OpenFront Nation Strike Protocol: Surviving the v35 Nuke-Then-Attack AI

Direct answer: don’t be the cheap target inside the Nation’s strike band

Against a Hard or Impossible Nation in v0.35.0-beta1, stop expanding the moment you sit inside its strike band — the range where your troops are about half of the Nation’s — because it will soften you with a nuke and then attack the tick the bomb lands. Read its border Defense Posts as a telegraph, keep your valuable structures behind depth rather than on the contact line, cover your core with SAMs so the blast is wasted, and counter only when the Nation has already committed troops. In one line: do not present a cheap, close, high-value target. That single habit is what the v35 Nation is built to punish, and the whole guide below is the discipline around it.

That one habit matters more than any individual counter. The v35 Nation is not a fixed script that always picks the weakest neighbor; it is a sequence of human-like decisions that ends in a coordinated nuke-and-attack. If your expansion keeps your own strength hovering right at that half-troops line, you are literally handing it a target that clears its eligibility test. If you pull back, keep value behind a buffer, and let the blast fall on empty or worthless land, the follow-up attack has nothing left to finish. Everything else in this guide — the blast math, the relations change, the Defense-Post telegraph, the bot-annexation economy — is context for when and how to apply that one habit. Use the S-T-R-I-K-E framework below before you make the next capture: check the Strength band, watch the Telegraph, remember Relations no longer punish a counter, Insulate your value, Kill the follow-up stack, and pick your Exit or counter.

Each letter is a check you run before the next capture. Strength: if taking the next structure would put you at or above half the Nation’s current troops, stop — you are now eligible for the strike. Telegraph: if the Nation is placing Defense Posts along that border, expect the strike there within minutes. Relations: a defensive counter costs nothing in escalation on Hard and Impossible, so hit the incoming stack, but do not convert it into a war of attrition. Insulation: move the next high-level structure you build one buffer behind the contact line, not onto it. Kill the follow-up: keep a declared reserve sized to contest the Nation’s post-blast stack, because the bomb spends its payload and leaves it thin. Exit: if you are already inside the band with exposed value and a live telegraph, withdraw behind the buffer and tighten the SAM core before the Nation’s save matures. If three or more of the six checks read ‘bad’ on a position, that structure is not worth the capture; take the cheaper one on the next border instead.

What v35.0-beta1 changed and the version boundary you are reading

This guide is written specifically for the v0.35.0-beta1 beta, which shipped the Nation AI overhaul as PR #5899, titled “Nations play more like winning humans.” The beta tag is the exact point where the behaviors described here live: the tag commit is ab5cc61, and the Nation AI commit fac309acc sits sixteen commits below it, so the constants quoted in this guide are the tagged source, not a draft or a backport. Treat this as a beta boundary, not a final one. OpenFront is in Early Access and the v0.35.0 line can still change between this beta and the final release, so if a number you see in-game disagrees with this guide, trust the game and the newer release note, then report the drift.

The overhaul is broad, but the decision-relevant core is one: Hard and Impossible Nations now run a strike protocol. They select a bordering rival that is about as strong as themselves, drop a nuke on it, and launch a follow-up attack the tick the bomb lands, so the explosion does the damage and the attack only has to mop up. Around that single change the PR ships four companion behaviors that re-price your counterattack and your exposure. First, relations: fighting back against a Nation no longer angers it on Medium, Hard, or Impossible (only Easy keeps the old escalation), and only manual embargoes lower a relationship, which makes disciplined counters and trade disruption far cheaper than in v34. Second, defense posts are now placed far enough behind the front that they finish building in time, and Hard and Impossible additionally fortify a hostile border before they attack it — a readable warning. Third, bot annexation: Hard and Impossible take Bots in an order that walls other players in, so those players get annexed for free, and they send only the troops a Bot attack needs. Fourth, the economy: Nations save later, keep building income while saving, refuse to save for a MIRV their SAMs would stop, bias toward ports on coastlines, and buy several upgrade levels at once. Hydro-only Nations are gone, and the Impossible leader no longer nukes player number two just for the slot. The deltas that matter for your decisions, side by side:

Behavior	v34	v0.35.0-beta1 (#5899)
Escalation from countering	Angers the Nation on every difficulty	No anger on Medium, Hard, or Impossible; only Easy keeps it
What lowers relations	Passive trade could degrade them	Only manual embargoes lower them
Defense Post timing	Often finished too late to defend	Built in time; Hard and Impossible fortify the target border before attacking
Nuke sequence	Attack on its own schedule	Bomb-then-attack the tick the bomb, or the last MIRV warhead, lands
Bot attacks	A full force	Minimum troops needed; on Hard and Impossible the order walls other players in
Nuke-back	None or rare	Nations nuke back the players whose bombs are currently flying at them
Economy while saving	Stopped building income	Keeps building income, saves later, buys several upgrade levels at once
Hydro-only Nations	Possible	No longer exist

Every section below is a player-facing read of one of those changes.

The strike protocol: about-as-strong, soften, then attack on landing

The heart of the change is three numbers that together define a strike band. A Nation only considers a strike when the target has at least half of the Nation’s own current troops — the source constant is STRIKE_MIN_TROOP_RATIO = 0.5. That is the “about as strong as them” line. A tiny neighbor you can walk over does not qualify for a strike; it falls to a plain land attack instead. So a Nation that is five times bigger than you is, on this test, not about to nuclear-strike you for being small — it is more likely just to expand into you. The strike is reserved for a rival whose size is comparable.

The second number is the follow-up requirement. On Hard and Impossible, even a target that clears the half-troops line is only struck if the Nation can field a follow-up attack worth at least half of the target’s troops — STRIKE_MIN_STACK_RATIO = 0.5. This is why a strike is a commitment, not a spam: the Nation checks that the post-blast attack still has teeth. If the Nation is committed elsewhere, low on free troops, or the blast would leave it unable to send that follow-up stack, it holds the bomb. The third number is spacing: the Nation keeps a 10-tile margin between the blast and its own land (STRIKE_BORDER_MARGIN = 10) so that the front it uses for the follow-up attack survives the explosion and the troops still reach you. A strike also waits about ten ticks (STRIKE_SPAWN_TICKS = 10) for the bomb to spawn before it is dropped.

Put the three together and you can read a strike as a two-stage action you can partially interrupt. Stage one is the bomb, which does the damage but exposes nothing of the Nation’s troops. Stage two is the follow-up attack the tick the bomb (or, for a MIRV, the last warhead) lands. The Nation’s own comment in the source is blunt: a MIRV leaves the launching player “next to no troops to defend with,” which is exactly why the follow-up is what you must disrupt, not the bomb itself. Your counter is therefore aimed at stage two: deny the follow-up stack, or make the blast land somewhere its damage does not open a gap the follow-up can exploit. A MIRV behaves the same way but at higher stakes: because a MIRV spends its payload on warheads, the launching Nation arrives with almost no troops behind it, so a well-timed counter into the emptied border can win the exchange before the last warhead even splits. This is the single most important reframe from v34: the bomb is a setup, and the follow-up is the kill.

Read the blast: coverage, margin, and what a wasted bomb tells you

A Nation does not nuke you just to hurt you; it nukes you when the blast pays. The source encodes a minimum share of the blast that must fall on enemy land — MIN_BLAST_COVERAGE — and it differs by bomb: an Atom Bomb needs at least 0.25 (a quarter) of its blast on your land, while a Hydrogen Bomb needs at least 0.4. Below those thresholds the Nation prefers a land attack, because the explosion would mostly hit water, fallout, or neutral land and the follow-up would have nothing to finish. This is a concrete lever for you: if your exposed footprint is small and mostly behind open water or fallout, a big Hydro on you is a poor trade and the Nation will often choose not to spend it. A blast that hits nothing but enemy land earns a large bonus in the Nation’s tile score (FULL_BLAST_COVERAGE_VALUE = 50_000), which is why a dense, solid, mostly-enemy border is the worst place to be.

Two more numbers make the blast math actionable. The first is the crown rule: when a player holds more than half of the non-fallout land (CROWN_NEAR_WIN_SHARE = 0.5), every FFA Nation — regardless of difficulty — nukes the crown position. If you are one big expansion away from winning the map, assume the bomb is coming to the crown and plan a backup position before you cross that line. The second is density: a target needs a minimum sum of structure levels (MIN_LEVEL_SUM_FOR_HIGH_DENSITY_NUKE = 5) to be treated as a high-density nuke target. That rewards concentrating a small number of high-level structures behind cover rather than spreading many low-level ones across the border — a spread of level-1 structures is cheap for the blast to clear and gives the Nation a satisfying full-coverage hit, while a compact core of high-level structures behind a SAM umbrella is the version that makes the blast less worthwhile.

Use the blast read as a wasted-bomb detector. If a Nation commits a nuke but the blast mostly misses your structures — water, fallout, or unowned land eats the radius — the follow-up is likely to fizzle, because the Nation chose the bomb on the assumption that a large share of it would open your land. When you see the blast land short of that coverage, that is your cue that the follow-up is weak and a disciplined counter into the soft front can be decisive. Conversely, when the Nation holds its bomb even though you are inside the strike band, read that as a good sign for you: the Nation computed that the blast would not cover enough of you to justify the commit, and it is probably choosing to expand or to wait. The blast math is the Nation’s own cost function, and it leaks a fair amount of intent into the map.

The relations change: countering no longer angers a Nation

Before v35, a big fear in Hard and Impossible games was escalation: you counter a Nation’s push, and the Nation treats you as a priority and doubles down. That dynamic has changed. In the #5899 overhaul, fighting back against a Nation no longer angers it on Medium, Hard, or Impossible — the old “anger” escalation is effectively removed on those difficulties, and only Easy keeps the behavior. Separately, the only action that now lowers a Nation relationship is a manual embargo: a passive trade stance does not degrade relations the way it could before. The practical consequence is that disciplined counterattacks and trade disruption are now much cheaper than in v34. You can hit a Nation’s incoming attack and reclaim a border, and you no longer have to assume it will pivot to you out of spite.

This does not mean countering is free, and it does not mean the Nation stops targeting you. The Nation still runs its normal priority list — retaliation against the player currently hitting it, very-weak targets, already-pressured victims, alliances, betrayal, and the new strike branch. If you are also the current attacker, you are the retaliation target by definition; the change is that a defensive counter or a trade disruption does not add a hidden resentment multiplier on top. The guide to use is: counter when you are the defender or when you can afford to be briefly the attacker, but do not chain counters into a prolonged offensive that keeps you the Nation’s live target.

There is a second relations-side behavior that matters a lot for nuke defense. In v35, Nations nuke back the players whose bombs are currently flying at them. That is a direct, in-flight retaliation, not a personality trait. If you launch a MIRV or a Hydro at a Nation, expect a return nuke aimed at you while your bomb is still in the air, on every difficulty where that Nation can field one. The source is specific that, while it is saving up for its own MIRV, a Hard or Impossible Nation only nukes back the players actually hitting it, and does so with Atom Bombs. So the window where you are least likely to draw an immediate nuke-back is before the Nation has a live nuke ready; the window where you are most likely to draw one is the moment you launch into it. Time your nukes for when the Nation’s nukes are committed or unavailable, and never assume a launch is a one-way trip.

Defense posts as a telegraph and your SAM exposure

The most readable early-warning signal in the overhaul is the Defense-Post telegraph. The source places Nation Defense Posts far enough behind the front that they finish building before the attack they defend, and Hard and Impossible additionally fortify the border with a hostile neighbor before they attack it. Translation for the player: if a Nation starts spending Defense Posts along the exact border it is about to hit, a strike or a major attack is imminent on that border. Watch the build queue and the placement, not just the troop counts. A Nation that suddenly puts up a line of Defense Posts facing you is telling you, in map-space, that this is the direction of its next big commitment.

The flip side of that same change is your own SAM exposure, and the numbers here are the ones to memorize. The SAM launchers a Nation builds per City scale with difficulty — SAM_RATIO_BY_DIFFICULTY: Easy 0.15, Medium 0.20, Hard 0.25, Impossible 0.30. So an Impossible Nation is running a meaningfully denser SAM umbrella than a Medium one, and it upgrades those SAMs once every one of its structures is already under cover (the SAM_UPGRADE_WHEN_COVERED behavior) rather than adding more spread. The implication for you: a late-game Impossible border is a wall of upgraded SAMs, and a nuke on the Nation’s own core is a poor trade against it. Your own SAM budget should be sized to waste the Nation’s blast, not to mirror the Nation’s count. Concentrate SAMs on the compact core of high-level structures behind your buffer line, so that when the Nation drops a bomb, the blast is intercepted or lands short and the MIN_BLAST_COVERAGE test fails.

Two related economy numbers sharpen the exposure window. While a Nation is saving up, the perceived cost of its income structures is scaled down by difficulty — INCOME_STRUCTURE_SAVING_WEIGHT: Hard 0.6, Impossible 0.4 — meaning Hard and Impossible keep building income even while saving for a big target, so their gold and structure count keep climbing during the save. And a Nation holds at most 3 Missile Silos (MAX_MISSILE_SILOS = 3), with the first one at a 0.4 City ratio (FIRST_MISSILE_SILO_RATIO = 0.4). So you can cap your nuke-defense planning: a Nation will never field more than three silos, and the first silo appears fairly early relative to its city count. Build your SAM coverage to the point where the Nation’s first or second silo is not worth firing, because past that it has almost no nuke stock left.

Bot annexation and the economy you are reading against

The bot-annexation change is about who else the Nation is eating while it builds toward you, and it changes the board geometry. On Hard and Impossible, the Nation takes Bots in an order that walls other players in, so that those walled-in players get annexed for free once they have no open border. It also sends only the troops a Bot attack needs rather than a full force. For you, this has two consequences. First, the “safe” neutral Bots you might be grinding are often being consumed by the Nation in a specific sequence that leaves you walled in behind it — so if you and the Nation are competing for the same Bots, the Nation may be deliberately letting one Bot survive to wall you. Read the Nation’s Bot order as a map of the borders it wants closed. Second, because the Nation spends only the minimum troops on Bots, its free-troop pool for the real strike against you stays larger, which is exactly the pool the STRIKE_MIN_STACK_RATIO follow-up test draws from.

The economy change is the other half of the board you are playing against. The Nation now saves later and keeps building income while saving, which means its opening is more expansion-focused and its save window is later than in v34 — your danger is not the first minutes but the mid-to-late game when its save matures. It refuses to save for a MIRV that its own SAMs would stop, so a Nation you are already covering with SAMs is less likely to invest in a MIRV aimed at you. On a coastline it cuts back less on factories in favor of ports (the coastal factory multiplier rises with difficulty: Easy 0.33, Medium 0.45, Hard 0.55, Impossible 0.60), and in team games it builds more ports outright (TEAM_PORT_RATIO = 1 versus 0.75 off-team) because teammates are safe trade partners. Buy-several-upgrade-levels-at-once and the stop-stalling-with-high-starting-gold fixes mean the Nation’s structure levels jump in bursts rather than slowly, so a sudden level-up wave from the Nation is a signal that its save is maturing and it is about to become a live strike threat. The practical read: track the Nation’s gold-to-structure ratio and its coastline port/factory mix, and treat a burst of multi-level upgrades as the moment to tighten your own buffer before it crosses into the strike band you must avoid. Treat a save that is still open as a window to build your own buffer and SAMs, not a window to expand, because the Nation’s gold and structure levels keep climbing while it saves and the moment it closes the save is the moment it becomes a live strike threat against your current, not your projected, strength.

Scenario 1: Hard Nation on a FFA border with about half your troops

Set the scene with concrete numbers. You are in a FFA game against a Hard Nation. You have 1,200,000 current troops; the Hard Nation has 2,400,000. You sit at exactly half — you are inside the strike band, because STRIKE_MIN_TROOP_RATIO requires the target to have at least half the Nation’s troops, and 1,200,000 is precisely half of 2,400,000. You have four level-3 Cities and two level-2 Ports on the border, a compact core, and a clean land contact to the Nation. The Nation has 2,400,000 troops and, having just cleared its Bot order, has a large free-troop pool. Assume it has one Atom Bomb live and no other nukes.

Run S-T-R-I-K-E on this position. Strength: you are dead-center in the band — bad. Telegraph: the Nation has just placed a line of Defense Posts along your contact line, which is the fortify-before-attack signal — bad. Relations: you have not attacked it and you are not under embargo, so there is no escalation penalty if you defend — neutral, but it means you can counter cheaply. Insulation: your four level-3 Cities and two Ports are on the contact line, not behind depth — bad, they are exactly the high-density, solid, mostly-enemy blast the Nation wants. Kill the follow-up: the Nation can field a follow-up of well over half your 1,200,000, so the follow-up test passes — bad. Exit: the honest move is to not be here. The correct play is to stop this border cold: pull the border Cities and Ports back one buffer behind the front before the Nation’s save matures, concentrate SAMs on the compact core so the Atom fails the MIN_BLAST_COVERAGE test (an Atom needs a quarter of its blast on your land), and keep a declared reserve so that when the Nation commits its follow-up you have troops to contest the soft front. If you cannot withdraw, the fallback is to make the blast miss: keep the exposed footprint mostly behind open water or fallout so the full-coverage bonus the Nation scores never materializes, and hold your counter for the tick the follow-up arrives, when its troops are committed and its own front is the 10-tile buffer you can punch through. The scenario’s lesson: sitting at exactly half the Nation’s strength with your best structures on the contact line is the single worst position v35 punishes.

Scenario 2: Impossible solo / HvN — denying the MIRV follow-up

Now the harder endgame. You are in an Impossible FFA or a Humans-vs-Nations game. You are down to the last two major players; you and the Impossible Nation. You have 900,000 troops; the Nation has 1,800,000 — again at the half line. The Nation has finished building its structures: at most 3 Missile Silos, and it has committed a MIRV at your border. Its SAM umbrella is dense (0.30 SAM per City, upgraded once covered), so a nuke on its core is a bad trade. The MIRV is the threat: a MIRV spends its payload on warheads, so by the time the last warhead lands the Nation has “next to no troops to defend with” — but the follow-up attack is launched the tick the last warhead lands, and on Impossible the STRIKE_MIN_STACK_RATIO still requires it to field half your 900,000, i.e. 450,000, which it can still muster because it saved its core troops.

The denial playbook here is different from Scenario 1. You cannot out-SAM a MIRV’s warheads individually; you must deny the follow-up stack and the landing geometry. First, keep your value compact and behind depth so the MIRV’s warheads, which you cannot fully block, land on low-value or partially-worthless ground, degrading the FULL_BLAST_COVERAGE the Nation scored the strike on. Second, and most important, keep your own 900,000 not committed to the exposed border: hold the majority as a declared reserve behind your SAM core, so that the moment the last warhead lands and the Nation’s follow-up leaves it with a thin front, you launch into the softened Nation border rather than the MIRV. Because the Nation has spent its nuke payload, its nuke-back is exhausted for this exchange, and because a MIRV leaves it with almost no troops, your 900,000 reserve is often the stronger force in the follow-up fight. Third, use the relations change to your advantage: your counter is a defensive counter into a Nation that is committed, and on Impossible that does not add an escalation penalty — you are not “angering” it, you are hitting it when its follow-up stack is already spent. The scenario’s lesson: against an Impossible MIRV you do not win by blocking the bomb; you win by keeping your troops whole, making the blast land short, and striking the emptied border the instant the follow-up commits. If the MIRV splits into several warheads, prioritize holding the border you will actually defend over holding every exposed structure, because a MIRV’s payload can cover more ground than a single bomb and your reserve only needs to beat the follow-up on the one line you commit to; the other lines can fall as long as the compact core behind the SAM umbrella survives the landing.

Failure modes and the counter-moves that fix them

The most common v35 failure is expanding to the exact half-line — you push your troops right up to half the Nation’s size because “one more City” feels safe, and now you are inside the strike band with your strength pinned there. The fix is the stop-line discipline from the framework: define in advance the troop ratio at which you stop the border (well below half), and treat crossing it as a hard stop, not a soft one. The second failure is leaving your best structures on the contact line. The v35 Nation scores a full-coverage blast on a dense, solid, mostly-enemy border, so your high-level Cities and Ports belong behind depth, not on the edge. The fix is the compact core behind a buffer: keep a small cluster of high-level structures inside your SAM umbrella and let the border be low-value. The third failure is ignoring the Defense-Post telegraph. A Nation that puts up a line of Defense Posts facing your border is committing to that direction; the fix is to treat that placement as an alarm and either withdraw the exposed value or pre-position your SAM core before the strike matures.

The fourth failure is assuming a nuke is a one-way trip. In v35 the Nation nukes back the players whose bombs are flying at it, so launching into it draws a return nuke. The fix is to launch only when the Nation’s nukes are committed, exhausted, or unavailable, and to keep a backup position so a nuke-back does not delete your crown. The fifth failure is misreading the relations change as “counters are free.” A defensive counter is cheap; a chained offensive that keeps you the Nation’s live attacker is still the retaliation target by definition. The fix is to counter and disengage, not to convert a counter into a war of attrition. The sixth failure is over-investing in SAMs to mirror the Nation. You do not need to match its 0.25–0.30 per-City count; you need just enough SAMs to make the Nation’s blast fail MIN_BLAST_COVERAGE, so concentrate them on the compact core rather than spreading them thin. And the seventh, most subtle: trusting the beta numbers as final. These are v0.35.0-beta1 values and can move before the final release; the fix is to treat the structure of the behavior (band → soften → follow-up, telegraph, nuke-back) as stable and the exact constants as the thing to re-verify against each new release note.

Mode and map adjustments, and the completion test

The strategy sharpens or softens by mode. In FFA, the crown rule is live: the moment you cross half the non-fallout land, every Nation nukes the crown, so plan a second position before you go for the win. In Team games, Nations build more ports (ratio 1.0) and back up a teammate’s fight first before hitting the enemy with the most structures — so a Nation in your team is a port-and-trade partner to protect, and the enemy Nations prioritize the teammate who is actively fighting, which changes who you should shield with nukes and SAMs. In Humans-vs-Nations, the Impossible Nation is a single coordinated opponent with up to three silos and a dense, upgraded SAM umbrella, so the MIRV-denial playbook from Scenario 2 is the primary endgame, and your whole game is about reaching the midgame with a compact, well-buffered core rather than a spread, exposed one. In No-alliances lobbies the Nations are still active and the relations change still applies, so the telegraph and the nuke-back still hold even when diplomacy is disabled.

On map shape, the blast math is the dominant variable. On maps where your border to the Nation runs along a coastline or open water, you can make the Nation’s Hydro fail the 0.4 coverage test by keeping your exposed footprint small and behind the waterline — a big Hydro on shallow, mostly-water land is a poor trade it will often refuse. On maps where your border is a long, solid land contact with no buffer, you are presenting exactly the full-coverage blast the Nation scores, so the first job is to create depth (withdraw behind a second line, build the buffer) before you think about offense. On compact maps where borders are short, the 10-tile margin is a large fraction of the border and the follow-up is easier for the Nation to field, so the stop-line discipline matters more; on open maps with long coastlines you have more room to make blasts waste themselves.

The completion test for this guide is concrete and you can run it on your own map: can you name, right now, the troop ratio at which you stop the Nation border, and is that ratio safely below half the Nation’s current troops? Do you know which of your structures are inside the SAM umbrella and which are exposed on the contact line? Can you read the Nation’s Defense-Post line as a direction of intent, and can you say whether its next nuke would pass its own MIN_BLAST_COVERAGE test on your layout? If you can answer all four, you are not the cheap target the v35 Nation is built to punish, and you have the follow-up-denial plan ready for the moment it commits. For the broader behavior tree, structures, alliances, and nuclear branches, see the Nation AI mechanics reference; for the v34 pressure model this build changes, read Nation Pressure; for choosing what to nuke first, use Threat Assessment and the MIRV guide, and for the numbers behind your own nuke budget, the Nuke Calculator.