OpenFront Outgunned Defense: Hold, Whittle, or Fall Back When a Bigger Neighbor Pushes You

Direct answer: do not answer a big push with your whole bar

The single highest-value move when a much larger neighbor sends a full push at your border is the one you want to do and the one you should not do. You want to answer it; you should not answer it with your entire committed stack at the same moment. The in-the-moment fork is three ways, and you pick the branch by reading the ratio the incoming push represents rather than by reacting to how much of your own bar you feel like spending. The ratio is the defender’s total army divided by the size of the push coming at you, and because the attacker’s per-tile loss is clamped at a 0.6 floor and the attacker’s push speed is clamped at a 0.82 floor, a neighbor who is two or three times your size is not pushing for free and is not pushing at a scary extra speed either. That single correction, that the community’s “two times their troops so they can’t touch me” intuition is wrong and the community’s “they are overwhelming me” fear is also wrong, is the whole reframe this page is about. If their total army is meaningfully larger than yours, the winning move is almost never a head-on full counter; it is to hold your bar behind a wall, whittle the incoming push with small five to fifteen percent clicks as it crosses, and only commit a real counter after the push has spread and thinned. If you are down by more than about two to one, the correct branch is often to fall back behind a defense post or to concede the weaker front and defend the one you can actually hold. Everything below is the mechanism behind that fork and the specific numbers that tell you which branch is yours on your screen.

The reason this is a decision and not a fight is that the defender’s cost of simply existing is fixed by the game and is independent of the attacker. You bleed your average troops-per-tile, your total troops divided by your number of tiles, no matter whether the neighbor sends a probe or sends everything. So the fight is never “who spends more”; it is “who’s bleed is cheaper and whose regrow curve is on the fast side of the peak.” Once you read those two numbers, the three branches stop feeling like a guess and start being a lookup. The rest of the page is written so you can run the arithmetic against your own two population bars and your own tile count in a match, not so you can file it away as theory, and it is organized around the exact sequence you should follow the next time a bigger bar starts moving toward your border.

The defender’s bleed is the whole game

The number that decides an outgunned fight is not the size of the push; it is the bleed, and the bleed is the defender’s total troops divided by the defender’s tile count, computed fresh every tick and applied to every tile the attacker conquers. In the v0.34.24 source this is a one-line formula: defenderTroopLoss = defender.troops / defender.numTiles. The detail that most players miss is that this value does not reference the attacker’s stack at all. A neighbor with 220,000 troops pushing your 580-tile territory, and a neighbor with 80,000 troops pushing that same territory, make your bleed the identical number, your 110,000 divided by your 580, which is roughly 190 troops per tile. The attacker’s size changes how fast their push arrives and how much they lose per tile; it does not change how much of your average army you surrender for each tile they take. So when you feel “overwhelmed” because their bar is huge, the actual thing you are losing is not scaling with their bar. You are losing your average density, and your average density is a function of your own army and your own land, not of them.

This reframes the whole defensive problem into the two things you can actually control. The first is your own density: because bleed is troops over tiles, the fastest way to cut the absolute troops you lose per conquered tile is to raise the number of tiles behind the same army, that is to keep land and not cede it cheaply, and the second is your own total: because the bleed is your total over your tiles, shrinking your total and your land at the same time keeps the ratio roughly flat while you are still losing the land you need to regrow. The two levers point in opposite directions, and the correct outgunned-defense play is the one that holds your land so the tile count does not collapse, because the moment your tile count collapses your average density spikes and every subsequent tile the attacker takes costs you a jump, which is exactly the spiral that ends a match. Holding land behind a wall is therefore not a passive act; it is the active management of the one number the attacker cannot see or force, your average density, and the entire decision fork that follows is about how you spend your bar to keep that number from spiking while their push arrives.

The panic-send trap and the 0.6 loss floor

The instinct that loses outgunned matches is to click-attack the instant the enemy bar fills, to put your entire committed stack across the border to “stop them.” That instinct is a trap, and the trap is that you are answering a push that is already priced at the cheap, fast end of the curve with a stack that is priced at the slow, expensive end. Run the arithmetic on a 2:1 situation, you hold 110,000 and they commit 220,000 at your border, so the ratio the game reads is your 110,000 over their 220,000, which is 0.50. Their attacker loss multiplier is within(0.50, 0.6, 2), and because 0.50 is below the 0.6 floor the clamp returns exactly 0.6. Their push therefore loses 0.6 times a parity push per tile, and their push speed is within(0.50, 0.82, 7.5) / 8.55, and because 0.50 is below the 0.82 floor the clamp returns 0.82, which is a push at roughly 1.22 times parity speed. So the “overwhelming” push is only 0.6 times a parity cost and only 1.22 times parity fast, and past 2:1 the attacker’s loss and speed cannot improve at all because both are pinned at their floors. That is the correction to the community’s “two times their troops so they can’t push free” line, which the source flatly contradicts; a two-times army is not a free push, it is a 0.6 cost push at 1.22 speed, and a three-times army is the identical 0.6 cost at the identical 1.22 speed because the floors already cap them.

Now the second half of the trap, what your full counter actually does. When you answer their 220,000 with your own 110,000 full stack, the ratio that prices your push is their total 110,000 over your 110,000, which is 1.00, and your loss multiplier is the full 1.0 with no floor saving you, and your speed is the full 1.0 with no speed discount, because you are at parity in the worst possible direction. Your counter and their push meet in the merge, and the side whose push is 0.6 cost and 1.22 speed simply outruns and outsurvives the side that is 1.0 cost and 1.0 speed, so your counter is nullified and you have spent your entire army to cancel a push you could have whittled with a fifth of it. The counter that actually works is the delayed counter, the one you run after the push commits and its troops spread over newly conquered tiles, when the stack crossing the border is thinner than their headline number and your ratio against the remainder is finally fundable. The panic-send trap is therefore not “the enemy is too strong”; it is “you met a floor-capped push with a non-floored push,” and the fix is to never commit a full stack at the moment of arrival, which is the single highest-leverage defensive habit on this page.

The regrow sweet spot: 42% of cap is faster than full

The number that makes “buying time” a real weapon instead of a resignation is the shape of the regrow curve, and it is the least-known fact on this page. Troop growth in v0.34.24 is (10 + troops^0.73 / 4) * (1 - troops / max), a curve that starts fast when you are empty, peaks somewhere in the middle of your bar, and flattens to zero as you approach your cap. The peak is at roughly 42.2 percent of your maximum, the point where the raw regrow rate times the (1 - troops/max) decay is at its maximum. The consequence is the counterintuitive one that the whole outgunned-defense fork depends on: a defender who has been pushed down to about 40 to 45 percent of their cap regrows troops faster per tick than a defender who is sitting at 90 percent near their cap. If you are outgunned and you let a push take you from 100,000 down toward 45,000, you have not merely lost troops, you have moved yourself onto the fast side of the regrow peak, so the time you are buying is not just delaying the attacker, it is re-arming you faster than you would have re-armed from a full bar.

This is why the community’s “40 percent regrow threshold” is a real mechanic and not folklore, and it is the number that separates hold from surrender. If your total drops below the 42 percent peak and keeps falling toward empty, you are on the wrong side of the curve and regrowing slowly, which is the sign that you should fall back and stop ceding land rather than keep holding a wall that is draining you past the peak. If you are holding at 50 to 70 percent, you are near the peak and regrowing at close to your maximum rate, which is the zone where holding is genuinely profitable because every tile the attacker takes costs them their 0.6 floor loss while your bar is regrowing at peak speed behind you. The practical move is to manage your bar to the 40 to 55 percent band during an outgunned fight, to spend down to the peak rather than all the way to the floor, and to use the fast regrow on the near-full side of the curve to outlast a push that is already pinned at its 0.6 cost and 1.22 speed. The attacker has a fixed cost and a fixed speed; you have a curve, and the side with the curve wins the waiting game if you keep the bar on the right side of the peak.

The decision table: hold, whittle, or fall back

The fork earns its keep only as a lookup, so here is the v0.34.24 arithmetic on the four postures you actually meet when a bigger neighbor is coming, on plains with magnitude 80, no defense posts on the front, and no fallout. “Their total” and “your total” are the two population bars, “your : their” is the ratio that prices the incoming push, and their two columns are the clamped loss multiplier and the push speed the game actually applies to the attacker. “Your bleed” is your troops over your tiles, the average density you surrender per conquered tile, and it is the same on every row because it does not depend on the attacker. Reading the attacker’s two columns against your bleed is the whole point: the push stops being scary once you see it is pinned at a 0.6 cost and a 1.22 speed, and the branch you pick is just which side of the table you are on.

Posture	Their total	Your total	Your : their	Their loss	Their speed	Your bleed	Correct call
You are bigger	100,000	140,000	1.40	1.40x	0.71x (slow)	200/tile	commit, counter
Near-parity	120,000	110,000	0.92	0.92x	1.09x	190/tile	whittle small clicks
Outgunned 2:1	220,000	110,000	0.50	0.60x (floor)	1.22x (fast)	190/tile	hold + defense posts
Outgunned 3:1+	330,000	110,000	0.33	0.60x (floor)	1.22x (fast)	190/tile	fall back / concede front

The first row is the only one where a full counter is the obvious answer, because your total is the bigger number and the ratio of 1.40 puts the attacker at a 1.40 cost and a slow 0.71 speed, so your head-on push outruns and outsurvives theirs. The second row is the default fight this page exists for: you are slightly smaller, their push is a 0.92 cost and 1.09 speed, just over parity, so a full counter is a coin flip and the right play is to whittle the push with small clicks and hold your bar. The third row is the classic outgunned case, and it is the row the table exists to prevent you from misreading as hopeless: their push is pinned at the 0.6 floor and the 1.22 speed, so holding behind a defense post and whittling is fundable, the attacker is not getting a free push, and the 42 percent regrow behind you is re-arming you faster than they are spending. The fourth row is where holding stops being fundable, because past 3:1 the attacker is still at 0.6 and 1.22 and you are bleeding the same 190 per tile but your bar will drain past the regrow peak, so the correct branch is to fall back behind a post or to concede this front and defend the one where your ratio is better. Add the two real-fight adjustments: if the front holds a defense post, the attacker’s tile cost is multiplied by 5 and their advance slowed by 3, which moves the 2:1 and 3:1 rows firmly into the “do not push this front” zone, and if the defender is a bot the attacker’s loss is cut by 30 percent, which is why the bot neighbor is the front you re-route to rather than the one you hold against.

Defense posts, the 5x wall, and when to fall back

The defense post is the tool that turns a 2:1 outgunned fight into a holding action, and it is the one piece of infrastructure you should already have on the front before the push arrives. In v0.34.24 a defense post multiplies the attacker’s tile cost by the defensePostDefenseBonus of 5 and slows their advance by the defensePostSpeedBonus of 3, with a range of 30 tiles and a build cost that scales 50,000, 100,000, 150,000, 200,000 and 250,000 for the first through fifth post, each over 50 ticks. The two multipliers do different jobs. The 5x cost multiplier means the attacker spends five times as many troops per tile to take the posted front, so the push that was a 0.6 cost on open ground becomes a 3.0 effective cost on the posted front, and the push that was only marginally fundable is now strictly not fundable. The 3x speed slowdown means the push that was at 1.22 times parity speed is dragged down toward parity or below, so you are not fighting a push that is arriving faster than you can react to it; you are fighting a push that is arriving at a speed you can whittle.

The decision of when to fall back is the part most players get wrong, and it is the flip side of the 42 percent regrow peak. You should hold the posted front as long as your bar stays at or above the 40 to 50 percent band, because that is where your regrow is at peak speed and the posted front is absorbing the attacker at a 3.0 cost. The moment your total drops below the 40 percent peak and you can see the bar continuing to fall, the holding action has become a drain and the correct move is to fall back, to pull your bar off the posted front and re-form behind a second post or on the front where your ratio is better, rather than keep feeding the 5x cost until your average density spikes and the whole territory tips. Falling back is not losing the fight; it is moving your bar onto the fast side of the regrow curve and onto a front where the ratio is fundable, and it is the difference between the outgunned match that you lose by slow bleeding and the one you win by surviving to the point where your regrow curve and the attacker’s 0.6 floor cost finally flip the merge in your favor. The post is the wall, the 40 percent band is the line you do not cross, and the re-formed front is where you actually win.

Two-front squeezes and the map and mode shifts

The worst version of the outgunned-defense problem is the two-front squeeze, where two neighbors push at you at the same time and your bar can only be in one place. The source math says the answer is to split your bar by the two ratios and only hold the front where the ratio stays inside the fundable band, which in practice means you pick the front you can actually win, the one where your total over their push is closest to parity, and you concede the other. Holding two fronts at once against two pushes is holding two walls at the 190 per tile bleed with a bar that is being drained from both sides, and the moment one front is a 3:1 ratio and the other is a 0.9 ratio, the 3:1 front is the one you do not hold; you fall back behind a post on it or let it cross, and you put your whole bar on the 0.9 front where a whittle is actually fundable. The two-front case is therefore not a harder version of the same fight; it is the same fork run twice with one extra constraint, that your bar is a shared resource and you must starve the worse front to feed the better one.

Map and mode shift the numbers, and the shifts are where the specific knowledge lives. On coastal maps the front you defend is shorter than the front you attack, so your density on the coast is higher than your average, which makes the posted coastal wall the strongest version of the hold move and the open-land side the front you fall back to. On mountain and highland terrain the magnitude term rises from 80 on plains toward 100 and 120, a 25 to 50 percent increase in every loss number on this page, and a defense post on a mountain tile multiplies that further, which is why the posted mountain front is the one you never meet a push on directly; you fall back off it before the attacker reaches it. In the endgame and overtime the two ratios flip toward the player who has been winning, because that player’s total is the bigger one and the large-territory attacker bonus is strongest for the biggest territory on the map, so the overtime push is the parity push into the lowest-density front your opponent holds, and the full kill push is the one moment on this page where the slider goes to maximum. Run the fork on the overtime push exactly as you run it on any push, and the outgunned endgame stops being a wall you bleed through and starts being the arithmetic it always was.

Where to go next

The page you read next depends on which side of the outgunned line you are on and which of the two numbers you now own, the bleed and the regrow peak, and the order matters less than making sure each question this page defers gets answered by the page that owns it. If your question is the exact ratio itself, what army to commit and when to click it, the attack-ratio guide is the attacker’s version of the ratio ladder this page prices, and it is the page to open when you want the slider settings that exploit the density drop between pushes. If your question is the terrain and density under the numbers, the land-combat guide is the per-tile loss reference that this page quotes from, and it is where the terrain magnitudes and the defense post multipliers come from; open it when the two-second ratio read ever disagrees with what you see on the map, because the mismatch is almost always a terrain or defense-post term this page only references but does not derive. If your question is which neighbor is the one your ratio is computed against, the threat-assessment guide is the target-selection page that decides which of your neighbors is the bigger one, and this page is the defensive answer to the army answer that page gives.

If your question is the growth side of “when does my bar finally get bigger,” the population-growth guide is the growth-curve page that the hold move runs against, and it is where the peak regrow rate comes from, the number that tells you how fast the hold move closes the gap and when it is faster to hold than to fight. If your question is when you are already near parity with one neighbor and the fight is two-sided, the near-parity-combat guide owns the symmetric case that this page is the defensive mirror of, and the defense-posts guide owns the when and where of placing the wall this page tells you to fall back behind. If your two-front squeeze is the real problem, the ffa-trapped-middle guide is the positional page for exactly that, and the winning-overtime guide owns the flipped endgame where the outgunned fight becomes the survival race. Read these pages as a cluster rather than one at a time, because the outgunned fight is decided by the ratio, the bleed, the regrow curve, and the position all at once, and no single page owns all four; this one owns the fork and the bleed, and the others own the rest.