Nation Alliances: Request, Rely, and Prepare an Exit

Direct answer: an alliance is a window, not a wall

Request a Nation alliance only when you can name the action that the temporary friendly border buys: closing another front, moving a reserve, developing a safer route, or surviving a specific attack window. Check the room mode, the Nation’s actual border, your current relation and troop position, then ask. Acceptance is conditional, and a friendly relation by itself is not an active agreement. Even an accepted alliance expires or can be broken. Keep a second route and a force that can answer the former ally before spending the entire window elsewhere.

Use PACT as an editorial decision check, not a game rule. Purpose names the measurable gain during the agreement. Acceptance asks whether a request is currently possible and whether refusal would leave you exposed. Coverage identifies the border, trade route and other players the pact actually affects. Termination records its timer and the first visible signal to disengage. A player who cannot fill all four lines should not count a pending invitation as defense. The agreement can still be worth trying, but the move after a refusal must already be affordable.

This page owns the player’s reliance decision around a Nation: whether to request now, how much to commit while friendly, and what to do before the relation stops protecting the border. Alliance and betrayal strategy owns general request timers, mutual extension, embargo and human-player exit costs. Nation mechanics is the reference for AI behavior branches. Nation pressure owns the enemy-target and expansion-stop decision once a Hard or Impossible Nation is a threat. Here those facts become a plan for the period between an invitation and a possible hostile border, not another universal diplomacy glossary.

Hard and Impossible Nations can weigh valuable bordering targets they can safely beat, and they may consider the retaliation a betrayal would invite. That does not make an ally permanently safe or make every request acceptable. Treat the relationship as conditional: its value depends on the current border, the Nation’s other targets and the force it can bring.

The simplest useful completion test is written before clicking: “If the Nation refuses, I hold this border with this reserve; if it accepts, I will finish this one task by the time remaining; if it expires or turns hostile, I retreat to this position.” This is especially important for a fragile coastal or enclosed start. A pact may buy time to get out, but it cannot create a land route, a Port, or Troops. If your entire economy depends on the Nation never changing relation, the pact is carrying more risk than the one action it enabled.

Read the invitation before treating a Nation as friendly

First separate a Nation from a human ally and a fixed teammate. A Nation follows an AI decision queue. A human can make a promise or deliberately coordinate outside the game; the Nation has no such conversation. A fixed Team relationship is not a five-minute alliance that can be extended or dissolved with the normal alliance controls. In Humans versus Nations, the room’s side assignment and alliance settings also matter. A successful request in one FFA recording cannot establish the legal or useful action in another mode. Read the mode and the relation actually displayed in your match.

An outgoing request is an offer, not an immediate ceasefire. The general request stays active for about 20 seconds and the same target has a 30-second request cooldown measured from creation. If the Nation is already threatening the border, those seconds are a real tactical interval. Hold enough Troops to survive it; do not begin a long unguarded construction solely because the button was pressed. A refused or expired invitation is not a hidden alliance. If the request succeeds, confirm the active alliance and its expiry display before moving the reserve.

The Nation decision is not a single favorable-relation threshold. Requests made during spawn can be rejected immediately. Later checks can consider confusion, traitor state, perceived threat, Team mode, relation, existing alliances, elapsed game time and relative strength. Any early acceptance roll only applies after earlier checks have passed, so a percentage from one branch is never the overall chance that a request succeeds.

Observation in the live match	What it establishes	What it does not establish
Friendly-looking relation	A relation input may be favorable	An active timed alliance or guaranteed acceptance
Request still pending	The offer has not resolved	A safe border during the wait
Alliance and expiry shown	The current agreement is active	Protection after expiry or against every third party
Nation pauses nearby	Its current queue has not produced an attack here	A promise to keep the same target and strength

The table matters because players often merge those four observations into one impression of safety. Treat the border as having 100 available defense units and the next expansion as needing 70. If the Nation refuses, only 30 remain. Requesting first is sensible only if you can delay that commitment or cover the exposed interval with terrain, another defender or retreat space. If it accepts, the agreement can release some attention, but keep the reallocation reversible. Measure the opportunity it bought against the actual timer, not against the hope of another invitation.

The reason for a refusal is usually not visible in the match. Several earlier checks may have ended the decision before any favorable relation or strength comparison mattered. Read the live border and reserve instead of trying to reverse-engineer a hidden roll. A single success is not a promise for the next request, and a single refusal is not proof that the button failed.

Spend the friendly interval on one reversible objective

The correct first move after acceptance is not “ignore that side.” It is to identify the newly freed capacity and a finish line. A pact can make one contact temporarily non-hostile, but it does not remove a neighboring human, repair an exposed City, or make a sea passage safe. Take a board snapshot: current Troops, Gold, the Nation’s contact tiles, the nearest other enemy, active attacks and the timer. Decide how much force may leave the allied border without making the return impossible. The reserve is an option to respond, not idle waste. If the Nation is the only buffer between two human opponents, its presence can also affect the path others must take; losing that buffer changes more than one edge.

Choose one task with an observable end. “Connect a second land pocket and put the City two borders behind the front” can finish before an alliance expires; “become strong enough to win” cannot. A Port can open a return path only if a legal route and a defendable destination exist. A Factory can generate useful traffic only if its network connects worthwhile stops. A defensive post is valuable only over the contact you expect to hold. The structure itself is not the completion test. The completion test is whether the position still has a route and reserve when the former ally is classified differently.

For an accepted pact lasting the ordinary 300 seconds, write the remaining time at the moment you start the task. That is the general default, not a promise for every custom room; a lobby may set a different duration or disable alliances. The final 30 seconds can offer mutual extension. Do not wait until then to start moving troops home. Extension requires both parties and a Nation’s response remains conditional. Plan the task to end early enough that refusal, an unavailable control or a changed board does not strand the investment. A deadline based on the live countdown is safer than memorizing “five minutes from my first click,” since time passed during the request.

Prioritize actions by reversibility. A small scout push on a second land route can stop if an enemy claims the target. A full troop send, a remote silo, or a Port whose only useful destination is across the allied border is harder to rescue. If the agreement is merely buying a recovery interval, keep Gold for the defense you may need at expiry. If the agreement is buying a bridge to another territory, establish the actual corridor before funding its second-stage economy. You can still invest aggressively when the board supports it; PACT asks for an exit from the investment, not perpetual caution.

The counterparty’s incentives can change while the clock runs. A Nation that was weaker when it accepted can later have more Troops, a shorter front and fewer alternative targets. The nation-pressure guide covers hostile candidate selection in detail. For the alliance decision, reassess after a major battle, a new shared border, the loss of a human buffer, an expiration warning or a visible change in the Nation’s force. A timer is not a shield against changing geometry.

Keep a visible ledger with three entries: benefit achieved, force that can return, and the state that ends reliance. At the midpoint, if benefit is still zero, stop adding irreversible costs. At the warning, if force cannot return, shorten the mission. At expiry, assess the new relation before treating an old treaty as current. That small discipline turns a diplomatic success into usable time rather than permission to forget the most dangerous side of the map.

Check strength, relation, and mode without inventing a guarantee

Acceptance and betrayal are related but distinct decisions. A Nation may decline because of an earlier eligibility or state check even when a player sees a pleasant relation. Conversely, acceptance does not freeze its later attack queue. Difficulty, existing alliances, mode and target priorities all matter. Read them as conditional filters. Neither a relation color nor a percentage from one branch is a complete probability model. Repeated clicks do not erase earlier checks or make an Impossible Nation in a Team context behave like an Easy Nation in FFA.

Avoid using troop strength as a magic safe number. Relative strength can influence the Nation’s response and later opportunity, but both counts change with normal expansion and battle. If you leave 40% of your force at home while the Nation loses half its army elsewhere, the relative balance differs from the moment you accepted. If it gains land or captures another army, the balance can move the other way. Track the visible trend and the width of the shared border. A compact boundary with an interior reserve has a different exposure from two remote protrusions, even when total Troops match.

The old story “the AI always hits the weakest neighbor” is wrong. The reverse story “an ally with more Troops always betrays me” is wrong too. Neutral expansion, retaliation, assistance and targets already under pressure can take priority. An attack may not be chosen on a given tick; a pause is not a contract. Make a plan that holds under either a quiet interval or a sudden new border rather than predicting the precise tick.

Embargoes are another separate state. A manual embargo is not simply a visual synonym for breaking an alliance, and trade networks may react differently from land attack classification. The general diplomacy guide explains the timing and exit costs. For a Nation, an embargo can change relation, and difficulty affects whether its own embargo clears automatically. Do not use an embargo as a “free warning shot” to test friendship while a trade route or second front depends on it. Verify the live relation and income columns after any trade or diplomatic action, and be ready for the route to stop paying without assuming a particular retaliatory attack.

You have 1,200,000 Troops, a Nation has 1,600,000, and a human opponent has a Port on the opposite coast. The Nation looks friendly, but no active timer appears. Classify this as “possible request,” not “safe side.” If the request fails, keep the Nation-facing reserve and challenge the Port only with surplus. If it succeeds, use the interval to make the coastal route independent, then move a reserve back before expiry. Do not derive a success probability from the 1.2-to-1.6 ratio; the other checks and the current mode still matter.

The way to improve the decision is to remove dependency, not to farm a better relation indefinitely. A defensible inland node, another exit, an allied human who can actually reinforce, or sufficient retained Gold gives the player choices after a Nation declines. If none exists, diplomacy may be the least bad gamble, but say so plainly. An uncertain agreement should never be displayed in your mental ledger as certain defense.

Scenario: a public FFA request fails at a crowded border

In a public FFA, a Hard Nation is immediately east, a human is to the west, and a neutral pocket is to the north. The Nation has 1.5 million current Troops; you have 1.0 million, a City one layer behind the eastern contact, and 600,000 Gold. The human is advancing toward the northern pocket. Your objective is to hold the City while claiming a second exit before the human does. The tempting shortcut is to request a Nation alliance and instantly send nearly every troop north.

At P, name the task: reach and hold the neutral pocket, not simply “ally with the Nation.” At A, check whether the request can be made and whether the Nation is actually evaluating it. A pending offer cannot cover the eastern border. If the Nation rejects, or the request times out, the safe branch is to keep enough defense at the City and take only the northern tiles that do not turn the east into an open corridor. A bounded probe can discover whether the human is already claiming the pocket. If the pocket is gone, do not compensate by making a larger blind commitment against the human while the Nation remains unallied.

At C, separate the two contacts. A successful pact can remove the Nation from the immediate ordinary hostile land-contact list, but it does not make the western human friendly and does not prevent a third party from using a different route. Check whether the City can still be defended from that human through the northern land. If so, move only the force needed for the contest and leave a return reserve. A City defended by an apparent ally on one side but open on the other is not safely behind the front. If a trade route also passes through the Nation relationship, record its income separately from territory access; a later embargo could remove the payout without first capturing the City.

At T, set the exit before spending. Suppose the alliance succeeds with four minutes remaining when the northern probe begins. Plan to complete the capture and restore the eastern reserve well before the final extension window. If the Nation loses a major battle, changes relative strength, or obtains a new shared border as you expand, reassess immediately rather than waiting for the clock. If the northern route proves too narrow to hold, cancel the investment and keep the City. A smaller durable border plus a reserve is a valid result even if the initial pact felt like permission to expand quickly.

The counterfactual is instructive. If no alliance is possible because the room disabled it, the whole PACT line for A is “not available,” not “try repeatedly.” If the Nation accepts but the human turns away, use the interval to make the eastern fallback stronger instead of chasing a now-unneeded race. If the Nation refuses yet attacks the other human, that is an observed opportunity, not evidence that it will continue to protect you. Take only the land you can defend when its attack ends. The scenario succeeds when the City and independent exit survive a refusal, expiry or changed priority; it does not require winning an acceptance lottery.

Scenario: a Humans versus Nations pact buys a recovery turn

In a Humans versus Nations game, your human side has two players on an island edge, an Impossible Nation controls the central mainland and a second Nation sits beyond a narrow coast. Your teammate has just lost a forward Port. You retain 900,000 Troops, one usable landing route and enough Gold for either a defensive rebuild or a risky second Port, but not both. The first Nation has 1.8 million Troops. The room’s teams, alliance options and map are authoritative, so check them before committing resources.

The purpose is a recovery interval: move your teammate’s surviving force out of the exposed landing, restore one protected Port or inland defensive node, and preserve a launch route. Do not describe the purpose as making the Impossible Nation “on our team.” Fixed team identity and a timed Nation alliance are different state. Check whether the current mode even permits the request and whether another earlier rejection check applies. Team mode has an additional refusal path that can be decisive at higher difficulty. If the displayed request is impossible or repeatedly rejected, stop budgeting the recovery around it and spend on the protected route you can control.

Suppose an agreement does become active. The first action is to move survivors to a line the teammate can help hold, not to build the risky second Port behind a single temporary Nation buffer. The Port is worthwhile only if it reconnects a real sea route and can be defended after the agreement. Watch the current alliance countdown and the Nation’s troop balance as the other Nation fights it. An apparently convenient Nation-on-Nation battle can make one side stronger and change which former ally becomes a valuable neighboring target. No fixed number of ticks certifies a safe window.

Use two budgets. Reserve A is the Troops and Gold needed to keep the island landing and home route alive after refusal. Budget B is the surplus that can rebuild the lost forward position while the alliance is active. If the recovery costs all of A, the pact is masking a forced gamble rather than financing a recovery. A narrow landing may also bottleneck reinforcements; the Transport landings guide covers route and wave execution, while Team roles covers who guards, earns and counterattacks. This page decides whether Nation diplomacy justifies moving those resources at all.

At the warning, the human teammate can prepare independently: move ships away from a hostile coast, keep an inland troop reserve, and identify the attack to cancel if the relation changes. A request to extend does not finish the job until both sides agree; a Nation’s response is still conditional. If the Port is not restored by that time, abandon the second-Port branch and secure the first route. If a Nation becomes hostile earlier, respond to the live border rather than arguing from the old pact icon. A good outcome is not necessarily victory: a rescued teammate, one defensible route and Gold left for the next turn meet the recovery purpose. A pact that expires with two exposed Ports and no reserve failed that purpose even if the invitation succeeded.

Prepare for expiry, refusal, and a selective betrayal

There are three different exits. A request can be rejected or expire without an alliance ever starting. An accepted alliance can end at its configured expiry unless mutually extended. The relation can also become hostile through a break or betrayal. Do not compress these into “the Nation attacked me.” The response depends on which state changed and when. A pending request needs ordinary border defense; expiry needs a planned return and a fresh classification; a sudden hostile transition needs immediate attack and route inspection. Preserve a screenshot or replay timestamp when diagnosing which occurred, but act on the live board first.

Selective betrayal is not immunity. Hard and Impossible Nations can assess an attractive allied target and the retaliation they would face; low allied Troops, traitor status and a uniquely weak bordering ally can also matter. These are conditions for consideration, not a countdown that guarantees an attack on crossing one ratio. A Nation already fighting another enemy may be quiet while that priority wins, then reconsider when its strength or options change. The player-controlled answer is a shorter exposed border, a reserve, and a route that does not require continuing permission from the former ally.

Do not voluntarily break a live alliance merely to simplify the map. The general diplomacy rules can impose a temporary traitor state after a manual break; natural expiry is different. If you plan an attack on the former ally, note whether the agreement is still active and whether a legal break carries a cost. A nuclear strike is not a loophole: alliance acceptance can remove relevant in-flight weapons and can block launches for a short interval afterward. “Accept now, immediately fire” is therefore an unsafe assumption. The Diplomacy and betrayal guide owns the complete timer and nuclear exit mechanics; use it before a deliberate hostile turn.

Counterplay from other players also changes the value of the pact. A human can take the neutral pocket you intended to reach, cut a rail connection, or pressure your exposed coastline while you watch the Nation. Another Nation can win a nearby battle and become the new contact. A teammate may expect you to cover a Port when you have already moved away. Keep the PACT purpose visible to your team and abandon it when the route is no longer independently useful. “We have an alliance” is status; “we can hold this bridge after it ends” is a plan.

At the first expiry warning, stop initiating an attack that commits the return reserve for longer than the remaining window. Check the opposing force on the former allied border, whether your structures have a second defensive layer, and whether the planned retreat has a legal land or water path. Ask for extension only if you can survive a refusal. If the Nation remains friendly after expiry, that is useful current information, but do not label it a renewed treaty without an active timer. If it turns hostile, cancel an unproductive expansion, defend the shortest route to the economic core, and reassess rather than reflexively attacking the entire Nation.

After the match, the right replay question is not “why did the AI hate me?” Mark the first moment at which you treated a pending request as protection, moved a reserve beyond return range, missed the expiry, or let a new border remove your fallback. Replay review separates what you knew then from what the replay later reveals. That distinction helps test one change next game: request earlier with a real refusal plan, complete a smaller task during the agreement, or save the return reserve. It cannot recover a private AI random draw from a video alone.

Modes, maps, and a repeatable next-match test

An ordinary FFA, a fixed Team game, and Humans versus Nations do not give the same alliance decision. In FFA, a Nation can be one temporary buffer among multiple human threats; the pact’s value is the independent route you create while it lasts. In Team, shared defense and victory thresholds change the opportunity cost of removing troops from a teammate’s front, and the Nation acceptance behavior has extra mode checks. In Humans versus Nations, verify side assignment and room settings instead of assuming that a Nation’s willingness in one custom match carries into another. A custom host can change alliance duration or disable the feature. Always begin with the live lobby.

Map shape changes the consequence, not the underlying acceptance rule. On a broad continent, a second inland lane can be established during a pact and held after it. On an island, a friendly Nation across water may not even be the direct bordering candidate you expect, while your only Transport route can be cut by a Warship. In a corridor, the former ally may control the single return edge; a treaty that ends before you establish an alternate exit is especially dangerous. Do not import a Europe FFA river crossing, a Port count or a video timestamp into another map. Read land contact, actual ship route and current owners.

During a match, use visible state: the request timer, the active alliance countdown, the current relation, the shared border and the forces that can return. When a recording or replay says that a Nation “agreed” but the interface shows no active timer, trust the interface. A successful ending cannot prove that the pact caused victory rather than a favorable human border elsewhere.

Players often ask about unexpected refusals, Nations used as buffers, and an alliance made while another route is contested. Those situations have the same practical answer: name the route or defense the pact is buying, then keep a plan that works after refusal or expiry. Do not turn a single replay into a universal opening.

For the next match, perform a small falsifiable test. Before sending a request, write the purpose, the border that remains defended during the pending interval, and the exact fallback after refusal. On acceptance, note the displayed expiry and invest in one task that can finish early. At the midpoint, record whether the task improved an independent route. Before expiry, bring back the reserve and decide whether extension is worth asking for without depending on approval. Afterward compare the planned and actual state. If the pact failed, change one controllable choice next game rather than claiming that a conditional AI gate is deterministic. That is a decision you can repeat across modes without pretending any Nation owes you an alliance.