How to think about units, roles, and builds in a game where everyone's roster is rolled, not chosen — and where no official stat table exists to rank against.
Because units arrive by rolling, the durable skill is assigning jobs and reading failures — every named-unit tier list without a source is noise, and every role plan survives updates.
Grounded in the official description plus dated community vocabulary; anything not enumerable in sources is labeled as such.
The only placement entity the official description names. Units come from rolling — automatically as you fight, per Eurogamer's September 29, 2026 description — and their value is the job they solve: group coverage, single-target pressure, a support effect, or income. No verified per-unit stat table exists in any dated source.
Traits are the documented progression layer attached to units, bought with Crystals (Pro Game Guides, July 6, 2026). Community sources add per-unit stat grades and Scroll-based reroll claims; the exact tables are not documented anywhere checkable, which the reroll-stats guide treats as a labeled boundary rather than a guessing game.
Waves of regular enemies plus a boss pressure test per run — the two failure modes every map produces. Wave counts, boss health, and enemy types are not enumerated in the official description or any dated tracker source, so the build answer is coverage of both failure modes rather than a stat check nobody can publish.
Servers hold up to 4 players (Roblox API maxPlayers = 4), so a map run can be solo or a small co-op defense. A squad is the six-ish set of jobs you field; the squad planner checks it for the two coverage requirements plus the over-stacking warning (too many supports).
The framework behind this site's tier list, checked 2026-10-09. Stats are omitted because no dated source publishes them; the judging criteria below are observable in any run.
| Role | Job | Look for | Swap when |
|---|---|---|---|
| Wave clear | Deletes groups of regular enemies before they reach the base. | Wide coverage or multihit attacks, an attack cycle that does not stall, and upgrades that stay affordable through the mid-waves. | Normal enemies leak on repeated runs while your boss damage is idle — your clear unit is under-leveled, misplaced, or outscaled. |
| Boss damage | Melts the single high-health boss health bar before its timer or push ends the run. | Sustained single-target output, enough range to stay on the boss, and an upgrade path you can afford during the run, not after it. | Waves are fine but the boss walks through — add or deepen a dedicated single-target slot. |
| Support / control | Slows, stuns, buffs, or otherwise makes the two damage roles better than a third damage unit would be. | An effect that scales with the rest of your lineup. One good support usually beats two mediocre ones. | Half your filled slots are support and runs still fail the damage checks — cut back to one. |
| Economy | Improves your income or roll flow so the team comes online earlier in the run. | A visible income effect early in the run. If the boost does not change when you can afford your next upgrade, the slot is wasted. | Your runs are lost to damage checks, not to being broke — trade the slot for damage. |
| Flex | The sixth-slot experiment: a second boss unit, a map-specific answer, or a unit you are testing for a future promotion. | Whatever the current map punishes you for lacking. | It has not earned its slot in three consecutive runs — promote the bench unit you have been meaning to test. |
Classify your roster by wave clear, boss damage, support, economy, and flex. A unit can do two jobs, but writing down its primary purpose makes duplicated and idle slots obvious — and makes advice from other players testable (“replace your weakest flex” beats “use the best unit”).
A build is a sequence of in-run purchases. A unit that only comes online after upgrades you cannot afford until the boss arrives is not a boss answer — it is a bench item with good marketing. Compare coverage, attack uptime, and the moment in the run where each slot starts solving its job.
Start from one dependable wave-clear slot and one sustained boss-damage slot. Add support or economy only while both checks still pass — the planner's warning fires when support reaches half your filled slots because that is the classic way balanced squads rot.
No. The developer publishes no roster with stats, and the game's rolling system means every player's collection differs. Community tier lists exist but name units without checkable sources, so this page builds plans around roles — wave clear, boss damage, support, economy, flex — which you can verify in your own runs.
Traits are the documented Crystal sink: Pro Game Guides explains that Crystals unlock traits for your units. Community sources also describe per-unit stat variability, and searches surface Scroll and Perfect Scroll claims, but no dated source documents exact grade tables — the reroll-stats guide tracks that boundary.
The role shape, not a named unit: one affordable wave-clear anchor, one sustained boss-damage unit, at most one support, and flex slots for whatever the current map punishes. Upgrade whichever role failed the last run. This survives every update because it does not depend on a stats table that does not exist.
When it stops solving its assigned job. Three runs of being carried means the slot is under-performing; a repeat failure type (waves, boss, income) names the role to change. Replace one slot at a time so you can attribute the improvement.
Rarity gates what rolls can deliver — the official description says luck raises the chance of rarer units — but rarity is not a build plan. A rare unit that solves nothing your squad lacks is bench; a common-role unit that fixes your leak is carry.
Wiki pages should point players toward the practical pages that explain what to claim, compare, build, or check next.
