Mech Arena

Reading intent in Mech Arena

Nine by seven, three waves, eight hit points

The arena is nine columns by seven rows with four fixed obstacles scattered across it. You start with eight hit points and face three waves of enemies that enter from the top row.

The waves escalate deliberately: two scouts, then three units including a tougher bruiser, then four units with two bruisers. Scouts fall to a single hit; bruisers take two.

The turn structure is the important part. Time only advances when you act. Each enemy displays its intended action, you move or fire, and only then do those intentions resolve. You always see the incoming attack before choosing your response.

Nothing here is unavoidable

Because enemy intent is published before you commit, every point of damage you take was avoidable in principle. An enemy marking a tile is announcing that it will strike that tile — if you are standing there when the turn resolves, you chose to be. Reading the markers before acting is not an optimisation, it is the entire defensive game.

That turns each turn into a concrete question: which tiles are threatened next, and which safe tile also improves my position? Moving purely to safety is easy; moving to a tile that is both safe and lines up your next shot is what wins waves without losing health.

Use the obstacles. Four fixed blockers break the arena into lanes, and standing behind one limits the angles you can be attacked from. Against the four-unit third wave, a position only two enemies can reach beats trying to out-damage all four in the open.

• Kill scouts first when a wave mixes types — they die in one hit, so removing them reduces the number of incoming intents fastest.
• Never end a turn on a marked tile unless the trade is deliberate and you have counted the damage.
• Fight near obstacles rather than in open rows; limiting approach angles matters more than dealing maximum damage.

What each action is worth

Damaging an enemy scores, destroying one scores considerably more, and clearing a wave pays a bonus larger than either. Because bruisers take two hits, destroying one banks two damage scores plus the destruction bonus, making them worth more in total than a scout.

The wave bonus is the reason to press rather than stall. Clearing a wave is worth twice what destroying a single enemy is, so playing for the clear beats grinding out safe individual hits.

Because there is no healing and no time pressure, the practical ceiling on your score is set by how cleanly you clear the third wave rather than by how fast you play. Four enemies acting each turn against eight hit points leaves very little room for a traded hit, so the points you bank there depend almost entirely on how much health survived the first two waves.

Common mistakes

Every one of these treats the game as real-time when it is not.

• Acting before reading the intent markers. They are shown precisely so you can avoid the damage; ignoring them converts a solvable puzzle into a guessing game.
• Trading hit points for damage. You have eight and no way to recover them, so a hit taken in wave one is still missing in wave three.
• Focusing bruisers first. They take two hits, so removing a one-hit scout reduces the number of enemies acting against you twice as quickly.
• Fighting in the open rows. The four obstacles exist to limit approach angles, and the third wave is much harder without one at your back.

What is different about this version

Enemy intents are computed and displayed before your turn, not rolled after it. That is a deliberate design choice recorded in the engine itself: every incoming action is made visible so the player can respond to it. There is no hidden roll deciding whether an attack lands.

Time advances only on your action, so pausing, reading, or walking away costs nothing. Combined with the published intents, that makes the entire game analysable at your own pace — and means a lost run can be traced to a specific decision rather than to variance.

Written by the Silverweb Games team from this game’s own implementation. Last reviewed July 31, 2026.