OpenFront Attack Ratio: What Percentage to Send, and When to Click

Direct answer: send a stack that beats the defender’s army, then size and time it

The attack ratio is the bottom-left “send” slider: each click commits that percentage of your selected troops to the border, and the loss and speed of the push depend on the ratio between your committed stack and the defender’s army, not on the percentage itself. The single decision that wins most fights is to choose the percentage so that the troops actually leaving your line beat the defender’s army on the border they will face, and then click at the moment your next push can still hold that ratio. In practice that means three habits instead of a fixed number.

First, for routine land against a weaker or distracted neighbor, send 25 to 40 percent and double-click at the point your growth can fund the second click. That range is where community consensus clusters, and it is the range that leaves you with troops behind the line for a follow-up while still landing a push at least as large as the defender can field at that border. Second, when a full-send is coming at you, do not answer it with your own 100 percent while it is still in the air: whittle it with 5 to 15 percent clicks as it crosses, and take the land only after it has committed and its army is spread over new tiles. Third, hold 100 percent in reserve for the two moments it is actually right: the opening full-send at match start when nobody has a real army yet, and the endgame kill push against a defender whose density you have already read. Every other full-send is a coin flip that the math below makes worse.

The rest of this page is the mechanism behind those three habits. It explains exactly how the game converts your slider setting into per-tile losses and push speed, why the numbers make one bigger push strictly cheaper than two smaller ones, why the endgame “900,000 into 1,000,000 in ten seconds” shock is the density system working as designed, and a concrete decision table you can run in your head before every click. If you only internalize one sentence from this page, make it this one: the ratio you want is not “a percent of my army,” it is “the fraction of the defender’s army my push will actually face when it lands.” Everything else on this page is how you compute that fraction before the click, not after.

The two ratios the game actually reads

The v0.34.11 combat formula, in Config.attackLogic, is driven by two numbers, and the second one is the one most players never think about. The first is the troop ratio: the defender’s total army divided by the troops in your push, defender.troops / attackTroops. The formula clamps this ratio between 0.6 and 2, which does two things at once. It means an attack that is smaller than 0.6 of the defender’s army (you are outnumbered by more than 1.67 to 1) pays exactly the same per-tile loss as an attack that is 0.6 of their army: beyond a certain point of being outgunned, the game stops punishing you more, because a hopeless push is already hopeless. It also means an attack bigger than twice the defender’s army (ratio under 0.5) pays no less than an attack that is exactly half their army: once you are overwhelming, the extra troops buy you speed, not a lower loss. The clamp is the entire reason a fixed percentage is not a fixed risk.

The second ratio is defender density: the defender’s total troops divided by their total tiles, defender.troops / defender.numTiles, their average troops per tile. Your per-tile loss is the terrain magnitude multiplied by the clamped ratio, multiplied by a base of 0.463 plus 0.0039 for every defender troop per tile of defending density. The defender’s loss in the same fight is exactly that average troops-per-tile number. So the same 10,000-troop defender is a wall when it is packed on 400 tiles (25 troops per tile, adding roughly 0.098 of loss per tile to your 0.463 base on plains) and a sieve when it is spread over 4,000 tiles (2.5 troops per tile, adding only 0.0098). The density term is why attacking a nation that just won a long war across the whole map feels different from attacking the same nation right after it spread thin to take islands.

Speed is the third output, and it is where your slider setting shows up directly. The per-tile speed cost is within(troopRatio, 0.82, 7.5) x within(troopRatio/20, 1, 50) / 8.55. The first factor floors at 0.82: an attack whose ratio is below 0.82 (you field more than 1.22 troops per defender troop) moves at full speed, and that is the 2x rule of thumb you see in every video: double the defender’s army and you are safely inside the fast zone with margin. Above 0.82 the cost climbs linearly toward 7.5, so an attack at ratio 2 (you are outnumbered 2 to 1) crawls at roughly half the speed of the same attack run at parity. Past ratio 20 the second factor takes over and the push slows to a fraction of a tile per tick. A defense post on the defending tile multiplies the terrain loss by 5 and the tile cost by 3, and a bot defender’s losses are multiplied by 0.7. All of this is why “what percent should I send” is really two questions at once: what army am I bringing to the ratio, and what density am I walking through?

Why one bigger push beats several small ones

The most counterintuitive result in the formula is that the per-troop cost of a push falls as the push gets bigger, up to the point where it is overwhelming. Take the community’s favorite example from the threads: you have 50,000 troops and the defender fields 25,000. If you push twice at 30 percent, your first stack is 15,000 troops against 25,000 of theirs, a ratio of 1.67, clamped down to 0.6. Your second stack is another 15,000, and by then the defender has regrown some of what it lost, so the second ratio is even worse. If you push once at 50 percent, your stack is 25,000, ratio exactly 1.0, and every troop in that push pays the full clamped 1.0 cost once. Both pushes take the same land, but the two-click version pays the 0.6 clamp on a stack that is smaller than the defender’s army and then pays again against a partially regenerated army, while the single push pays a flat 1.0 against the army as it actually was. The single bigger push is cheaper per tile and faster, because the speed factor at ratio 1.0 is lower than at 1.67.

The general rule is a ratio ladder, and it is worth committing to memory because it is the whole of the “when to click” decision. When the defender’s army is small relative to yours, you can afford to push at 100 percent of what you can muster and land at ratio 2 or better, which is maximum speed with no extra loss. As the defender’s army approaches yours, the right move is to wait for your next growth cycle and push at parity, because a parity push is the biggest push you can make that is not yet outnumbered. When the defender is bigger than you, any push is outnumbered and the ratio clamp means your loss does not improve no matter how small you make the stack, so the only rational small pushes are probes, and the rational big push is a 100 percent kill attempt where you accept the loss and bet on the land being worth more than the army. The middle of that ladder, the 25 to 40 percent range, is where parity is most of the time in a three-player match with a bot neighbor, and it is why that range is the consensus default.

There is a second-order effect the ladder explains: double-clicking is a ratio tool, not a speed tool. The first click of a double at 30 percent lands at a good ratio and takes the first strip of tiles; the defender then loses its average troops per tile across the whole front, so its density drops for the second click, and your second 30 percent push walks through cheaper density at a slightly better ratio. That is the real value of the 25 to 40 percent plus second-click pattern: the second push is not repeating the first, it is exploiting the density drop the first one created. If you are ever going to click twice, the second click should land within the growth window that funds it, because the density drop decays as the defender regrows, and a second click two cycles late is just a small, outnumbered push paying the 0.6 clamp against a nearly whole army again.

Density and size: why the endgame feels different

The most common question in the threads is not about the slider at all. It is “why could I take a 5,000-troop bot early with 10,000 of mine, but 900,000 of me went into 1,000,000 of them late game and lost 20 percent of the map in ten seconds?” The answer is the density term, plus one size system on top of it, and both are deliberate anti-snowball mechanics. Early, the 5,000-troop bot is spread across maybe 800 tiles, so its density is 6.25 troops per tile and the density term adds only 0.024 to your per-tile loss. Your 10,000 push is a 2 to 1 ratio, inside the fast zone, and the land is cheap to take. Late, the 1,000,000-troop defender that just won a war is packed onto maybe 300 tiles of hard border, so its density is 3,333 troops per tile, and the density term adds 13 to your per-tile loss on top of the 0.463 base. The same percentage slider that was a free land-grab at 6 troops per tile is now a massacre at 3,333 troops per tile, and the game is working exactly as designed.

The size system on top of that is the large-territory bonus, a smooth sigmoid centered on 300,000 tiles. A defender with more tiles than you gets a rising loss floor as it grows: the more land a nation owns, the more its per-tile losses are protected, because a giant empire is the thing every other player is hunting and the game subsidizes its defense to keep the match winnable. The attacker’s bonus is the mirror of it and the stronger one: attacking from a large territory is cheaper and faster, capping the gap at roughly 3.3x a small player’s rate. This is why the endgame of a match where one player has 60 percent of the map is a war of attrition that the two smaller players fight together: the giant’s density on every front is high, its size floor is at the ceiling, and the only way to bring its per-tile loss down is to pull its army off one front so the other front’s density drops.

The practical reading of all of this is that density, not army size, is what you are actually fighting, and you can read it in two seconds. Before any push, estimate the defender’s troops on the border you are facing divided by the tiles of front they hold there. If that number is under about 10, your density penalty is small and percentage barely matters; push at whatever ratio you can fund. If it is over 100, your density penalty is over 0.4, which is nearly double the base loss, and no percentage you send is going to make the land cheap; you either break their concentration first or you do not attack that front at all. The middle band, 10 to 100 troops per tile, is where the slider range on this page earns its keep, and where the 25 to 40 percent plus second-click pattern exists to exploit the density drop between clicks. Read the density, pick the push that lands at the ratio you can hold, and the endgame stop feeling like a wall and start feeling like arithmetic.

Choosing the ratio: a three-question framework

Before every click, answer three questions in a fixed order, because the order matters: the third question changes the answer to the second, and answering them out of sequence is how players over-commit. Question one is what is the defender doing right now, not what is their army size. If they are sending a full-send at someone else, their army is on the move and their home front is thin; that is a 40 to 100 percent push into the exposed front, the opening-weapon full-send the community describes against a distracted third player. If they are sitting and growing, their army is packed on its border and the right push is the parity push your growth funds, 25 to 40 percent with the second click on the next cycle. If they are spreading thin to take land themselves, their density on the front you are facing is dropping and your ratio is improving for free; hold your current ratio and let their own expansion do the work.

Question two is what density am I walking through, using the two-second read from the last section. Take their troops on your front and divide by the front’s tiles. Under 10, push at maximum fundable ratio. From 10 to 100, the 25 to 40 percent band is the default and the double-click is worth it because the first click’s density drop makes the second click cheap. Over 100, do not attack that front; either break the concentration on an adjacent front first or wait for them to commit troops elsewhere. This is the single most valuable sentence on the page for endgame play: a front that reads over 100 troops per tile is not a wall you push through, it is a wall you route around, and the routing move is almost always a cheap probe on a neighboring front that forces the defender to move troops, which drops the density on both fronts at once.

Question three is what would my next push be, and it is the one that turns a good push into a bad one or the other. If your next push after this one would be outnumbered (ratio over 1.5), then this push is the last one you can afford to send at a good ratio, and you should either send your biggest fundable push now and accept a one-shot fight, or hold and grow so the next push is the big one. If your next push would still be at parity or better, then the correct move is the 25 to 40 percent push now and the second click on the next cycle, because the density drop between clicks is worth more than the extra troops a bigger single push would bring. This third question is why the slider setting you see in replays is almost never a fixed number: the same player runs 100 percent at the start, 30 percent for the mid-game, and 15 percent when a full-send is incoming, and the framework above is the reason those three settings are the right settings, not three habits. Run the three questions on every push and the slider stops being a mystery and starts being a number you compute.

Numbers on the table: three concrete situations

The table below runs the v0.34.11 formula on three situations players actually meet, on plains with magnitude 80, no defense posts, no fallout. “Density” is the defender’s average troops per tile on the front, “ratio” is defender army over push stack, and “loss per tile” is the per-tile attacker loss the formula returns; defender loss is the density itself, and speed is relative to a full-speed parity push. The purpose is to give your eye three anchors: what a free push looks like, what a parity push looks like, and what an outnumbered push looks like.

Situation	Defender army	Defender tiles	Density	Your push	Ratio	Loss/tile	Read
Free land-grab	4,000	1,200	3.3	8,000 (50%)	0.5	~0.47	2:1, max speed, cheap
Parity mid-game	15,000	900	16.7	15,000 (30%)	1.0	~0.93	fundable parity, double-click
Outnumbered	30,000	1,500	20	10,000 (30%)	3.0 (clamped 2)	~1.05 + speed halved	probe only, or kill

The first row is the early-game land-grab and it tells you the percentage is almost irrelevant: at ratio 0.5 the clamp and the density term both put you at the cheap end, and the 50 percent stack is chosen only because it is what you can fund at that moment. The second row is the mid-game default this page keeps returning to: 15,000 troops against a 16.7-density front at exact parity, a loss per tile of about 0.93 on plains, and the second 30 percent click on the next cycle lands into the density the first click just dropped. The third row is the trap: 30,000 of them against your 10,000 push clamps the ratio to 2, so your per-tile loss is only marginally higher than the parity row, but your speed is roughly halved and the density is the same, meaning you buy the land at a real loss while the defender’s 20-troops-per-tile counter is only a small cost to them. The only versions of row three that are worth running are the probe, a small push whose purpose is to read their density and force a reaction, and the kill, a 100 percent push you only take when the land is worth more than the army you will lose.

Two adjustments to the table. If the front holds a defense post, multiply the loss by 5 and the tile cost by 3, which moves the parity row from 0.93 to roughly 4.6 per tile and makes the double-click pattern expensive enough that you should usually route around the posted front instead of paying for it. And if the defender is a bot, multiply its losses by 0.7, which is why attacking the bot neighbor is the cheap way to grow early: you are paying the bot’s density penalty in full, but the bot’s own counter-attacks cost it less, so the bot is the player you can afford to run the third-row probe against while you wait for the parity push on the human front. The table is a mental model, not a calculator; the game does not show you these numbers, but you can hold all three rows in your head and check every push against them in under two seconds, which is exactly the reaction time a push decision needs.

Mistakes the ratio makes: failures and the counter

The failure with the worst math in the whole system is countering a full-send at your own 100 percent while it is still crossing. The attacker chose their stack to sit at or near the 0.6 clamp against your army, which means their per-tile loss is at the cheap end and their speed is at the fast end; your 100 percent answer walks into the densest troops-per-tile the defender can field on the border they are holding, because a defender at full strength is exactly a defender whose army is packed. The two fights meet, and the attacker’s push is cheaper and faster than yours, so the counter is nullified and you have committed your entire army to a losing ratio. The community’s fix, and the one this page endorses, is the whittle: 5 to 15 percent clicks as the full-send crosses, timed to your growth turning yellow, which trades a small amount of your front’s land for a proportional drop in the incoming push’s effective army before it lands, and then the 15 percent follow-up after it commits, when its troops are spread over new tiles and its density on the old front is broken. The whittle is not a smaller version of the counter; it is a different fight with a different goal, and running it as a counter is how players who “should have won” lose the war.

The second failure is the fixed-ratio habit, and it shows up as 10 percent, 30 percent, or 100 percent players all reporting wins, because a fixed ratio happens to be right about one third of the time. The 10 percent player is safe against full-sends and starves against slow growth; the 30 percent player is the mid-game default and bleeds in the endgame; the 100 percent player wins the opening and dies to the second-largest army in the match. The fixed habit is a substitute for the three-question framework, and the substitute only costs you when the situation is one of the two-thirds the habit gets wrong, which is most of the match. The fix is not a better fixed number; it is the discipline of running the three questions, because the situation-dependent answer is the only answer that is right in more than one third of cases.

The third failure is the premature second click, and it is the one most players never notice because it looks like the correct move. The double-click pattern only works while the first click’s density drop is still in effect, which is roughly the next growth cycle; a second click two cycles late is a small push paying the 0.6 clamp against a defender that has regrown most of what it lost, and it is worse than holding. The check is simple: if the defender has already visibly regrown on the front before your second click is funded, the second click is a probe, not a push, and probes do not take land, they only read. Hold the line, let the density drop decay, and wait for a push you can send at parity; the land that the late second click was supposed to take is taken by the next real push, and the late second click has only cost you troops and speed for the privilege of reading a number you already had.

Map, mode, and endgame adjustments

The ratio framework holds on every map and in every mode, but the numbers shift, and the shifts are where map-specific knowledge lives. On coastal maps the front you are attacking is shorter than the front you are defending, because the water removes a strip of border from both sides, and a shorter front means the defender’s density on it is higher than their average density across the whole territory. The two-second read therefore underestimates the coastal front’s density, and the 25 to 40 percent band that is correct on an open land map is usually one band too high on a coastal one; the parity push that funds a double-click on open land often only funds a single click on the coast, and the second click should be held for the growth cycle after. If you are the defender on the coastal front, the same geometry works for you: your shorter front means your density is higher than the attacker’s model assumed, which is the mechanical reason holding the coast with a packed wall is so effective, and why the coastal defense guides pair with this one.

On mountain and highland terrain the magnitude term rises from 80 on plains to 100 on highland and 120 on mountain, a 25 to 50 percent increase in every loss number on this page, and the defense post on a mountain tile multiplies that magnitude by 5, which is why posted mountain fronts are the one front on the map you should almost never attack directly. The correct move against a posted mountain front is to break its adjacent open front first, force the defender to move troops off the posted tile, and only then push the mountain at the reduced density; pushing the posted front directly is paying the 5x on a front that is already at the high end of the density band, which is a loss number that no percentage you send can make fundable. The one exception is the endgame kill, where the land is worth the army and the 5x is a price you pay knowingly, and the check before that push is whether the defender’s army on the posted front has already been pulled off it by your pressure on the adjacent front.

In overtime, the two ratios invert: your own density is high because your territory is large and packed, and the defender’s density on the front is the number that decides everything, so the overtime push is always the parity push against the lowest-density front the defender holds, not the strongest front. The large-territory attacker bonus is your friend in overtime, because it is the bonus that is strongest when you are the biggest territory on the map, and it caps the defender’s size subsidy at a point where your attack is still faster than theirs. The overtime full-send, the 100 percent kill push, is the one situation on this page where the slider goes to maximum, and the check before it is the density read on the target front: if it is under 100, the kill push is fundable and the land is worth it; if it is over 100, the kill push is a coin flip against the size floor, and the correct move is to keep whittling the adjacent fronts until the target front’s density drops into the fundable band. Run the three questions on the overtime push exactly as you run them on any other push, and the endgame stop being a wall you push through and start being the arithmetic it always was.

Where to go next

The page you read next depends on which side of the ratio you are on and which question you now own, and the order in which you read them matters less than making sure each of the four questions this page defers actually gets answered by the page that owns it. If you are choosing what army to bring to the ratio rather than the percentage itself, the land combat guide is the per-tile loss reference that this page assumes you have already read, and it is where the terrain magnitudes and defense post numbers come from. That page is the one to open if the two-second density read on this page ever disagrees with what you see in the game, because the mismatch is almost always a terrain or defense-post term this page only quotes but does not derive. If your question is when a push is a threat at all, the threat assessment guide is the target-selection page: it decides which of your neighbors is the one your ratio is computed against, and this page is the percentage answer to the army answer that page gives. The two pages are a pair in the way a speedometer and a map are a pair: the threat assessment tells you which road you are driving on, and the ratio tells you how hard you can press on that road before you run into it. If you are holding the front rather than attacking it, the coastal boat defense guide is the defender’s version of the density read, paired with the winning overtime guide for the 2x max-speed rule in the endgame. The defensive pair matters most in the last two minutes of a match, when the question flips from “what push can I afford” to “what push can I survive,” and the overtime guide is the one that owns that flipped question. If your question is the economy side of “when can I fund the next push,” the population growth guide is the growth-curve page that the third question of the framework runs against, because the growth window that funds the second click is a population number, and the growth guide owns that number. The third question of the framework is the one most often answered wrong, and it is wrong most often because the player is estimating the growth window from memory instead of from the curve, and the curve is the thing the growth page exists to make visible. This page is the ratio page: it answers “what percentage, and when do I click,” and it hands the other four pages their questions the moment those questions become yours. The last thing to internalize before you return to the game is that the slider is a computed number, not a setting, and the three questions are the computation; run them on every push and the percentage takes care of itself, and the four pages above are the references that keep the three questions honest when the situation on the map stops being the situation you expected.