How battles work and how to read a battle report

Why the result surprised you

Almost every new commander has the same first battle: a fleet that looked clearly stronger comes home smaller than expected, or a raid that “should” have been decisive ends as a partial success. Nothing went wrong. Astera’s combat is designed so that power decides who wins, while classes, shields, retreats and a little randomness decide what the win costs. Once you know those parts, a battle report stops being a verdict and becomes intelligence for the next decision.

This guide follows a battle in the order the game resolves it, then walks through the report in the order it is written. Every number here is a current rule value; the full tables live in how a battle resolves.

Rounds, targets and the battle pace

A battle lasts at most eight rounds. In each round both sides fire at the same time. Every armed unit picks one living armed target at random and deals damage set by its attack, research and the class matchup. If a shot destroys its target with damage to spare, the remainder carries on to another randomly chosen target. A unit that was alive at the start of a round still fires in that round even if it is destroyed during it, so the side that loses a ship never loses that ship’s last volley.

Unarmed ships such as Couriers and the Garbage Collector are protected while at least one armed escort still fights. That is why a transport-heavy raid is safe only as long as its warships last: the moment the last escort falls, the holds are exposed.

Every hit is also scaled by one battle pace, fixed at the start. The pace exists so that battles take several rounds whatever hulls fight in them; a fleet of tier-four ships does not erase a defence in one volley. Several rounds mean several chances for retreats, carry-over damage and class advantage to shape the result.

Shields, Aegis and wounds

Every mobile ship carries a personal shield equal to 25% of its maximum hull strength before research. A Dart with 77 hull therefore enters battle with about 19 points of shield in front of it. The shield starts full in each battle and does not recharge between rounds. The Ship Shields research raises the base by 25%, 50%, 70% or 100% depending on its level. Ground defences have no personal shield.

A planet’s Aegis is a different layer: an active Aegis absorbs damage before any defending unit is hit. It cannot hold a planet alone, though; without combat-ready defenders the attacker still wins. Nullifiers exist to break it: at the opening of each round their shield bonus strikes Aegis five times harder, and that extra damage does not pass into ships.

Hull wounds are permanent until you repair them. A ship pushed below 70% of its hull may break apart, and the lower its hull, the more likely that becomes. Ships at 5% hull or less cannot start a combat mission at all. After a battle, the surviving ships come home with every wound they took, which is why the Repair Station matters as much as the Shipyard. See ship damage and repairs.

Power ratio: what each advantage really buys

The single most useful number before a battle is the power ratio between the two sides at the start. The weaker side manoeuvres: the wider the stronger side’s advantage, the less each of the stronger side’s hits does, so even a clear winner keeps losing ships. From an advantage of 4× upward the gap is too wide to evade and every hit lands in full. The game publishes the average outcome at each ratio:

Read the table as a price list. An even fight is ruinous for both sides. At 1.6:1 the winner still loses about 30% of its force. Doubling the defence brings the average loss down to about 13%, and at 3:1 it falls to about 3%. If your probe report suggests a 1.3:1 edge, the honest question is not “will I win?” but “is the loot worth losing more than a third of this fleet?”.

Two things move the effective ratio without adding ships. The first is class: Skirmishers deal ×1.6 to Bulwarks, Bulwarks ×1.6 to Lances and Lances ×1.6 to Skirmishers, with ×0.63 in reverse. The second is research, which raises attack and hull strength across every ship you own. A smaller fleet with the right classes can sit on the good side of this table.

Retreats: why ships disappear without dying

At the end of rounds one to seven, wounded ships on the weaker side may retreat. A ship at or below a hull limit leaves the battle with a set chance, and both the limit and the chance grow with the power ratio at the start and with the size of the weaker side. Near 1:1, both sides may retreat. Retreating ships stop firing and stop taking hits but keep their hull damage.

Retreats matter in three places. When deciding control, a retreated ship counts as having left the battle. For Dominion and for wreckage, it never counts as destroyed. And in your report, it appears as retreated alive, not as a kill. Unarmed ships and ground defences cannot retreat.

Separately, a defender can set a tactical-retreat posture that pulls the whole fleet out before a hopeless fight. It works only with at least five armed ships, enough deuterium and an attacker at least three times stronger than the defending line. Ground defences stay behind and still fight with Aegis, so a planet in retreat posture is not necessarily empty. Defence posture explains the conditions.

Decisive, partial and repelled

Every attack ends with one of three grades. Decisive means the defending force and any active battle shield were cleared, or the target was undefended. Partial means the attack removed at least 42% of the defending resource value from combat, through destruction or retreat, without clearing it. Anything less is repelled.

The grade sets the loot share, not the score. Against a commander in tiers 1–3, decisive takes 25% of the stock above Store protection and 15% of uncollected Works; partial takes half; repelled takes nothing. Commanders in tier 4 and above expose 60% and 30% on a decisive loss. Surviving cargo capacity then caps what actually comes home. After any battle, 60% of each destroyed type of ground defence is rebuilt free, rounded down, so a defence you broke today is partly standing tomorrow.

Reading the report, section by section

The report answers four questions in the order a commander asks them.

• What happened: the grade in plain words, such as “Your raid partly succeeded”, with one line explaining why the full-success condition was or was not met.
• What you learned about the enemy: the units that defended at the start of the battle and what you destroyed from them. It is not their whole fleet. Units that were away, or that retreated, are not revealed; the report says so, and a new probe is the way to find out.
• What happened to your force: what you sent, what you lost and how many ships survived, with wounds carried home.
• Loot and point changes: resources carried, wreck resources salvaged and the Dominion you gained or lost.

Below those answers, each round can be opened to show what you dealt, what you took and which units fell, and a rules panel shows the exact calculation. Use it when a result surprises you; it usually shows a class mismatch, a full Aegis or a retreat that changed the count.

Turning a report into the next decision

Start with the score line, not the title. A raid can win the battle and still lose Dominion if your permanent losses cost more than the loot and the defender’s losses combined. In an ordinary raid, your result is loot carried plus the value the defender lost permanently, minus what you lost permanently.

Then compare the loot with your cargo. If the report shows that your holds came home full while loot remained, the next raid needs another Courier, not another warship. If the loot was small, the target may simply be empty now; a second raid soon after the first usually brings less.

Finally, read the enemy section as dated intelligence. Destroyed units are a floor on what they had, not a full list. Ground defences will be partly rebuilt. Wounded defenders may be in their workshop. Watch the target before deciding to return, and probe again if the stakes justify it. A battle is the only information in this game that is never an estimate; use it.