OpenFront Stack vs Spread: Where to Place Cities, Factories, and Ports

Direct answer: there is no buff, so compute the four levers instead of choosing a side

The community’s most-upvoted answer to “stack or spread?” is that there is no hidden buff or debuff, and the v0.34.12 source confirms it: the game applies no layout modifier to your territory, so the cost of clustering versus scattering is purely geometric and you can compute it yourself. The single most useful reframing is that the question is not a binary. In v0.34.12 “stacking” is a tight cluster of buildings that sit at the 15-tile minimum distance from one another, and “spreading” is building those same structures across your territory with gaps of hundreds of tiles. The layout you choose is re-priced by four separate mechanics at all times: the blast radius of the nukes a neighbor can afford, the SAM sites needed to cover your footprint, the troop capacity your city levels produce, and the gold your trade network earns from the cities a train can visit. The correct move is to run a 60-second decision loop before you place each new building, not to pick a side once and hold it.

First, ask whether any neighbor is nuke-capable and what blast that means. An Atom bomb costs 750,000 gold and destroys up to a 30-tile radius; a Hydrogen bomb costs 5,000,000 and reaches 100 tiles; a MIRV costs 25,000,000 and drops many warheads at once. If nobody on the board can field a Hydro yet, the nuke lever is near zero and the other three levers decide. Second, ask how many SAM sites your current footprint needs and how many it will need after the next expansion. Third, check that your city levels, not your city positions, are meeting the troop cap you need. Fourth, check whether your train route is above or below the nine-city free threshold. The table below turns those four questions into immediate board decisions, and the rest of the page is the exact v0.34.12 mechanics behind each lever.

Board signal	What the v0.34.12 mechanics mean	Immediate decision
No neighbor has 5M gold	Hydro is not yet a threat; nuke lever is near zero	Spread for cap and trade; cluster only behind your SAM line
A neighbor holds 5M+ gold	They can field a Hydro with a 100-tile radius	Move new cities 100+ tiles from your existing ones
Your cities sit at the 15-tile floor	A single Hydro overlaps several of them	Price the overlap; add a SAM or move the next city
Your city levels fall behind your cap target	Position does not fix this; levels do	Keep building city levels, not city count
Your train visits 12+ cities	Each city past nine costs 5k per run	Extend the route only if the new city pays for it

If you internalize only one sentence, make it this: the layout that wins is the one whose four-lever cost is lowest this phase on this map, and that answer changes every time a neighbor crosses a nuke-cost threshold. The threat assessment guide tells you when a neighbor crosses that threshold; this page owns what to do with your buildings once you know.

The 15-tile floor: what “stacking” actually means in v0.34.12

The most important thing to understand before comparing layouts is that the game does not let you build two structures on the same tile, so the colloquial “stack a city on top of your city” is not what happens. The configuration function structureMinDist() returns 15, and the placement check enforces that any two of your structures must be at least 15 tiles apart (measured as a squared distance, so the effective minimum gap is 15 tiles). That means the tightest possible “stack” is a cluster where every building sits exactly 15 tiles from its neighbors, and the loosest possible “spread” is buildings separated by hundreds of tiles. Everything downstream of that floor is measured in tiles from it: the nuke radii, the SAM range, the train station range, and the blast overlap that decides whether one bomb takes one city or six.

This floor has two consequences that people miss. First, it sets a minimum, so you cannot get closer than 15 tiles no matter how much you want to concentrate; the question is never “can I overlap?” but “how far apart, from 15 tiles outward?” Second, it means the two extreme layouts are not “same tile” versus “across the map.” They are a 15-tile-gap cluster versus a 500-tile-gap scatter, and the difference in nuke blast overlap between those two layouts is the entire nuke-risk calculation. A cluster of eight cities at the 15-tile floor fits inside a circle of roughly 60 tiles; the same eight cities spread across a 400-tile nation sit, on average, more than 3.6 tiles apart from the territory edge but hundreds of tiles from one another. When a Hydro with a 100-tile radius lands on the cluster, it can plausibly touch all eight cities; when it lands on the spread, it touches maybe one. That is the whole of the “stacking is bad” intuition, and it is a geometric fact, not a balance opinion.

The floor also interacts with the structures you build around your cities. A SAM launcher, a defense post, and a factory all obey the same 15-tile rule, so a “stacked” city that is also a “stacked” SAM and a “stacked” factory is a three-structure cluster at 15-tile gaps, and that cluster is exactly the kind of target a Hydro is sized to erase. The guide’s scenarios below work out the blast overlap for both layouts, but the floor is the constant that makes both layouts well-defined in the first place. If you only remember one number from this page, make it 15: it is the distance that separates “stacked” from “spread” and it is the unit every other number on this page is measured in.

Nuke radii: the cost of clustering, computed from the tag

The nuke lever is the one that makes the community’s “stacking is bad” prior correct, and the v0.34.12 nukeMagnitudes function gives the exact radii in tiles. A MIRV warhead has an inner radius of 12 and an outer radius of 18; an Atom bomb has an inner radius of 12 and an outer radius of 30; a Hydrogen bomb has an inner radius of 80 and an outer radius of 100. The inner radius is the zone that is destroyed with certainty, and the outer radius is the ring around it where destruction is partial: in the land nuke execution, every tile inside the inner radius is removed, while tiles in the ring between the inner and outer radii are destroyed with a coin-flip probability, which is how a single Hydro can erase a dozen or more cities that sit inside its 100-tile outer ring. The cost ladder that decides which nuke a neighbor can field is also in the configuration: an Atom is 750,000 gold, a Hydro is 5,000,000, and a MIRV is 25,000,000 plus 15,000,000 for every MIRV already launched in the match, so MIRVs get dramatically more expensive the longer the game runs.

The geometry of the radii is where the layout decision lives. A Hydro’s outer circle of 100 tiles covers an area of roughly 31,400 square tiles; the inner 80-tile circle covers roughly 20,100. If your eight stacked cities sit inside a 60-tile cluster, all eight are inside the inner 80-tile circle of a Hydro that lands anywhere near the cluster, so a single well-aimed Hydro takes all eight with certainty. If the same eight cities are spread across a 400-tile nation, the 100-tile Hydro touches at most a handful of them, and the partial ring means even those are not all lost. An Atom is the early-game version of the same trade: its 30-tile outer radius is large enough to take two or three cities at the 15-tile floor but not an entire spread, so an Atom is the threshold weapon that separates “cluster is survivable” from “cluster is not.” That is why the community splits its advice by phase: before anyone can field a Hydro, the Atom lever is the binding one, and after a player crosses 5M gold, the Hydro lever takes over and the cluster stops being survivable.

The MIRV is the late-game complication. A single MIRV costs 25M and drops many warheads, each with a 12/18 radius, so a MIRV aimed at your cluster lands multiple 18-tile blasts across the cluster area, and a MIRV aimed at your spread lands multiple 18-tile blasts across the nation. The MIRV is the one weapon that makes even a spread layout dangerous, because its warheads are small enough to pick out individual cities but numerous enough to cover a wide area. That is the entire reason the community’s “spread good early, stack good late” advice is not a contradiction: early, the Hydro is the binding threat and it favors spreading; late, the MIRV is the binding threat and it favors concentrating behind a bunker or a SAM umbrella that the MIRV warheads cannot all get through. The nuke calculator guide has the full blast-overlap math; this page uses the radii only to price the layout decision.

SAM coverage: the cost of spreading, computed from the tag

The SAM lever is the one that makes “spreading is always better” too strong, and the v0.34.12 configuration gives the exact cost. A SAM launcher costs 1,500,000 × (numUnits + 1) gold, capped at 3,000,000, so the first SAM is 1.5M, the second is 3M, and every SAM after that is 3M. That cost curve is steep enough that the number of SAMs you need is a first-order expense, and the number of SAMs you need is set by your footprint: the wider your cities are spread, the more SAM sites it takes to keep them all inside an interception range. The range function is samRange(level) = 150 − 480 / (level + 5), so a level 1 SAM reaches 70 tiles, a level 3 SAM reaches exactly 100 (the Hydro outer radius), a level 5 SAM reaches 110, a level 10 SAM reaches 140, and the range asymptotes to 150. Each SAM upgrade takes 45 seconds, so a SAM you upgrade mid-game is temporarily at the lower range.

Now price the two layouts. A tight cluster of eight cities at the 15-tile floor fits inside a circle of roughly 60 tiles, so a single level 3 SAM (range 100) placed at the cluster center covers the entire cluster; the SAM cost is 1.5M. The same eight cities spread across a 400-tile nation need to be covered at level 3 range of 100 tiles, and a 400-tile nation cannot be covered by one 100-tile SAM; it takes roughly three SAMs at the 1.5M, 3M, and 3M cost curve, a total of 7.5M. The spread therefore carries roughly 6M more in SAM cost than the cluster at the same city count, and that 6M is the “SAM coverage premium” that the nuke lever is trying to pay off. The break-even question is whether the nukes you avoid by spreading are worth more than the 6M you pay in SAMs to cover the spread. Early, when no Hydro is in play, the avoided nukes are near zero and the spread is strictly more expensive; late, when a Hydro can take six of your cities, the avoided nukes are worth far more than 6M and the spread is strictly cheaper.

There is a second-order effect the cost curve creates: the SAM cap at 3M means that past the second SAM, every additional SAM costs the same, so the marginal cost of covering a larger spread is flat at 3M per SAM, and the marginal benefit of spreading past the point where your SAMs already cover your cities is zero. In other words, there is a natural “SAM saturation” footprint for each number of SAMs you can afford, and the optimal layout is to spread your cities to just inside that footprint, not to the edge of the map. The sam launchers guide covers the placement and upgrade timing; this page uses the SAM cost only to show that spreading is not free, and that the free part ends exactly where your SAM coverage ends.

Troop capacity: the lever position does not touch

The counterintuitive lever is the one that does not depend on layout at all, and it is the one players most often get wrong. The v0.34.12 maxTroops function is 2 × (tiles^0.6 × 1000 + 50000) + (total city level) × cityTroopIncrease(), where cityTroopIncrease() is 250,000. The first term scales with the number of tiles you own, and the second term scales with the sum of your city levels, not with the number of cities or with where those cities are. That means a nation with thirty cities at level 7 has the same troop cap from the city term as a nation with six cities at level 35, regardless of whether those cities are stacked at the 15-tile floor or spread across the map. The city term is 21 × 250,000 = 5.25M in the stacked case and the same 5.25M in the spread case; only the tile term differs, and the tile term is set by how much land you own, not by where your cities are.

This is why “stacking” and “spreading” are not troop-cap strategies at all. The community sometimes treats building more cities as the way to raise your cap, and that is true, but it is true whether you stack them or spread them. The position of the city contributes nothing to the cap; the level of the city contributes everything. The practical rule is to separate the two decisions completely: the count and level of your cities is a troop-cap decision, and the position of your cities is a nuke-risk and SAM-cost decision. A player who stacks cities because they think it gives them more troops is confusing the two levers, and the fix is to build city levels for the cap and choose city positions for the blast and the SAM. The population growth guide covers the growth side of the cap; this page only needs the fact that the cap is level-driven so the layout choice is free of it.

The tile term is the one layout-adjacent piece, and it is small. A nation of 400 tiles has a tile term of 2 × (400^0.6 × 1000 + 50000) ≈ 2 × (3020 × 1000 + 50000) ≈ 6.54M, and a nation of 4000 tiles has a tile term of 2 × (4000^0.6 × 1000 + 50000) ≈ 2 × (6310 × 1000 + 50000) ≈ 13.22M, so the tile term roughly doubles between 400 and 4000 tiles. That is a real difference, but it is set by how much land you own, and a player who spreads their cities does not automatically own more land. The two are correlated, because a spread layout usually implies a bigger nation, but the cap is driven by the land and the city levels, not by the gap between the cities. Treat the tile term as a byproduct of your expansion, not as a layout lever, and the layout decision reduces cleanly to the nuke and the SAM.

Train-network gold: the lever that rewards spreading

The trade lever is the one that quietly makes spreading cheaper, and it is the one the community understates. The v0.34.12 trainGold function pays a base of 10,000 gold for a self route, 25,000 for a team or other route, and 35,000 for an ally route, and then subtracts 5,000 for every city the train visits beyond the first nine, floored at 5,000. The comment in the source says the first ten cities are free (the code subtracts nine, so the tenth is the first penalized city), which means a train that visits nine or fewer cities pays the full base, and a train that visits ten or more cities pays the base minus 5k per extra city. The trainStationMinRange is 15 tiles and the trainStationMaxRange is 110 tiles, so a train station can reach cities within 110 tiles, and a train that runs a long route visiting many cities is the mechanism by which a spread layout turns its spread into income.

Now price the two layouts on the trade lever. A cluster of eight cities at the 15-tile floor sits inside a single train station’s 110-tile reach, so one train visiting the cluster visits all eight cities, which is below the nine-city free threshold, and pays the full 10,000 self base. A spread of thirty cities across a 3000-tile nation requires multiple train stations, and a train that runs the full spread visits well over nine cities, so it pays 10,000 minus 5,000 per extra city, floored at 5,000. The spread therefore pays less per run than the cluster, which seems to contradict the “spread pays” intuition, but the resolution is that the spread runs more trains and more routes. The cluster’s single train is the only route, and its 10k per run is the total trade income; the spread’s many trains and many routes each pay less but multiply, and the total is higher. The trade lever does not reward a single spread route; it rewards a network of spread routes, and the network is what the cluster cannot have.

This is the lever that justifies the “spread for the trade network” advice, and it is the one that interacts with the SAM lever in the opposite direction. The SAM lever says a larger footprint costs more in SAMs, and the trade lever says a larger footprint earns more in train gold, so the optimal footprint is where the marginal train gold of one more spread city equals the marginal SAM cost of covering it. The train network guide covers the route planning and the station placement; this page uses the train gold only to show that spreading is not purely a cost, that it is also an income, and that the break-even is a per-city marginal comparison, not a global “spread is better” statement. The two levers are the same decision seen from two sides, and the player who prices both of them per city has the answer.

Scenario 1: the Hydro on a stacked cluster versus the same nation spread

Work the nuke lever end to end on a concrete nation so the radii stop being abstract. A mid-game player owns 3,000 tiles and has built thirty cities, most at level 7, for a total city level of 210. The tile term of the cap is 2 × (3000^0.6 × 1000 + 50000) ≈ 2 × (2884 × 1000 + 50000) ≈ 6.17M, and the city term is 210 × 250,000 = 52.5M, so the max troop cap is roughly 58.7M. Now a neighbor with 5M of gold fields a Hydro and aims it at the player’s home region. In the stacked layout, the thirty cities sit at the 15-tile floor inside a cluster of roughly 150 tiles across, and the Hydro’s 100-tile outer radius covers the entire cluster. Every one of the thirty cities is inside the outer ring, and the eight or ten that are inside the 80-tile inner circle are destroyed with certainty while the rest are destroyed with a coin flip, so a realistic outcome is twenty to twenty-five of the thirty cities gone in one strike. The player loses their home base, their SAMs that were covering the cluster, and a large fraction of their troop cap, and the match is effectively decided.

In the spread layout, the same thirty cities are scattered across the 3,000 tiles, and the Hydro’s 100-tile outer ring covers only a small fraction of the nation. A realistic outcome is four to six of the thirty cities inside the outer ring, with two to four inside the inner circle destroyed with certainty. The player loses a fifth to a sixth of their cities, keeps the remaining twenty-four to twenty-six, and their troop cap drops from 58.7M to roughly 45M, which is a real loss but not a match-ender. The same 5M gold, the same Hydro, the same nation: the only difference is the layout, and the difference is the outcome. This is the entire community “stacking is bad” argument made concrete, and it is the scenario every new player should run in their head before they place a city at the 15-tile floor.

The follow-up question is what the player should have done differently, and the answer is the SAM lever. In the stacked layout, the cluster of 150 tiles could have been covered by two level 3 SAMs (range 100) for a total of 4.5M, and a level 3 SAM intercepts an incoming nuke within its range, so the 4.5M SAM investment would have saved the 5M Hydro from landing at all. In the spread layout, the SAMs cannot cover the whole nation, so the spread is the layout that requires accepting some nuke loss, and the cluster is the layout that requires the SAM investment. The two layouts are not better or worse in the abstract; the cluster is cheaper to defend if you can afford the SAMs and the spread is cheaper to defend if you cannot, and the player who cannot afford 4.5M of SAMs should spread because the spread’s nuke loss is bounded while the cluster’s nuke loss is total.

Scenario 2: the 2v2 trade line and the per-city marginal cost

Now work the trade lever on a team map so the train gold stops being abstract. In a 2v2, each team has a home region and a trade line that connects its cities to its ports, and the team’s income is the sum of the train gold across every route it runs. A team with twelve cities on its home region runs a self route that visits all twelve, which is three cities past the nine-city free threshold, so the route pays 10,000 minus 3 × 5,000 = 5,000 gold per run. If the team spreads three of those cities an extra 50 tiles apart, the train can visit four more cities on a second route, and the second route pays 10,000 minus (13 − 9) × 5,000 = 5,000, so the team’s total trade income goes from one route at 5,000 to two routes at 5,000 each, a total of 10,000, which is double. The marginal city that the extra spread enabled is worth the second route, and that is the “spread pays” mechanism on a team map.

The same layout has a SAM cost, and that is the break-even. The extra three cities spread 50 tiles apart require one more SAM to cover them, at the 3M cost (the team already has its first two SAMs), and the extra trade income from the second route is 5,000 per run. Over a match that runs the route a hundred times, the extra trade income is 500,000, which is far less than the 3M SAM, so the spread is not worth it on the trade lever alone. But the spread also spreads the team’s cities across a wider area, which reduces the nuke lever: an opponent’s Hydro now touches fewer of the team’s cities, and the avoided nuke loss is worth more than the 3M SAM if the opponent is nuke-capable. The 2v2 scenario is therefore the same four-lever trade as the FFA scenario, just with the ally route base of 35,000 and the team’s shared SAM pool, and the conclusion is the same: the layout is worth it when the avoided nukes plus the extra trade routes exceed the SAM cost, and it is not worth it when they do not.

The practical rule for a 2v2 is to let the opponent’s nuke capability set the spread. If the opposing team is below 5M gold, they cannot field a Hydro, the nuke lever is near zero, and the trade lever alone does not justify the SAM cost, so keep the cities clustered and the trade line short. If the opposing team crosses 5M, the nuke lever turns on and the spread’s nuke avoidance becomes worth more than the SAM cost, so spread the cities and add the SAM. The four islands team coordination guide covers the 2v2 coordination side; this page uses the 2v2 only to show that the four-lever trade is the same in team play, and that the ally route base of 35,000 makes the trade lever slightly stronger in team play than in FFA.

Failure modes and counterplay: when each layout breaks

The most common failure is the early stack. A new player who reads “stacking is bad” and then stacks their first five cities at the 15-tile floor in the opening minutes is doing the opposite of the advice, and the reason is that the opening is the one phase where the nuke lever is zero and the cluster is actually optimal. Before any neighbor can field an Atom at 750k, there is no nuke threat, the SAM lever is the only cost, and a tight cluster covered by one 1.5M SAM is the cheapest defense on the board. The failure is not the stack; the failure is the timing, because the player should stack early and then spread as the neighbors cross the nuke-cost thresholds. The counterplay is to treat the layout as a phase variable: cluster in the opening, begin spreading once a neighbor reaches 750k, and spread wide once a neighbor reaches 5M.

The second failure is the SAM-only defense against a Hydro. A player who clusters and covers the cluster with SAMs is protected against an Atom, because the Atom’s 30-tile radius is inside the SAM’s 100-tile range, but the Hydro’s 100-tile radius is exactly at the level 3 SAM range, and a Hydro that lands at the edge of the SAM range is not intercepted. The counterplay is to upgrade the SAMs past level 3, because a level 5 SAM reaches 110 and a level 10 SAM reaches 140, both well past the Hydro’s 100, so the SAM umbrella only works against a Hydro if it is upgraded. The third failure is the MIRV in the late game, and the counterplay is the MIRV bunker: a bunker is the one structure that survives a nuke, and a city stacked inside a bunker is the one stacked city that a Hydro or a MIRV cannot take, which is the entire reason the community’s “stack good late” advice exists.

The fourth failure is the spread that is not a network. A player who spreads their cities but does not build the train stations and the rail lines to connect them has the nuke benefit of the spread without the trade benefit, so they pay the SAM cost of the spread and earn none of the trade gold. The counterplay is to treat the train network as a first-class build, not an afterthought: the spread is only worth it if the rail line is built, and the rail line is only worth it if the spread is built. The fifth failure is the cluster that is not behind a SAM, which is the early stack with no defense, and the counterplay is the 1.5M SAM at the cluster center. The water nukes guide covers the nuke variants that change the radii, and the island defense guide covers the bunker and the coastal version of the cluster; this page owns the land-based layout decision and the four levers that price it.

Decision table and mode adjustments: the 60-second loop you run before each build

The table below is the whole of the guide, compressed. Run it before you place each new building, and the answer is the layout for that moment. The inputs are the phase (opening, mid, late), the strongest neighbor’s gold, and the nuke they can field; the output is the layout and the SAM count. Opening, no neighbor at 750k: cluster at the 15-tile floor, one SAM at 1.5M, build city levels for the cap. Mid, a neighbor at 750k to 5M: begin spreading, each new city at least 30 tiles from the last, two SAMs at 4.5M. Late, a neighbor at 5M or more: spread wide, each new city at least 100 tiles from the last, three or more SAMs, and stack any city you can afford to put in a bunker. The troop-cap column is always the same, because the cap is level-driven: keep building city levels regardless of the layout.

Phase	Strongest neighbor gold	Nuke they can field	Layout	SAM count	Troop-cap action
Opening	below 750k	none	cluster at 15-tile floor	one at 1.5M	build city levels
Mid	750k to 5M	Atom (30-tile)	begin spreading, 30-tile gaps	two at 4.5M	build city levels
Late	5M or more	Hydro (100-tile) / MIRV	spread wide, 100-tile gaps	three or more	build city levels, bunker the core

The mode adjustments change the table but not the levers. In a 2v2, the ally route base of 35,000 makes the trade lever stronger and the shared SAM pool makes the SAM lever cheaper, so the break-even spread comes earlier in team play. On a compact map, the territory is small enough that the spread is capped by the map edge, so the cluster is the default and the spread is only for the cities you cannot cluster. On a large world map, the spread is the default and the cluster is only for the bunker core. In No Alliances, the nuke lever is stronger because no neighbor is treaty-protected, so the spread is the default from the opening. The no alliances mode guide covers that mode’s border logic; this page only notes that the four levers are the same and the nuke lever is weighted higher.

The final rule is the one that ties the page together: the layout is a phase variable, not a fixed preference, and the phase is set by the strongest neighbor’s gold. The player who chooses a layout once and holds it for the whole match is the player who clusters in the late game and loses to a Hydro, and the player who re-runs the 60-second loop before every build is the player who clusters early, spreads as the nukes turn on, and bunkers the core in the endgame. That is the entire strategy, and it is the reason the community’s “no buff” answer is correct: there is no hidden modifier, so the only thing that can price the layout is the four levers, and the four levers are all in the v0.34.12 configuration, and they are all on this page.

How this differs from the other layout guides

This guide is the one the other layout guides do not own. The tall versus wide guide is about territory posture, the shape of the nation you own, and it does not ask where the cities are. The map strategy guide is about the opening and the expansion direction, and it does not ask the nuke or the SAM question. The port versus factory guide is about the production split between ports and factories, and it does not ask the layout question. The attack ratio guide is about the send slider, and the land combat guide is about the border fight, and neither of them asks where the buildings are. This page owns the single decision that all of those pages assume but none of them answer: given the territory you have and the production you are running, where do the cities, the factories, and the ports go, and how does the answer change when a neighbor crosses a nuke-cost threshold?

The reason the decision deserves its own page is that the four levers are in tension with each other, and no single page about territory or production can hold the tension. The nuke lever wants the cluster, the SAM lever wants the cluster to be covered, the trade lever wants the spread, and the troop-cap lever does not care, and the player who only thinks about one lever makes the layout decision on one lever and loses on the other three. This page holds all four levers at once, prices them against each other in the two scenarios, and gives the 60-second loop that resolves the tension in real time. If you already know the territory posture and the production split, this page is the missing decision, and if you do not, read the linked guides first and come back to this page for the layout, because the layout is the last variable you set and the one that the nuke and the SAM are waiting on.

The final note is the version note. Every number on this page is from the v0.34.12 tag, commit 7c27263, and the constants are in the game configuration and the nuke execution. If a future release changes the nuke radii, the SAM cost, the train gold, or the troop-cap formula, the four levers change and the table on this page changes with them, so the freshness summary at the top of the page is the place to check before you trust the numbers in a new release. The layout decision is stable across versions because it is a geometric trade, but the exact break-even points move with the constants, and the constants are the only part of this page that the game can change.