a medium map, and I do it all the time.

Originally posted by jairgraham:

Originally posted by Siontific:

...

So you say you can wipe out an enemy AI on Glorious with about 20-25 units on turn 60? I've never seen an AI with so few units, not even Babylon. 45 units ALL GAME and you're fighting an AI on turn 90? You can't be serious. In my experience a war on turn 90 means fighting 40-50 late game units minimum. Even assuming your warriors etc are all upgraded to swords at that point it sounds, unfeasible, to fight that many units efficiently. A turn 90 war to me means having two clusters of 5 mangonel/onagers at choke points guarded by a much larger mix of infantry and archers with Calvary in the flanks for cleanup and looking for gaps in the endless stream of units so I can gradually and painstakingly take boarder cities one at a time.

I'm very serious. I didn't kill just one nation in that game; I killed two - The Hittites and Assyria were both dead by turn 106 on standard, unmodified glorious settings.

Of course, other games can indeed be more of a challenge. On a different game where I won on turn 153 on The Great, I built 120 units all game. Clearly that one took a bit more effort. Looking at a third game, this time on Glorious again, it's turn 148 victory, and I build 150 units in the game, having the highest unit production of any other nation in the game. (Though #2 was at 148 units) - for units killed, I killed 276 units in that game; a very bloody one indeed.

But here's the thing; ultimately this is player preference, but I find it FUN to have massive wars with a bunch of units. You see tedium; that's okay, we aren't going to like the same stuff, but if anything I think it's actually TOO EASY to win a war these days, and have made some complaints as such. But players like myself are in a certain minority since you can't really balance a game around it's best players. Not to toot my own horn or anything, but I just mean as a general design philosophy; if 5% of the audience think wars are too easy while 95% of the audience things wars are too hard, hard, just right, or even impossible... well then you don't go changing the game to make wars even harder to satisfy that 5%

So I think you're in a particularly unique position:

The diplomatic element of this game is very easy to "exploit" - this isn't necessarily a bad thing, but you can leverage a lot of tools to keep the computer nations off of your back so that you can sit back and chill and develop your empire. Do this right, and you can just win the game without conflict (Caravan spam used to be a core strategy of mine I relied on).

On the one hand, you seem to dissatisfied with how easy this is.

Conversely, wars take a lot of effort. If you want to go to war via conquest you either need to be a solid tactician, know how to leverage power spikes, and know how to super-charge your economy; or all three. The reason for this is, as you've noted, the computer nations will established a veritable mountain of units that you need to climb over in order to kill them.

So on the other hand, you're dissatisfied with how much of a hassle that is.

It's a tricky spot to be in, for sure.

[Siontific]

One final tidbit:

I just found out that there has been a bug with Computer nations at peace that seems to make it harder for them to declare war. This is fixed internally, and may have a lot of impact on any current criticisms about how the A.I. does or does not choose to declare war against a human player.

In any event, play a game with ruthless A.I. on and see how it goes; maybe even try it on the test build to see the effects of the bug fix.