The best kittens, technology, and video games blog in the world.

Friday, February 12, 2021

More Empire Total War advanced tips

king of the castle by allenthepostman from flickr (CC-SA)

Last month I wrote Empire Total War advanced tips but then I realized I have a lot more.

How to finance your empire

Empire economics is much more challenging than in earlier games, where you could just loot your way out of anything. And the game doesn't make it intuitive how to do economy well.

Economics of trade ships

The first type of income to consider are trade ships. Each ship costs 600, has upkeep of 50, and produces 17 goods (20 for the first one in each node). Trade goods price start very high, but they'll drop to about 7-10 if countries are actively trading. Game has some sort of supply and demand model, but I don't really understand how it works.

For some assumption I made here, let's say each ship takes 2 turns from being built to reaching its trade node (and you're paying upkeep at this time while getting zero income), extra ships don't affect prices, and there are no war fleet or other costs in maintaining your trade fleet. These are fairly reasonable assumptions most of the time, especially if you eliminate pirates early.

Max number of ships is 20 ships per node in 20 nodes, so 400 trade ships active. It's really easy to grab half the nodes before anyone else figures thing out, so you should have no trouble having space to send your ships to. Or if you're late to node grabbing (let's say playing Russia), you could just grab a few by force.

With all 400 ships, max trade income from trade ships is, their upkeep is 20000, and based on trade prices income is:

  • 7 gold - 48020, net 28020
  • 8 gold - 54880, net 34880
  • 9 gold - 61740, net 41740
  • 10 gold - 68600, net 48600

And newly built ship will fully repay itself in, based on trade price of its goods:

  • 7 gold - 12 turns
  • 8 gold - 10 turns
  • 9 gold - 9 turns
  • 10 gold - 8 turns

This applies to Indianmen and Dhow (which are 100 gold cheaper to buy, but have same upkeep). There are also two hybrid war/trade ships Fluyt and Galleon which will not really even pay for themselves as their upkeep is higher than what they'd be making at typical prices. The game's idea is that you want to mix in a few of such hybrid ships to protect fleet of nearly defenseless Indianmen, but I never found it to be useful - they have low gun range, and low speed, and protecting slow trade ships is hard in the first place.

I found Galleons really hard to fight, as they are a rare ship type which has guns from sides, front, and back, preventing coming close from any angle, and their weird extra dangling mast is very hard to remove, so they keep a bit of mobility (so it's hard to approach diagonally too). They still have low range, so they eventually lose to long range shooting, but battles against Galleons just keep dragging on forever.

Economics of conquest

It might seem there's a more obvious way - conquering land!

Rome 1 and Medieval 2 had easy looting based economy and conquering provinces and looting them was a great source of income. There's no loot in Empire, in fact you need to pay to repair damage you caused.

At least taxes should work right? Unfortunately administrative cost formula is really brutal, getting up to -57% at world conquest (at least it's no longer -80% it was at release).

Let's assume for a moment every province is worth equally much, 2500 total wealth at 40% tax before admin costs (so nice round 1000 income per turn). 40% is 30% from normal tax rate, 9% from max level minor province government building, and 1% from just slightly above average minister.

Here's by how much your income would increaes per turn for your Nth such province, averaged for each range:

  • 1-4 - gain 1000 tax per province, no penalty yet
  • 5-20 - gain 713 tax per province
  • 21-40 - gain 562 tax per province
  • 41-60 - gain 454 tax per province
  • 61-80 - gain 377 tax per province
  • 81-108 - gain 308 tax per province
  • 109-137 - gain 254 tax per province

Conquering provinces worse than your average is going to cost you money rather than make it - penalty to every other province can easily be more than new province's income.

Economics of industry buildings

I mentioned it before but industry investment has terrible ROI. If you got that full 40% base tax, here's how long it would take to repay various buildings:

  • farm 1 (cost 300, wealth +100) - 5 turns
  • farm 2 (cost 500, wealth +50) - 17 turns
  • farm 3 (cost 700, wealth +50) - 23 turns
  • farm 4 (cost 1000, wealth +50) - 33 turns
  • farm 5 (cost 3000, wealth +50) - 100 turns
  • gold mine 1 (cost 1500, wealth +1200) - 2 turns
  • gold mine 2 (cost 3000, wealth +200) - 25 turns
  • gold mine 3 (cost 4500, wealth +200) - 37 turns
  • silver mine 1 (cost 1250, wealth +1000) - 2 turns
  • silver mine 2 (cost 2500, wealth +200) - 21 turns
  • silver mine 3 (cost 4000, wealth +200) - 33 turns
  • iron mine 1 (cost 1000, wealth +700) - 2 turns
  • iron mine 2 (cost 2000, wealth +100) - 33 turns
  • iron mine 3 (cost 3000, wealth +100) - 50 turns
  • textile industry 1 (cost 1000, wealth +300) - 8 turns
  • textile industry 2 (cost 2000, wealth +150) - 33 turns
  • textile industry 3 (cost 4000, wealth +150) - 67 turns
  • textile industry 4 (cost 6000, wealth +150) - 100 turns
  • iron industry 1 (cost 1000, wealth +400) - 6 turns
  • iron industry 2 (cost 2000, wealth +200) - 25 turns
  • iron industry 3 (cost 4000, wealth +200) - 50 turns
  • iron industry 4 (cost 6000, wealth +200) - 75 turns
  • port 1 (cost 1250, wealth +200) - 16 turns
  • port 2 (cost 2500, wealth +100) - 63 turns
  • port 3 (cost 5000, wealth +100) - 125 turns
  • port 4 (cost 10000, wealth +100) - 250 turns

I'm assuming average yield for resources, and very poor wealth for industry - since all newly spawned towns will be very poor and it never changes; while resource tiles all already exist and have about even mix of yields.

And similar figures for other types of buildings. It gets a bit better for better slots. This number also gets a bit better with tech, as there are modest cost discounts and modest wealth increases (increase is shown in city summary, but not in tooltip for individual building, just to confuse you a bit more).

But of course it gets drastically worse with your administrative efficiency - at 15 provinces you'll be losing 1/5, and at 40 provinces you'll be losing 1/3 of that income, so return will be a lot longer.

There are other considerations:

  • high tier buildings cause unhappiness
  • farms are weird are you get unhappiness at level 3 but then it disappears at level 4
  • farms cause population growth which unlocks new building slots, and that's extremely useful - at least as long as there are undeveloped slots left
  • ports are needed to export plantation goods (tooltip in city screen tells you if you need to upgrade), and to trade with more countries - you're not really upgrading them for wealth bonus

So you should always prioritize level 1 buildings, as their returns are great. Farms are worth upgrading up to level 4.

High tier mines and industry are rather questionable. You'll need to build or conquer one of each type to unlock their associated research, but there's really little point in building tier 3-4 industry beyond that.

Industry also generates town wealth every turn, however that is extremely slow process, town wealth reduces itself when taxed, and income from generated town wealth is also affected by the same administrative efficiency as everything else. You get a lot more town wealth from enlightenment technologies anyway.

Economics of plantations and conquest

So far it seems like trade fleets should be your main investment as long as you're small or medium, while conquest and industry have poor returns. Is there anything else?

It turns out there's one more great source of income - plantations, and they don't suffer from any scaling issues.

Average yield tier 3 plantation produces 60 goods (sugar only has extra tier 4 building for 90 goods), at price of 7 you get 420 town wealth and 420 trade income for it. Town wealth suffers from 40% tax and then administrative efficiency, but for trade income you get 100% of it.

So every such plantation gives you 588 gold if you're tiny, and 490 when you're at world conquest stage.

By my count there seem to be 72 plantation slots (including fur slots; fur has slightly different build cost, but works mostly the same) worldwide, giving you another 36000 gold per turn.

Nice thing about them is that they completely change administrative efficiency formula. If we assume each region has on average half a plantation in it, in addition to 1765 non-plantation wealth, your gain will be, once you fully upgrade them:

  • 1-4 - gain 1000 tax+trade per province, no penalty yet
  • 5-20 - gain 773 tax+trade per province
  • 21-40 - gain 654 tax+trade per province
  • 41-60 - gain 568 tax+trade per province
  • 61-80 - gain 508 tax+trade per province
  • 81-108 - gain 453 tax+trade per province
  • 109-137 - gain 404 tax+trade per province

It's much reduced penalty, final province being 38% of your first, rather than 25%. Of course this is only true as long as you have at least one trade partner (which you won't when you finally conquer the world).

Economics of plantations

And of course you should upgarde all your plantations! Assuming 7 cost, 40% base tax, aand -33% admininstrative cost penalty (40 provinces), returns are:

  • plantation 1 - 8 turns
  • plantation 2 - 19 turns
  • plantation 3 - 22 turns
  • plantation 4 - 28 turns

At goods price 10:

  • plantation 1 - 5 turns
  • plantation 2 - 13 turns
  • plantation 3 - 15 turns
  • plantation 4 - 20 turns

20-28 turns from top tier sugar building is crazy better than 100 turns you get from investing in top tier industry building at this country size.

Plantation building costs get lower with technology, but then they get 30% higher once you abolish slavery. Abolition of slavery is fairly late game, and by that time your plantations are likely going to be all upgraded anyway. Even if they're not, it's still much better deal than industry.

One small thing. To get full benefit of plantations you might need to build and upgrade some commercial ports - and on small islands held by the pirates this also means destroying military ports first. If you go to city and mouseover trade goods, tooltip will tell you if you need to upgrade.

Town wealth

Resource towns (mines, farms, plantations, fur traders, and so on) have 5 different resource yield levels, and they never change.

Other towns (factories, schools, inns, ports) have 5 wealth levels from very poor to wealthy.

They are supposed to grow with town wealth, but it is very bugged. Inland provinces grow as they're supposed to, but if there's any port in the province, none of the towns will experience any wealtnh growth.

Port village that didn't emerge yet do not block growth, only after they emerge.

All this is confirmed by a lot of testing. Since most provinces in Empire have a port, people think town wealth just doesn't change period. It's supposed to, and it sort of does.

I think ports slots that are empty or damaged stop blocking town wealth growth, suggesting that destroying ports for a while, and rebuilding them later might be extremely profitable, but this is based on very small sample size.

Demanding surrender

Now for some non-economic tips. When you siege a settlement there's demand surrender button, which lets you take settlement without a fight.

AI will accept it only if you have overwhelming advantage and it's not a capital. Like full stack vs 2 garrison units kind of advantage.

As far as I can tell:

  • it will never accept if it's its capital
  • it does not consider reinforcing troops, so 1 sieging units with 19 unit reinforcement won't make it surrender, even if 20 units sieging would
  • it generally won't surrender if there are any non-garrison troops present - I vaguely recall it is possible, and they then withdraw, but it's been forever since I last had this happen

There's small downside to demanding surrender - AI capital province gets some unrest from war weariness for every battle lost, and that makes it more likely to peace out - but any successful demand for surrender doesn't count for this. So if you want to peace out AI, it might be better to fight (autoresolving at those odds is fine).

Capital switching

If you take faction's capital, but it has more settlements in its home theatre, another one of its settlements becomes its capital.

If you want to take multiple settlements of a faction, and it makes little difference to you otherwise, try to take capital last.

Capital gets extra garrison units, and never accepts demand of surrender, so having non-capitol fight followed by capitol fight is slightly easier than two capitol fights.

High tier royal palaces

Some buildings can only be built in faction's capital. That doesn't mean just its original capital.

One thing you could do is take someone's capital, then wait for them to build top tier royal palace in their new capital, take that and so on. You keep any such buildings on capture.

Those bonuses aren't that big, so I don't think it's really worth the hassle usually, but it's a thing that happens.

Walking through towns

Walking on roads is much faster than walking on open terrain, and if you click somewhere, your troops will follow roads to get there, getting you close to maximum possible speed.

There are however some exceptions to this. Notable if there's a town on a road, your units will not automatically go there - they'll take off-road detour and waste some movement. For friendly or enemy towns, you can go into the town and then continue the road, and usually save about one action point on this. You cannot go into neutral down (allied or if you have military access), so that's always a bit slower.

Walking into enemy town also automatically damages it for no extra movement cost. If you want to minimize this damage, there are ways.

Damage to town buildings by troops

Hostile troops entering a town damages it, unless it was already damaged. This seems to be in proportion to stack size. So if you need to enter a building, but you'd rather not damage it too much (because you plan to conquer that province and will need to pay for repairs), just pick one unit from your stack, move it there, then move the rest of the stack to join it.

In costs zero extra movement to do it this way.

Or you could just walk around and not cause any damage, usually the movement speed difference is tiny.

Damage to city buildings by troops

Much more expensive damage happens when you win a siege - cities, especially 5/6 slot cities, tend to have a lot of expensive buildings which get severely damaged by winning the siege.

Again, as far as I can tell, damage is proportional to number of units in sieging stack - but without counting reinforcements. So if you really want to save some money, just siege the city with one unit, have everyone else stand just behind it. It will make deployment a bit of a hassle, but AI is usually fairly passive while defending, so it won't try to attack you while you're gathering troops.

Nothing you do during the battle has any consequences to building damage, except damage to fort walls (and that drops fps to 0, so even if you play with forts, you should do your best to avoid it).

Damage to buildings by rebels

When workers riot, they damage one of the buildings by about 80%, and if it was an expensive building it will be expensive to repair.

There are ways to optimize this. Rebels seem to only target government building (civilian or military), or building in industry slot (so industry, school, religious school, or tavern). I don't think they ever target anything else like farms, roads, forts, plantations, ports, mines, or military recruitment buildings.

If none of the buildings have relevant kind, nothing get damaged.

If target is already damaged, rebels just damage it a bit extra. If you expect riot after riot, you might want to just not repair it until province is under control. Malta is a good example - the only valid target is government building, which will likely be very expensive royal palace, and it takes forever to get its -30 resistance to foreign occupation down. In the meantime you can have multiple riots trying to damage it. After first such, you don't have to repair it right away, and you'll save tons of money.

If you plan to destroy a building (like extra university in province you don't want one), of course don't repair it - destroy it as normal.

Interestingly, if the building is being upgraded, the upgrade is not cancelled, and once upgraded, it will be repaired for no extra cost. This means you can freely upgrade buildings you conquered from technologically inferior enemy without worrying about first round of riots.

Zones of Control

Every unit has zone of control around it. Land units only on land, naval units only on water.

You can mostly ignore them and just charge at the nearest stack or settlement, but knowing finer point of how zones of control work can be very useful.

You cannot walk through unit's zone of control. If you step into enemy zone of control, they have an option to start battle, in which you'll be the attacker (so any battle timer works against you).

Units with overlapping zones of control will reinforce each other - even if their overlap is tiny. This means one stack in Berlin and one stack in Dresden will reinforce each other in battle - even if that battle is actually on the opposite side of whichever unit that is involved. At least it doesn't chain any further.

Attacking units without reaching them

If you just walk into unit's zone of control, it can start a battle. You can generally predict if AI will do so or not.

If you want fight enemy army, but you also want to go somewhere else, you don't always have to attack the army itself. Just walk a bit into its zone of control, and if they accept the battle, you'll save a lot of movement.

You can even move there with part of your force (even just one unit), and have your main force join battle as reinforcements, at zero movement cost. Reinforcement range is very high, but I recommend saving game before such tricks before you get used to it.

Walking through enemy zone of control

If you're trying to get somewhere and there's small enemy force on the way you want to ignore, you can just step into their zone. If they refuse the battle, you can keep walking to the place you want (like city you want to attack) without their zone counting anymore.

Zones of Control and blockades

If you're attacking a unit, it doesn't get an option to start battle even though you walked into its zone of control. Now this seems like a fairly pointless distinction, but if you're attacking a settlement you can start a siege without actually starting a battle. Likewise, if your ship attacks some ships in port, it will just start a blockade without any battles.

The sieged / blockaded units can sally out on their turn, but by that time we might have done what we wanted. And they cannot attack you while you withdraw.

Units under siege or blockade don't have zones of control. This is extremely useful. Let's say you want to invade Malta, but you have just two Sloops, they have a full stack of 5th Rates in their port, and you really don't want a naval battle. Nothing could be simpler - attack their port with one of your Sloops, this will stack a blockade. Then use another to drop your army. Using "Fast naval landing technique" I described in my previous post you can move army that landed and attack their city on same turn - wiping out the faction also wipes out their navy.

Zones of Control and sieges

You can do something similar with units in cities. Let's say there are 2 full stacks next to each other - one in city, one just outside, and you'd really rather not fight them in one battle, as you only have one stack there.

Simple, just use 1 unit to attack the city, and keep it under siege. Then use your stack to attack the stack outside with your other 19 units (plus the sieging unit will join as reinforcements anyways, so it's 19+1 vs 20 battle!). Units under siege cannot join to reinforce anyone, no matter how tiny the sieging force is.

After that you can have the rest of your units join the siege, and attack in a 20 vs 20 battle. Or withdraw if you're too battered. This is usually far easier than 20 vs 40 battle which can get really messy.

Units that cannot reinforce

Some units cannot reinforce. Most important of those are all fixed artillery - early game that's the only kind you can make (and AI never disbands theirs, so you can often find some even really late), and all natives' artillery. If there are two small AI stacks close to each other, attacking one without fixed artillery lets you wipe out the rest of the units more easily.

All survivors, including that artillery, will then withdraw, usually far away that they can't reinforce any nearby battles. Even if reinforcements were in the settlement.

Another unit type that cannot reinforce are garrison units. Western countries have really awful ones (half-size Firelock Armed Citizenry), but non-Western countries sometimes have units that are at least servicable support for a real army. Especially the full-size melee units, even if they're trash, can charge your artillery or overwhelm your infantry while real enemy army does their job. Pirate Mob is probably the best of all such units.

So imagine the following situation, enemy has:

  • city - 2 Milita, 5 Garrison, 2 Fixed Artillery
  • outside city - 3 Militia

If you attack the city, you face 5 Militia, 5 Garrison, and 2 Fixed Artillery you'll need to approach. It won't be easy.

Or you could attack that other stack, facing just 5 Militia with no support. And after that, face just 5 Garrison.

If your army is like 1 General, 3 Militia, 1 Fixed Artillery, then you can probably win those two separate battles without too much trouble, but it would be quite impressive if you managed to win the big one.

This is mostly very early game technique, but if you're playing with a mod that makes garrisons stronger (and it's not like they can get worse) like Darthmod, it can be useful much longer.

Surviving all cavalry attacks

This is really rare, but if enemy attacks your settlement with just cavalry, you can start timer and cheese the win. If there's any building on the map - and there usually are some - just put your Firelock Armed Citizenry inside and wait it out. Cavalry cannot dismount and get there, so you win by timer.

Similarly if you're on defensive, have infantry superiority, but enemy has a lot of cavalry, and there's no artillery on either side, go inside buildings, and you'll be fine.

It's a rare situation, but I've had it happen to me.

Research and happiness

You will invariably get Common Land Enclosures (-1 lower class happiness) very early, as it unlocks a lot of tech tree.

Mid/late game enlightenment technology Secular Humanism cuts religious unrest by half, which can be huge if you try to conquer Muslim lands, as it will cut -10 into just -5. Most most of the world conquering different flavor of Christianity, it just cuts -2 into -1 or such.

All other techs that claim to affect happiness only actually do anything in provinces with schools in them, and that's going to be tiny fraction of your provinces, and so you really shouldn't worry too much about them.

Notable is Abolition of slavery as it has huge swing of -4 for nobility, 0 for middle class, and +4 for lower class, but this only applies to provinces with schools, and it's modified by government's clamor for reform (-80% for republics, -25% for constitutional monarchies).

So abolishing slavery for constitutional monarchies is just awful (-3 nobility unhappiness for every province with school), for republics it's slightly useful (+0.8 lower class happiness for every province with school), and it's quite useful for absolute monarchies (-4 nobility and +4 lower class happiness for provinces with schools, but nobility happiness is usually so much higher than lower class it's always worth it).

Social Class happiness

Government types mainly differ by happiness of their social classes.

Let's also say that every province gets -2 lower class happiness from industry buildings in it. This is much more varied, but it should give a general idea.

So here's your happiness budget (happiness bonus, repression, and tech bonuses) based on government type. Lower of the number is the important one, so sorted by best to worst for happiness:

  • Constitutional Monarchy: +4 nobility, +4 middle
  • Republic: +4 middle, +2 lower
  • Absolute Monarchy: +7 nobility, +0 lower

Republics also get drastically less unhappiness from schools, while absolute monarchies get the most, but since it's concentrated in a few provinces, it's usually no big deal to manage it.

Ministers and happiness

As a consequence of these happiness numbers, the most important bonus ministers give are traits that change population happiness, it matters far more than their management skill which gives fairly insignificant bonus to various costs.

You should prioritize the following:

  • Absolute Monarchy - lower class only (nobility will always be happy no matter what, it's almost impossible to have unhappy nobility and happy lower class)
  • Republic - mainly lower class, then middle class
  • Constitutional Monarchy: both about equally

Middle class buffs are harder to get, there's a lot more traits that affect nobility or lower class, and often traits give +1 to one and -1 to the other, which is good or bad depending on which class tends to be less happy for you.

For absolute and constitutional monarchies ministers you setup stay there until natural death.

For republics you'll suffer from new government and a few turns of godawful ministers every elections, that is every 8 years / 16 turns. You should really focus on fixing that every time. If you're not confortable with this micro, constitutional monarchy will be best government type for you.

Research overflow

The best way to get ahead with research is by building or capturing more schools, and I don't bother with this micro myself, but it could be useful, especially very early game.

The game shows how many turns it takes to research something, but it's not how it works behind the scenes. It actually adds research points from a school and gentlemen in it to progress bar, and when it's full, the technology is researched. Any excess is lost.

So if technology costs 38 points, but your school generates 17 a turn, you'll do it in 3 turns, and waste 13 points on last turn.

By making a lesser school (lower tier or fewer gentlemen) do those last turns, while your good school does something else, you could reduce that waste, and research faster.

The advantage is farly small, and the game doesn't display information you need to do this micro, so I don't really recommend it as a regular thing.

Research specialization

This one is much easier to pull off. Schools don't get any traits, but gentlement in them can get traits, and researching technology in specific tree (military, industry, enlightenment) is likely to give them traits that gives them more research points of the same kind.

So it makes sense to do most of your military research in one school, most of your industry research in another, and most of your enlightenment research in yet another, instead of mixing it up.

It's a small effect, so you don't have to do it, but it takes very little effort.

Battle AI

It's very useful to know if AI will attack you or stand back in battle. You want AI to leave you alone if you're waiting for reinforcements, and you want AI to come to you if you have nice formation facing it setup.

This seems to be primarily based on two factors:

  • artillery superiority, counting only unlimbered artillery
  • if AI has any troops that can hide anywhere (mostly native American auxiliaries), it will try to stand back and ambush you no matter how much pounding it gets from your artillery


Anonymous said...

Very nice list!

I actually put all gentlemen/scholars into one college. I try to use this college for only the last turn or two of researching a tech. The idea was
to give as many opportunities as possible for gentlemen to gain traits. Other schools/universities do most of the researching and this incidentally keeps me from ever wasting many research points as I pay attention and put the scholars on the job.

taw said...

Anonymous: Nice. Microing research ending for gentlemen traits is a powerful idea, if you have patience for it.

Eric said...

Thank you for this, you may wish to add that Coffee doesn't work in the most current patch. As in, you can have 20 coffee plantations but 0 will be sold. So no matter the price, it is a DEAD trade good that has 0 production.

taw said...

Eric: I've never heard of this bug. Are you sure you don't have those plantations on islands with military shipyards?

You need trade ports to get your goods home, and a lot of Caribbean islands start with military shipyards prebuilt, making their plantations completely useless. You need to destroy these shipyards and build proper trade ports.

There are a few more locations without any way to transfer goods. Like Spanish Florida early game, which only gets port slots like 10 years in, and you need to put the proper trade ports there.

I'm guessing that's what's going on, not coffee issue.

Eric said...

Just checked it, nope. I have a coffee plantation in Panama, a trading port and 0 exports in my trade screen. Unable to post a screenshot, or I would. Coffee is bugged. Scroll down and look at all the goods this person has and look at his coffee being 0. I've seen multiple examples of it. Load up as Spain I guess and see if Panama has the same result for you.