City-States and Envoys: Where Your Influence Should Go

Spending the First Three or Four

The first envoys are worth more than any later ones, because the one-envoy tier is the cheapest bonus in the game and it lands in your capital immediately. One envoy in a scientific city-state adds science to your capital and every Library; the cultural, religious and trade versions do the same through Amphitheaters, Shrines, and Markets or Lighthouses, while militaristic and industrial ones add production toward units, or toward buildings, wonders and districts, in the capital and in cities with the matching building. On turn thirty your capital is most of your empire, so a single envoy in each type you have met is a real fraction of your economy.

So the opening pattern is breadth before depth: one envoy into each type you have met, prioritised by what your capital is about to build. If a Library is queued, the scientific city-state goes first; if a Shrine and a religion are the plan, the religious one does. Only after that first spread should you stack toward three, and the target should be the city-state whose suzerain bonus you actually want, not the one that already has the most envoys.

Two adjustments change the pattern. With the Diplomatic League card slotted, every first envoy counts as two, so the breadth phase is the time to run that card and the moment you begin stacking is the moment to swap it out. And when a militaristic city-state sits between you and a hostile neighbour, early envoys there are about earning the suzerainty needed to levy its units later, so that stack should begin before the war does.

Which Types for Which Victory

The three-envoy and six-envoy tiers pay through buildings, and that decides which types are worth stacking. Three envoys in a scientific city-state add science to every University and Consulate; six add more to every Research Lab and Chancery. The cultural ladder runs through Art Museums, Archaeological Museums and Broadcast Centers; the religious one through Temples and worship buildings; the trade one through Banks, Shipyards, Stock Exchanges and Seaports. Industrial and militaristic tiers run through Factories and Power Plants, or Armories and Military Academies. If your plan never builds the buildings, the upper tiers are worth almost nothing to you, whatever the suzerain bonus says.

That gives a clean mapping. A science plan stacks scientific city-states first and industrial second, because Universities and Factories are on the build order anyway, and it treats Geneva, whose bonus raises science in every city while you are at war with nobody, as the best suzerainty on the map. A culture plan stacks cultural and trade city-states, since museums and Banks are its core buildings and gold buys the theatre infrastructure; Kumasi, which pays culture on every trade route sent to a city-state, is worth suzerainty for a culture player with spare traders. A religious plan stacks religious city-states, and a domination plan stacks militaristic and industrial ones, with special attention to whichever militaristic city-state can be levied against the first target.

The exceptions are suzerain bonuses strong enough to override type. Auckland's production on shallow water tiles is worth suzerainty for any coastal empire, best claimed before the coastal cities grow. Nan Madol's culture for districts on or beside water does the same for a culture player with a coastline. Hattusa's steady supply of strategic resources you have revealed but not improved can carry an early war when you lack iron. These are the cases where a type you would normally ignore deserves three envoys.

• Science: scientific first, industrial second; Geneva above every other suzerainty if you can stay out of wars.
• Culture: cultural and trade; Kumasi for empires with idle trade route capacity.
• Domination: militaristic and industrial, with the militaristic stack timed to be levy-ready before the first war.

The 1/3/6 Ladder Versus Suzerainty

Envoy tiers and suzerainty are two different rewards paid from the same pile. The tiers are yours forever at one, three and six envoys, no matter what anyone else does. Suzerainty is relative: you hold it while you have at least three envoys and more than every other civilization, and a tie leaves the seat empty. Suzerainty is what unlocks the unique bonus, the city-state's luxury and strategic resources, the right to levy its army, diplomatic favor each turn, and the city-state's participation in your wars. Losing it costs you all of that at once, while the tier bonuses stay.

This creates two different kinds of city-state in your envoy plan. The first is a tier target, where the bonus you want lives on the ladder: three envoys in any scientific city-state pay the same University science, so pick the one nobody else is contesting and stop at three. The second is a suzerainty target, where the bonus you want is unique: for Geneva or Auckland the tier bonus is a side effect, and the number of envoys you need is whatever beats the AI, which may be three or may be nine. Decide which kind each city-state is before stacking; the most common waste is five envoys in a city-state whose suzerain bonus you never needed.

Six is a real threshold, not a vanity number. The six-envoy bonuses attach to late buildings and multiply across many large cities, so a science empire with ten Research Labs feels the sixth envoy in a scientific city-state more than it felt the first. The right time to push from three to six is the era when the relevant buildings start appearing, not earlier.

Out-Competing the AI, and When to Let One Go

The AI does not spread envoys thoughtfully. It tends to pour them into the two or three city-states closest to its capital, which means the contested seats are usually predictable from the map on turn twenty. A city-state wedged between you and a neighbour will be fought over; a city-state on your far side, out of the AI's reach, will not. Uncontested city-states are cheap suzerainties you can take late at exactly three envoys; contested ones need an early decision, before the AI's lead outgrows what a quest or two can close.

Winning a contested seat is about concentration and timing. Save envoys and drop four in a single turn to take the seat before the AI reacts; trickling them in invites a one-for-one race you cannot win on influence alone. Amani is the tiebreaker; assigning her to the contested city-state adds two at once, and her doubling promotion turns a four-envoy stack into eight. Quests in that city-state should be completed even when inconvenient.

Letting go is a skill of its own. When an AI sits five or more envoys ahead in a city-state whose suzerain bonus was never central to your plan, the envoys you would spend catching up are worth more as three-envoy tiers in two uncontested city-states. Envoys already placed are not wasted, since the tier bonuses remain, so the loss is only the seat. Let it go, and remember that suzerainty changes hands whenever the count does; a war that costs that AI its envoys can reopen the door later at a fraction of the price.

Protecting a City-State

A suzerainty is an asset, and assets get attacked. The first line of defence is the levy: as suzerain you can pay gold to take control of the city-state's entire military for thirty turns, and those units fight as yours, gain your promotions and can be moved into the city-state's own walls. Levying is cheapest early, when the army is small, and it doubles as a reserve for your own wars, which is why militaristic suzerainties near a border are worth holding even when their tier bonuses are marginal.

When a neighbour declares war on your city-state, you have three options in rising order of cost. Levy its army and let the city-state defend itself with your promotions, which is usually enough against a lazy AI attack. Move your own units into its territory; as suzerain you have open borders there, and a couple of ranged units on its walls stop most sieges cold. Or declare war on the attacker. The Diplomatic Service civic adds a casus belli specifically for defending a city-state you are suzerain of, which cuts the grievances the war costs you and makes this the correct answer against a warmonger who would otherwise eat the city-state and then turn to you.

Prevention is cheaper than any of those. Alliances do not shield a city-state directly, but an allied neighbour will not attack one you hold, so an alliance with the most aggressive civ next to your city-states removes the likeliest aggressor. A modest garrison of your own nearby also changes the AI's calculation, since it weighs local strength before declaring, and one Encampment city in the middle of a cluster of suzerainties protects all of them at once.

When Conquering a City-State Is Right, and What It Costs

Taking a city-state is occasionally correct and usually expensive. The bill has four lines. Every envoy you had there is lost the moment you declare war, and every civilization with envoys there resents you. Grievances pile up with the whole table, more so if another civ held suzerainty, and the conquest can trigger an emergency that hands your rivals a shared war goal. The unique suzerain bonus disappears for everyone, including you, for the rest of the game. And the captured city is a modest one, with no district plan and unhappy citizens for a long while.

Against that, the cases where conquest wins. The clearest is a city-state whose land you need more than its bonus: one sitting on the only river valley near your capital, or holding the mountain range your second Campus was planned around, is worth more as your city than as anyone's suzerainty. The second is a city-state already lost to a rival suzerain who is using it against you, especially one that keeps harassing your traders and settlers. The third is civilization-specific: Germany under Barbarossa fights city-states with a combat bonus and wants the extra city for its extra district slot. Do it early, before walls go up and before the table has invested envoys there, and the grievances fade within an era.

If you do take one, take it cleanly: declare with a casus belli you can justify rather than a surprise war, finish the siege in a few turns so the emergency window is short, and never raze it, since a razed city-state gives you nothing for the same grievances. Liberating a city-state that someone else captured is the opposite play and often the better one: it restores the city-state and your envoys there, earns goodwill with everyone who had envoys, and typically leaves you as suzerain.

• Conquer when the land itself is the prize, when the city-state has become a rival's weapon against you, or when your civ pays a discount on the fight.
• Never conquer for the tier bonuses you already hold; they vanish with your envoys the turn you declare.
• Prefer liberation to conquest whenever a rival has already done the dirty work.

A Worked Example: Six Envoys by Turn Sixty

Take Korea under Seondeok on a Pangaea map, aiming at science. By turn sixty the scouts have met four city-states: Geneva to the north, first met by you; Hattusa to the east, already holding two envoys from a Persian neighbour; Nan Madol on the southern coast; and Auckland on a bay your third city will share. Six envoys have arrived, two already placed: one in Geneva from the first-meeting bonus, one in Hattusa from a Seowon quest completed without noticing. The capital has a Library.

The four unplaced envoys go like this. Two more into Geneva, reaching three: that is suzerainty, because nobody else has met it yet, and its bonus is the one a peaceful science game wants above all others. One into Auckland for the one-envoy industrial tier, which speeds every district in the capital now and marks the city-state as a future suzerainty once the coastal cities exist. One into Nan Madol, whose single-envoy culture is worth more to a science player than it looks, because civics unlock the Seowon's doubling card and the governments that make influence faster. Hattusa is left alone: Persia is two ahead, its bonus is a war bonus, and the tier reward for a scientific city-state is already covered by Geneva.

From there the plan writes itself. Amani, once earned, goes to Auckland the turn the third city settles, taking suzerainty on the spot. The next three envoys push Geneva to six as Universities appear, or answer Persia if it starts contesting Geneva rather than Hattusa. And if Persia ever attacks Geneva, levy first and declare the protectorate war second; a science game that loses its best suzerainty on turn ninety has lost more than the war would cost.

Related pages

• Geneva (scientific)
• Auckland (industrial)
• Kumasi (cultural)
• Apadana placement
• Kilwa Kisiwani placement
• How District Adjacency Works