Compare defense by cost and outcome

Make the routine raid unattractive

Start with a practical goal: protect the production that accumulates between collections, together with vulnerable stationary assets such as satellites and crawlers. Defense is fixed to its location. A stronger bunker does not make a parked fleet or a research savings bank safe, and it cannot replace fleetsaving.

Think like a raider. Their possible return includes loot and recoverable debris; their costs include lost ships, flight fuel and any missile preparation. Your job is to make ordinary attacks unattractive at the amount you actually leave behind. Even then, someone may attack for rivalry, military destruction or amusement. Unprofitability reduces an incentive; it does not grant immunity.

Defense does not slow an incoming fleet's timer. It may make an attacker choose a different fleet or abandon the idea, but once the attack is flying, your response window comes from the mission timing. Keep movable wealth mobile, collect regularly, and preserve fuel for leaving.

Read the universe before choosing a build

Record the separate ship-debris and defense-debris percentages. Zero defense debris does not mean zero profit: your ships, satellites, crawlers and stored resources can still be the prize. In a defense-into-debris universe, the bunker itself can contribute to the attacker's return. Do not apply the ship percentage to defenses or calculate defense debris from gross battle casualties without checking the simulator's repair treatment.

Also record whether deuterium enters debris, hostile fleet speed, fuel consumption and ACS availability. Faster hostile fleets change your response window. ACS can change the scale of fleet that reaches you. After a merge or settings change, re-run your tests instead of assuming last month's defense target remains useful.

Use opponents that actually exist nearby. A layout tested only against cargo raids tells you little about cruisers, bombers, a mixed fleet or Deathstars. Keep those scenarios separate so you can see which threat the next investment addresses.

Give each part of the defense a job

The cheap layer spreads random targeting across more units. Heavy guns supply stronger individual hits. That is a useful planning model, not a physical front line: an attacking shot can randomly hit a Plasma Turret even while thousands of Rocket Launchers remain. A large quantity of weak guns also cannot automatically overcome a shield threshold.

Understand rapid fire before adding more fodder

Rapid fire is a chance to take another shot in the same round after hitting an eligible unit type. It is not a damage multiplier or a promise to destroy that many units. For a rapid-fire value r, the repeat chance is 1 − 1/r. At r = 10, the chance is 90%. The extra shot selects a target again; it need not hit the same unit. Destruction of the first target is not required.

The familiar examples are Cruiser versus Rocket Launcher at 10 and Destroyer versus Light Laser at 10. An Ion Cannon has rapid fire 2 against a Reaper. Mixing targets can interrupt a particular attacker's firing chains, but only if the added type is not another favorable target for that attacker. Simply adding one of everything is not an optimization.

Bombers have broad anti-defense rapid fire, including Gauss Cannons and Plasma Turrets in the post-version-7 rules. Plasma avoids Deathstar rapid fire; that does not make it immune to Deathstars or all rapid fire. Read the current in-game rapid-fire panel when testing a matchup. Very old tables predate ship additions and balance changes.

Shields protect their own unit

Each combat unit has its own shield and hull. A dome is another target with a large shield, not a shared force field that lends shield points to every turret. In the standard combat model, shields regenerate between rounds while hull damage persists through that battle. Surviving units finish the battle repaired; destroyed defenses are handled by the separate reconstruction step.

Small shots can bounce without reducing a strong shield. Compare one shot with the target's effective shield, rather than summing your entire fleet's attack. Older references disagree at the exact 1% boundary, and simulator authors have documented a less-than-or-equal implementation. Avoid planning a valuable fight around equality; check the current simulator and relevant live results.

Weapons, Shielding and Armour Technology change the corresponding stats. The classic technology multiplier is 1 + 0.1 × level. Structural integrity is not the combat hull number: the classic hull starts at structural integrity divided by ten. Modern class and lifeform bonuses must also be included, preferably through a current import.

Count permanent losses after reconstruction

Defenses destroyed in ordinary combat normally reconstruct at about 70%; the Engineer raises the normal recovery figure to about 85%. These are baseline recovery expectations, not guaranteed counts for a small batch or a substitute for checking current modifiers. The Engineer halves the expected permanent share of destroyed defense from 30% to 15% under that model. IPM destruction does not receive this reconstruction.

Losing the battle does not itself erase every defense. Count what was actually destroyed. For each type, remaining = starting − destroyed + repaired. Do not forget reconstruction or add it twice to a result that already includes it. Keep ship wreck-field recovery separate.

Repeated waves still cause attrition. Re-simulate the next wave with the post-repair remainder, not the original bunker. New construction, replenished missiles, changed fleets and different research can all invalidate a simple forecast. A draw can still be costly if the attacker loses nothing and your defenses suffer permanent losses.

Protect the bunker from missile preparation

IPMs attack defenses independently of a fleet battle. ABMs intercept automatically at one interceptor per incoming missile, consuming both; they do not stop attacking ships. Surviving IPMs bypass shields. The attacker’s Weapons Technology and defender’s Armour Technology matter. Neither a dome nor the Engineer’s ordinary reconstruction bonus provides missile protection.

Missile Silo level 2 unlocks ABMs; level 4 unlocks IPMs with the relevant technology prerequisites. Capacity is ten slots per silo level: an ABM uses one and an IPM uses two. A moon has no missile silo, but its planet's ABMs also defend it. Keep enough capacity and resources to replenish protection rather than assuming one full silo lasts indefinitely.

After interception, the selected primary defense type is attacked. Do not rely on a guaranteed spillover order after it is exhausted. The archived tutorial describes random follow-on targets, while a later developer discussion disputes a fixed order and says the full algorithm was not published. Exact mixed-defense missile predictions deserve verification.

Use missile formulas as estimates, not launch guarantees

The classic range is 5 × Impulse Drive level − 1 systems within the same galaxy. Impulse 6 therefore gives a 29-system range. The official circular-universe announcement also allows IPM range to wrap when circular solar systems are enabled. Circular galaxies alone do not let a missile cross into another galaxy. Check actual target availability in Galaxy view. Hostile fleet speed affects missile flight time; archived guides disagree on the base expression, so use the current dispatch timing.

A classic damage estimate is surviving IPMs × 12,000 × (1 + 0.1 × attacker Weapons level). Compare that with effective defense hull, not shield points or the displayed structural integrity without conversion. This is an estimate from the older mechanics model: verify applicable modern bonuses and rounding before using it for an exact kill count.

Missiles are spent even when intercepted and cannot be recalled after launch. For a defender, the useful lesson is that armor, stocked ABMs and the ability to replenish all matter; shield upgrades alone do not solve missile exposure. A determined attacker can send repeated salvos or coordinate several launch sites.

Develop defense by measured need

1. Write the job: protect a particular collection interval and specified stationary assets. Set a resource ceiling that leaves room for mines, research, cargo and flight fuel.
2. Check the technology tree and purchase a small working defense rather than racing to an expensive endgame recipe. Add affordable cheap units and consider domes as they become available. Keep the early economy moving.
3. Introduce heavier guns when they change the outcome against relevant attacks. Test the next batch against the same threats and stock level used for the baseline.
4. Compare resource values consistently. If using a hypothetical 3:2:1 trade ratio, metal-equivalent cost is M + 1.5C + 3D. State the ratio and also inspect the real Crystal and Deuterium requirement; equal totals can hide a resource bottleneck.
5. Compare more units with a combat-research or defense-lifeform upgrade. An upgrade can improve existing defense, but nominal percentages alone do not show whether it crosses a useful survival or damage threshold. Include cost, completion time and the locations actually affected.
6. Prepare ABMs as defense value and local missile risk grow. After each completed batch, update the saved layout and re-test. Stop expanding when the next spend does not materially improve the protection you need.

Make the simulator answer the real question

Use a simulator maintained for the current game version, with the in-game simulator as a useful starting point. A tool being listed in an old guide is not evidence that it still works. Large-battle calculations and simulator inputs have received fixes, including live version 12.5.4 in January 2026. A PTS announcement alone does not establish behavior on your live server.

Import a recent report where possible, then inspect what arrived. Missing research or bonuses must be treated as unknown, not as zero. For a costly decision, compare several runs and plausible unfavorable inputs. Record variation in losses and draws, not just the prettiest average. A narrow simulated margin is a reason for caution.

1. Check every attacking and defending fleet, all defense counts and both domes. Include satellites, crawlers and ships that will actually be present at impact.
2. Verify Weapons, Shielding and Armour levels, player class, relevant lifeform bonuses and Engineer status. Avoid manually adding a bonus that the import already includes.
3. Check ship and defense debris percentages, deuterium-in-debris, fuel settings, coordinates and loot assumptions. The resources present at impact may exceed those on the earlier spy report.
4. Compare attacker win/draw/loss, attacker ship losses and fuel, debris they could recover, your permanent losses after reconstruction and the cost to restore production assets.
5. Run cargo raids, local fast fleets, bomber-heavy attacks and Deathstar scenarios separately. For follow-up waves, carry the post-repair result forward and account for missile preparation first.
6. Save the assumptions and compare them with real reports when available. Investigate a mismatch before committing to another large build based on the same model.

When an attack is already coming

Confirm the destination, arrival time and what is actually exposed. Move ships and resources through a suitable fleetsave while you still can; do not wait for a new defensive batch to rescue a clearly losing position. Verify that any construction you are counting on finishes before impact.

Check the missile inventory after a salvo, not just the old espionage or build screen. Ask alliance members for help only with enough time and accurate details to coordinate. A defending fleet can change both combat and debris profitability, so it belongs in the simulation rather than being treated as free protection.

Afterward, read the report and inspect the location. Record actual repaired defense, surviving production assets, missing cargo capacity and replacement costs. Repair the operational problem first. The goal is to resume a sustainable routine without leaving the same profitable opportunity behind again.

Hold the budget and attacker constant

Make one change at a time. If you compare two cheap layers, keep the heavy guns, domes, technologies, stationary ships and resources identical. Price both alternatives in the same resource valuation. Otherwise you may simply be proving that a more expensive layout performs better.

For an equal raw-cost comparison, 1,000 Rocket Launchers cost 2,000,000 Metal, while 1,000 Light Lasers cost 1,500,000 Metal plus 500,000 Crystal. Both total 2,000,000 unweighted resource units, but they are not equal if Crystal is valued more highly. At a 2-Metal-to-1-Crystal planning valuation, the lasers cost 2,500,000 Metal-equivalent. Eight hundred lasers then match the 2,000,000 Metal-equivalent budget.

Run both candidates against a cruiser-heavy fleet, a destroyer-heavy fleet and the mixed force most plausible near you. A change that interrupts one rapid-fire chain may worsen another. Retain a choice only when its benefit matches your threat and replacement budget.

Test the attacker’s recovery assumptions

Estimate what the attacker can actually take: usable loot plus debris they recover, minus fleet replacement cost, fuel and missile preparation. Use the same resource valuation on every term. A positive number in a simulator is not the same as debris safely collected.

Suppose the simplified values are 6 million loot, 4 million debris, 5 million losses and 1 million fuel. Full debris recovery gives 4 million net. Half the debris gives 2 million. No debris gives 0. If the target moves the loot as well, the operation becomes a 6-million loss under these assumptions.

This is a sensitivity test, not a prediction of enemy behaviour. Some players attack for reasons other than profit. It is still useful for deciding whether leaving extra resources or cargo behind turns an otherwise marginal attack into an attractive one.

Separate the first battle from restoration

Record both military survival and how long the economy needs to recover. A layout may preserve most defense after reconstruction but lose satellites, crawlers or cargo that kept production and collection working. Keep those replacement jobs and resources in the plan.

Use post-repair defense counts for each follow-up wave. Normal reconstruction, Engineer recovery, ship wreck fields and IPM damage are different systems. Never add an assumed repair percentage to a simulator output that already includes it.

A practical comparison records attacker losses, your permanent loss, surviving power and production assets, time to restore collection, and the amount still exposed afterward. This often gives a clearer next purchase than the win/draw label alone.