Hold the Flood — Enemies and Encounters

CMB101 / CMB102 · r11 · 2026-09-07 · Combat and Enemy Designer · In review. r11 makes the three corrections PMR009 asks of §0A, and two of them are withdrawals. The "sheds once" claim is narrowed: it stops re-harvesting the same body and does not stop a player being rewarded for repeated fresh arrivals, so the honest answer to the vision's cut condition is that charge never leaves the district — killing yields fuel that is discarded at extraction and is worthless unless an intervention spends it. The "ten hits is not durability" defence is withdrawn, and so is the placement rule that would have forbidden free bypasses. The Heavy is instead tested against the retained scene PMR009 names, where it fails as an obstruction and can only be optional fuel — and as written it is not earning its place there. Sources read: Design Document r14 ec1b6c §§3–6, 8, 12–16, App. A; Enemies r1 c63228; the shared-rule contract Power and Attraction r2 69024c and its r1 seed 8afc3c; Campaign 34dc09 (DES104); Design Document r15 5f9fbb; Design Review REV104 r1 80a882; Colony COL101 a1046d; Cutter; Equipment; Colony; Pumping District; World and Mystery r20 7699de; Expansion Brief; Design Index dd01c7; Tasks eb8c03; PMR003; Design Review REV103 r5; Brief; Glossary; Reference/Content Scale REF107; External Critique EXT102; Wiki Conventions. Design Index · Expansion Brief · Glossary · Combat Rules

Nothing here has been built, prototyped, simulated or played, and no asset has been located, inspected or created. Every number is a proposal. The two encounters below are written scenes checked on paper against declared rules; they are not playtests and they establish nothing about fun, balance or performance.

0A. The roster against the vision's decision test

The Game vision gives crowd combat a job — "make reaching, reading and operating machinery a survival decision; let positioning and harvested charge create room to act" — and a cut condition: "an enemy adds only durability, or a reward encourages killing indefinitely instead of choosing an intervention." PMU008 tells domain designers to answer for their systems on the Visionary's actual changes and "not defend a feature just because it is already documented". Run against the eight, before anyone asks.

The second half of the cut condition — narrowed in r11, because PMR009 is right that r10 claimed too much. r10 said the roster passes "structurally" because an organism sheds once. That rule prevents re-harvesting the same body. It does not prevent a player being rewarded for killing repeated fresh arrivals, and a surge schedule that delivers forty to a hundred and forty bodies three times an expedition supplies plenty of those. The claim as written was false.

What actually answers the cut condition is a different rule, and it is a better one: charge never leaves the district. Killing produces a resource that is discarded at extraction, so it cannot be banked, cannot become refuge progress, and is worth exactly nothing unless it is spent on an intervention in the same expedition. Salvage is what the colony wants, and killing does not produce salvage. So the reward for killing is not points — it is fuel, and fuel is only worth having if there is something to do with it. That is the vision's own sentence — "let positioning and harvested charge create room to act" — and it is the honest form of what r10 was reaching for.

The residual exposure, stated rather than argued away. Nothing stops a player spending an expedition farming fresh arrivals instead of intervening. What makes that a bad trade is that they leave with nothing to show for it. Whether it feels like a bad trade in play is not something this page can establish, and if it turns out that fighting is more attractive than intervening, the fix is on the intervention's side, not by making enemies less rewarding.

The first half — "adds only durability" — the test each role has to pass individually.

The Heavy, tested against the retained scene — and one defence withdrawn

Withdrawn in r11: "ten hits is not durability because there is no damage curve." PMR009: "Ten-hit durability still is durability even without a damage progression." That is correct and my argument was a dodge — the absence of a stat ladder changes who can shorten the ten hits, not whether the ten hits are durability. The role has to earn its place on what an encounter does with it.

Also withdrawn: the placement rule I proposed. r10 asked that a Heavy be placed only where it blocks something the player wants. PMR009 rightly refuses it — "do not forbid free bypasses everywhere to save its role" — and a free bypass is usually good design. Withdrawn without replacement of that kind.

The test PMR009 gives instead, applied to the scene it names. A Heavy earns its place in an encounter by being a useful obstruction or an optional fuel opportunity that is honestly labelled optional and has a stated benefit. The retained scene is Campaign §3.2's middle return: the objective is the Exchange doorway and its drums; the rail starts powered, the pass is open, and the stairs bypass the rail-blocking Heavy outside pursuit range except at their ends.

• As an obstruction, it fails there. The stairs go to the same destination for free. That is not a flaw in the scene — the bypass should stay.
• So in that scene it can only be optional fuel — and that only works if its five units have something to buy in that expedition. Under the rule above, charge is discarded at extraction, so a Heavy killed for fuel that is never spent is a fight that bought nothing, which is precisely the defect X37 found in the same scene's Coil plan.

So my answer is not a defence of the Heavy: in that scene, as written, it is not earning its place. It earns one only if the expedition contains something the five units pay for, and if the page says plainly that killing it is optional and what the fuel is for. That is a statement about the scene and the labelling, not a demand that Campaign remove a bypass.

The alternative I would still take if the Lead prefers it: merge the Heavy into the Anchor — one idea, a body that will not move out of your way, sited dry and wet.

Nothing else in the roster is near the line, and I am not proposing to cut anything today. No new role, no quota change (PMR008: no new roster or composition quota). If the Visionary's next pass changes what an expedition is for, the roles that would break first are the two that assume a lit-branch routing game at all — Grounder and Carrier — and I will say so then rather than wait to be asked.

0. Reconciliation with DES104, which landed while this was drafted

What arrived after this role read its sources and before this revision saved: the Lead's Campaign 34dc09 (DES104), the shared-rule contract Power and Attraction r2 69024c, Design Document r15 5f9fbb, the critic's REV104 r1 80a882, and the Colony designer's COL101 a1046d. r1 of the six role pages and r1 of Combat Rules were written against r14 and the r1 seed. This section says what that changes. It changes the name of one role and the page to cite for the rules — but the Campaign page proposes four enemy roles of its own, and two rosters is not a thing this project can afford to have.

Where the rules now live. Power and Attraction r2 is the shared-rule contract, adopted by the Lead from PM's seed and validated against r14. It is now the page to cite for conduction, power, attraction and charge, and this page and Combat Rules cite it. Nothing written here against GDD §4 is invalidated: r2 restates the same rules, states the retarget-or-spread precedence in the same words, and confirms that pursuit and press are separate things.

The one thing the contract settles in this role's favour, and it is worth naming. Contract §7 lists what a district may not add: "anything that makes a lamp draw, a coil select intake, an organism shed twice, or a device work without current." The Grounder does none of the four — it makes a device stop working for want of current, which is the rule applied, not broken. R9 passes the Lead's own published test. It still needs the Lead's and World's explicit ruling; it no longer needs an argument about whether it is the kind of thing that could be allowed.

0.1 Two rosters, and the merge

0.2 Heavy — where I was wrong, and the one clause I still ask for

R8 says there are no damage numbers, and §1 below says none of my six has more health than another. A body that takes ten hits looks like exactly the stat enemy the Expansion Brief warns about, and that was my first reading.

It is not, and the Campaign page's argument persuaded me: a Heavy is a walking large mote, five units in one pickup where the cutter earns one at a time, and it blocks a route while it does it. That is a systems job, not a difficulty knob.

Corrected in r4. r3 supported this with the Campaign page's "about nineteen kill motes before a two-bank rack's first discharge", which rested on most-room collection. D023 selects shelter-selected pickup priority instead, under which the player fills one chosen bank first — about ten kills, not nineteen. So the case for the Heavy is real but smaller than r3 made it, and it should rest on the large mote and the blocked route rather than on a number that has since changed.

The clause I asked for is withdrawn in r4. r3 asked that the Heavy's resistance be stated as "cannot be cleared by the cutter in useful time" rather than as ten hits. PMR005 declines it and is right: replacing a number with a phrase makes the weapon and crowd consequences harder to inspect without removing the durability that caused the concern. Keep the explicit provisional ten, and let a reader-facing document summarise it. The underlying worry — that a published ladder is what Equipment prices weapons against — stays as the scoped recommendation in R8, which is a recommendation and not a ban.

0.3 Static and Carrier — one role, and which half of each to keep

Both are "the thing that comes for your charge". The merge I recommend, with what each side gives up:

• Keep the Carrier's targeting sentence — the nearest reachable loose charge — because it is the beacon rule with one noun changed, so it costs the player no new concept. Static's "drawn to stored charge within a short range, over a lit beacon" adds two ideas at once: a range, and an explicit priority override. The precedence stack already supplies the override as line 2, so only the range is new.
• Keep Static's short range. This is the half the Carrier page got wrong, and I withdraw the unbounded version. Without a bound, a player carrying twenty units is hunted district-wide for as long as they hold it, which turns a good tension — power on your back is bait — into a tax on ever being rich. A short range makes it a proximity cost: carry charge and you cannot walk through a crowd, but you can still walk past one.
• Keep R5, the mote theft, which neither the Campaign page nor the Static sketch has. It is what makes a harvest a contest rather than a stopwatch, and it gives the role a second job so it is not only a punishment for being well-supplied.
• Drop "sparks toward it." A ranged attack is a new combat verb, and the closed grammar in Combat Rules §6 has none. If the Lead wants the spark, it should be specified as damage at range with its own telegraph, and it needs a ruling rather than a clause.

The Lead decides, as integrator. This page proceeds on the merged version; the only beat that depends on it is Encounter B's, where the surge enters the channel the surveyor is standing beside, so a bounded range changes nothing there.

0.4 A cue-budget concern — narrowed in r4, because r3 overstated it

r3 reported the contract's "enemy variants by size and tint" and Campaign's "readable by size, silhouette or spark, not by tint alone" as a contradiction between the Lead's two pages. PMR005 is right that they are not — both can be true — and the finding is withdrawn as stated.

The concern underneath it stands, and is smaller and more useful. Four different facts want a cue: which role a body is, whether it still has charge to shed, whether current reaches it, and whether it is about to discharge. Brightness cannot carry more than one of them. My proposal is glow = remaining yield, encrustation = non-conduction, flicker = imminent discharge, silhouette and position = role. The Lead should state which redundant cues carry which fact. No readability claim on this page is verified, and prose cannot verify one.

0.5 Two gifts from DES104 that this page did not have

• The prod (Campaign §4: "pushes and never kills") is the best answer to the Spark anyone has proposed, including me: a Spark is only dangerous when killed, and the prod is a weapon that does not kill. It also answers the Spark's hardest constraint, that the cutter cannot be withheld (R6) — the prod gives the player a way to move a body out of reach instead of shooting it. Under R6 it must publish its targeting rule: what it selects, whether it can be withheld, and what its line looks like. And one ruling I make on my own side rather than route: displacement is movement, not a status. It has no duration and nothing reads it off a body, so R7's list stays closed at three.
• The cable reel — and a correction to what r2 and r3 said about it. They quoted the Lead's original sketch, "lays a conductor between two wet points across dry ground", and used it to argue R10 mattered more. Both are wrong now. R10 is withdrawn, and Equipment U6 revises the reel to a one-cell, six-metre reel limited to exposed power terminals on an already traversable path — deliberately not an arbitrary wet-point connector, so that water and valves keep their significance. Against that definition the reel is a power-routing tool, not a conduction-topology one: it can bring a terminal's power somewhere a beacon or hatch can use it, which is a Grounder interaction, not an arc one.

0.6 The merged roster, and where its cost is accounted

Eight roles across the campaign: Walker, Floater, Heavy (Campaign's), Carrier (merged), Anchor, Crust, Spark, Grounder (this page's). The per-district cap in the composition budget is unchanged — at most three distinct non-Walker roles in a district, at most two in a surge — so no expedition asks the player to read more than four kinds of body at once.

The presentation cost is accounted once, on Campaign §7, which the Lead names as the single D025 list and which REF110's audit runs against. The answer there: the roster costs one base body and one licensed edit, with the rest poses, a scale and engine-drawn cues — not eight creatures. This page's own competing accounting is removed rather than left marked superseded (PMR007 §3 asked for the removal; the text is in page history).

If the roster must shrink, my cut order stands, and it is a design judgement rather than a costing. (1) Crust — its job, pricing greed inside a stun, is the least load-bearing of the new roles, and the once-per-organism rule already bounds the coil without it. (2) Merge Heavy into Anchor — they are one idea, a body that will not move out of your way, sited dry and wet. I would not cut further: the remaining six each delete a different assumption, and below six the roster stops covering the four clauses.

0.7 Eight roles against what comparable games ship — answering X32

EXT102's X32 counts citations of REF107, Content Scale, and records Enemies 0, Mechanics/Combat 0. The complaint is fair against this page, because the roster size is one of the numbers REF107 is about. Citing it here, with what I make of it.

What REF107 records (developers' own published lists, not this role's research): Halls of Torment ships 6 stages against 70+ monsters, 74 abilities and 60 items with 240 rarity variants; Vampire Survivors ships 5 normal stages in its base game. Its observation: "the number that is small is the environment count, and the numbers that are large are the things that differ within an environment" — and set beside this project, the two axes point in opposite directions, with a district count above both comparables and per-run variety an order of magnitude below. REF107's own caveat, which it insists on and I carry: those totals are years of post-launch accumulation, and comparing a first release to them is a category error.

Where this roster actually sits, stated plainly. Eight roles, against 70+. That gap is real and I am not going to argue it away by redefining the units.

What I think follows, as a position rather than a defence.

• The variety in those games is per-run, and it is bought mostly with the player's side of the table — abilities, items, rarity variants — not with the bestiary. This design's equivalent axis is Equipment's twelve entries and three families, and that is where a like-for-like comparison bites hardest. It is not mine to answer, but it is the honest place to point.
• On the enemy side, this design deliberately buys variety with combination rather than count: which roles are live, in what water, on what topology, under which panel setting. Eight roles under the composition budget give a bounded set of legal encounters per district, and the district's topology multiplies it. That is a design bet, not a measured equivalence, and nothing here shows it produces the felt variety 70 monsters produce.
• And there is a tension inside my own pages that this comparison exposes, which I would rather name than leave for a reviewer. The composition budget caps a surge at two non-Walker roles and a district at three — I wrote that cap to protect readability, and it is also a cap on exactly the per-encounter variety REF107 is measuring. The two goals genuinely conflict, and I still prefer the cap: a crowd of a hundred and forty in which four things are happening is not variety, it is noise, and this design's legibility is the thing every reviewer has said is its strength. But the cost is real and the owner should see it as a cost.
• If the owner wants more variety, the cheapest honest lever is combinations, not bodies — raising the per-district cap, or placing the same eight roles against more topologies — because a new body costs a silhouette from an asset budget whose first item is still an unapproved request (Campaign §7), and a new combination costs nothing.

What this is not. REF107 proposes no count and neither does this. Nothing here establishes that eight is enough, or that combinations substitute for bodies; both are unverifiable in the design phase. It is a comparison recorded where the number is set, which is what X32 asked for.

1. What a role is for, in this game

The existing proposal has one enemy type with size and tint variants (GDD §15). The Expansion Brief asks for "enemy combinations that change movement or priorities, rather than only more health or a new tint", and REV103 r5 pre-registers the test it will apply: a combination must pose a precedence question that a single enemy type cannot.

So every role here is defined by one sentence it makes false — a thing the player could otherwise safely assume about a crowd. That is the whole justification for a role's existence. None of the six has more health than another, and none moves faster.

The roster is eight, and the ownership is settled. Walker is the existing enemy re-specified as a role; Anchor, Carrier, Crust, Spark and Grounder are new here; Floater and Heavy are the Lead's, adopted, with the Floater transferred to Combat by contract r3 §7 and the Heavy still defined on Campaign. r2–r5 of this page said "six" in places while the merged roster was eight; r6 removes that drift, which PMR006 asked for. It remains an initial roster, not a quota — I would rather ship eight that each delete an assumption than twenty that each add a number, and Combat Rules §0.6 states which of the eight cost a new drawn asset and which cost a pose.

Why this set and not another. The design's spine has four clauses — conduction, power, attraction, charge. A roster that covers all four with the fewest bodies is the one that teaches the spine rather than decorating it. Anchor, Carrier and Grounder each deform attraction in a different direction (ignore it, replace its target, remove its source). Crust and Spark deform conduction in opposite directions (will not carry current, will not stop carrying it). Charge is deformed by Carrier a second time, through R5. Power is deformed by Grounder. Nothing deforms conduction inside the rack, and that is deliberate: the rack is Equipment's, and an enemy that reached into it would be reaching across an interface.

The credible alternative I rejected. The obvious sixth role, and the one most survivorlikes would pick, is a tracker: an enemy that, once it has seen you, pursues without limit and never drifts back, converting your mistakes into a growing tail. It is a good role and it would work here. I left it out because it is the one role on my shortlist that adds a clause rather than subtracting one — "some bodies ignore range" is a new fact about pursuit, where the other five are all "this body answers a different question". It is also the role partly covered by a Carrier: a player carrying twenty units acquires an escort within a Carrier's range, and the ways to shake it — spend the charge, or leave the range — are decisions rather than a footrace. Corrected in r4: r2 said "an escort that does not give up", which was true only of the unbounded Carrier that §0.3 withdraws. The bounded Carrier does give up, and a tracker would still be a genuinely different thing. If a later authorized phase shows the roster needs unrelenting pressure, the tracker is the first thing I would add and I would add it knowing it costs a clause.

2. The eight roles at a glance

Full pages carry each role's trigger, action, threat, counterplay, machinery and equipment interactions, placement conditions, telegraph, asset note and open questions. This table is a summary; the pages win.

3. Encounter A — the invited yard

Contrast: defensive, mid-expedition, one surge, preparation-light. Roles live: Anchor + Crust (2 — within the composition budget). Where: the Pumping District yard channel — the district that already exists, so the encounter needs no new geography. What it teaches: that a stun is not the same as a cleared floor.

Declared setup. The player has learned the electrician play in GDD §5 and is running that rack: coil, battery, insulation, two bundles, link battery→coil, battery full at 10 units. Panel thrown from causeway to yard at 7:30, so the yard beacon is steady and the pump-hall hatch is open. Four Anchors have been standing in the yard channel since the player arrived. The first surge — 40 bodies — has been holding at the causeway beacon since 3:00.

Checking the scene against its own rules

• Composition legal? 54 Walkers of 60 is 90%, above the 85% floor. Two non-Walker roles are live (Anchor, Crust), at the limit of two per surge. Anchors were standing before the expedition began, as their placement rule requires. No Sparks are placed, so the dry-ground rule is not engaged. ✔
• Arithmetic. 54 + 6 = 60 entering. At 8:25, 40 Walkers + 4 Crusts = 44 in the water; 14 Walkers + 2 Crusts = 16 still entering or on the bank; 44 + 16 = 60. ✔
• Stun count. Conducting bodies in the connected water = 40 Walkers + 4 Anchors = 44. Crusts excluded by their own rule. ✔
• Shed. 44 ÷ 10 = 4 complete tens → 4 large motes → 20 units. ✔ (The "one per complete ten" rounding is a proposal; the source states the ratio but not the rounding.)
• Bankable. Capacity 10, so 2 motes. This reproduces GDD §6 exactly: "a discharge spends ten, a stunned surge of forty sheds twenty, and a single battery can hold only ten of them". ✔
• Anchors shed? They were stunned in connected water, so yes — and they are therefore spent for the expedition, which is why the channel is worth nothing on a second visit that hour. ✔ Consistent with once-per-organism.
• Is the channel energised, and by what? Settled. By the panel, from 7:30, for as long as the yard is lit — the coil's two seconds sit inside that and add no hazard the player was not already in. ✔ r3 had the water go inert after the discharge, which contradicted its own setup (PMR004 found it); r6 priced both readings while the ruling already existed; r7 carries the ruling.
• Do the Anchors take damage standing in energised water all expedition? No. Under the intensity rule in §0B, energised water harms the surveyor and does nothing to organisms; only a bank-strength discharge reaches a body, and it stuns rather than damages. ✔

What the encounter is actually for

The six Crusts convert "collect everything" into "collect exactly two and leave." The third and fourth motes are worth nothing to a one-battery rack, so a player who understands the battery cap takes two and walks out before the Crusts close. A player running the harvester rack — two batteries, GDD §5 — genuinely wants all four, and must therefore accept Crust contact to get them. The composition prices the loadout without a single stat existing anywhere in it. That is the whole argument for R8 in one scene.

What preparation changes. Rewritten in r4; r3 had this wrong. Without insulation the player cannot run this encounter at all: the yard is lit, so the channel is energised, and every second in it costs one health from five (§0B's intensity rule and open question C10). The insulation slot is not a hedge, it is the entry fee — and that is the shape GDD §5's electrician rack already has. Without knowing what an Anchor is, the player throws the panel expecting a clear channel and is wrong. Without knowing what a Crust is, they step into the mote line expecting 44 stopped bodies and meet four that never stopped, and that mistake costs health rather than only the harvest, because a Crust resists the stun a spent Walker still receives.

4. Encounter B — the failing causeway

Contrast: offensive, late-expedition, the whole precedence stack live at once, preparation-heavy, and it turns on a decision made twenty minutes earlier at a shelter. Roles live: Grounder, Carrier and Spark — three in the district, at the limit; two in the third surge, at the limit. What it teaches: that a plan can expire, and that hoarding charge is itself a threat.

Declared setup. Late campaign. The panel has been on causeway since 3:00, so the first two surges (40 and 80) both entered there and are held at the causeway beacon — about 110 bodies after kills, drawn as one ring. The player is in the yard, having come by the culvert, running the harvester rack (GDD §5): coil, arc, two batteries, insulation, no cargo. Both batteries are full: 20 units. Two Grounders climbed the causeway's power cable at about 10:30. The third dispatch is due at 12:00 and will bring 140.

The four options, priced

Under (a): the late payoff, checked

Checking this scene against its own rules

• Does the Carrier's twelve-metre range change this scene? No, and here is the distance. PMR005 rightly refused to take that on assertion. The surveyor is standing on the dry bank beside the yard channel — that is where the coil is discharged from and where the motes are collected — and the surge enters the channel and wades down it past that position. Bodies in the channel are therefore within twelve metres of the surveyor for most of their walk, so the Carriers acquire and the beat is unchanged. What the bound does change is the option that is missing from this scene: under the unbounded version there was no way to shake them at all, and under the bounded one walking away from the channel is a real fourth answer — which costs the harvest, the discharge and the burst, i.e. the whole expedition's payoff. That is the choice the range exists to create. ✔
• Composition legal? 122 of 140 Walkers = 87%, above the floor. Two non-Walker roles in the surge (Carrier, Spark); three in the district counting the Grounders, at the limit. Grounders are placed on a cable run rather than counted in a surge. Sparks enter water, not dry ground. ✔
• Does the beacon go dark legally? A Grounder removes power from a run; a beacon with no power is dark; a dark beacon draws nothing. No new meaning is given to a beacon, no dispatch is caused, and the panel's monopoly on the intake is untouched. This is the argument that R9 sits inside World r20's contract, and it is the argument World should check. ✔ (pending C5)
• Does the crowd really spread, and is that step skipped? Corrected in r4. PMR004 is right that darkening a beacon does not by itself make a crowd spread: the crowd first re-evaluates line 3 and retargets to any other reachable lit beacon. It spreads only when there is none. Here there is none, and the reason is the panel: contract clause 2 makes the panel's positions exclusive — one branch at a time, or off — so with the panel on causeway the yard beacon is dark, and grounding the causeway's leaves nothing lit. That is why line 4 applies, and it is a fact about this district's topology rather than about the Grounder. ✔
• Stun and shed arithmetic. 100 in water → 10 complete tens → 10 large motes → 50 units. Coil battery capacity 10, arc battery full and being reserved for the burst → 10 units bankable at most, 40 not. ✔
• Does the burst fit inside the stun? No, and deliberately not. The stun is 8 s and the burst is 10 s, so the burst is spent from 12:33 on a recovering crowd, not a stunned one. This is the correction PMR003 item 1 required of GDD §13 and which r13 supplied; this scene is built on the corrected sequence rather than re-introducing the error. ✔
• Does the burst return charge? Corrected in r4. Not from the 100 stunned at 12:25, which are spent — but the 40 that arrived after the discharge have not shed, so chained kills among them do drop 1-unit motes. r3 said the burst returned nothing, which ignored the late cohort; PMR004 asked for the cohorts to be separated and they now are. Those motes are still lost, for a different and more interesting reason: the coil bank is full at 10 and the arc bank is being drawn, so nothing eligible can take them. ✔
• Bank states, end to end. Start: coil 10, arc 10 (20). 12:25 discharge: coil 10→0. Collection: coil 0→10 (2 large motes; the rest expire or are carried off). 12:33–12:43 burst: arc 10→0. End: coil 10, arc 0. r3 asserted two empty banks and drew a conclusion from it about the Carriers; both were wrong. ✔
• Is the closing dump legal? Yes, and it costs nothing: a discharge needs a full linked bank, the coil's is full, and contract clause 4 says charge left in a battery at extraction is lost at the dock. The 100 spent bodies shed nothing more; the 40 unshed ones shed 4 large motes. ✔

The contrast that makes the point

Run the identical composition against the arc hauler rack (GDD §5: arc, battery, insulation, three bundles). No coil, so no stun and no bulk harvest. No second battery. No beacon. At 11:40 that player cannot take option (a) profitably and cannot take (c) at all; their correct play is (d): culvert home through the thin edge, bank three bundles, and see none of the payoff.

They finish the expedition successfully. That is the design working, not failing: the rack chosen at the dock decided which game this expedition was, and both are legitimate. It is also the concrete answer to the Expansion Brief's demand for "a reason to prepare a different build, with an alternative viable approach and a real tradeoff" — the two racks do different things, not different amounts.

5. Early, middle and late — how combat grows without a number

The Cutter page asks what develops basic combat across a campaign when the weapon is fixed and there is no XP. The answer this roster gives:

Combat grows because more lines of the precedence stack are live at once, not because anything gets bigger.

What the owner should press on. Whether "four precedence lines live at once" reads as growth to a survivorlike audience the way a rising damage number does is unverified and unverifiable in the design phase. It is the central bet of this page and it is the same bet GDD §6 already makes about the arc. I believe it is the right one, because the alternative — a stat ladder — has to be invented from nothing here and would immediately compete with the thing the game sells. But it is a bet, and it is the first thing an authorized play phase should look at.

6. Elites and bosses — a recommendation against, and two substitutes

The assignment asks for an assessment only with a reason. Mine is against creature bosses and against stat elites, and for two elite-shaped things that fit this design.

Scoped, after PMR004. PM is right that the owner's brief bans none of this, and that a compact grammar is a proposal rather than proof the alternatives are bad. So read what follows as what a boss would cost this particular design and what I would build instead for less — not as a prohibition, and not as a claim that bosses are bad in general. If the owner wants one, the design can have one; the reasons below are the price list.

Against creature bosses, for three reasons specific to this game rather than to taste:

1. A boss is a damage race, and this game has no damage curve. No XP, a fixed cutter, and an arc bounded three ways (GDD §6). To make a boss killable you would either invent a stat ladder — contradicting the no-XP bet and the Expansion Brief's warning about damage-bonus catalogues — or make it a machinery puzzle, in which case it is a machinery puzzle, and the design already has those and does them better. The relay room is the proof.
2. A boss is understood once and then repeated. This design's promise is the opposite: "the same place means something different once you know what its light means" (GDD §7). A set-piece that is solved and then re-executed inverts the one thing the game is selling.
3. Asset cost. A boss is a unique body, unique animation and unique cues. GDD §14 already records one enemy sprite as the project's first likely approval request. A boss per district would multiply the one line item PM has been asked to worry about.

Against stat elites — a bigger, tougher, faster Walker whose only distinction is the numbers — because that is the "more health or a new tint" the Expansion Brief warns about. Narrowed in r4: this is not an argument against different hit counts, which R8 permits and which Campaign's Heavy uses well for a stated economic reason (§0.2). It is an argument against a body whose only answer to "what does it change" is a number.

For, instead:

(i) The declared surge. The forecast already tells the player the intake and the countdown. Let it also declare the next dispatch's composition — "third dispatch: carriers" — once the player has repaired the colony's forecast relay or has learned to read the cadence. Player action: repack at a shelter, or change the panel, before it lands. Visible information: one more row on the overlay that already exists. Consequence: you meet the surge you chose to meet. Cost: knowing costs a colony repair or a learned convention, and acting on it costs slots or a route. Concrete interaction: seeing "carriers" and deliberately spending a battery before the dispatch, so there is nothing left to walk toward. It converts an elite into a preparation decision, which is what this game is about; it gives Campaign a difficulty dial made of composition rather than numbers; it gives Colony something worth repairing; and it costs no creature and no asset.

(ii) The standing problem — a district's one persistent, named concentration that is a place, not a fight. The pattern: a district's lowest channel permanently occupied by a mass of Anchors, because Anchors go to the lowest wet ground and never leave it. It is never cleared and never respawns, because it never left. Player action: route around it, drain it with a field pump, or spend a battery on it. Visible information: it is standing there on arrival, every time. Consequence: it is the reason the district's shortest route is expensive. Cost: a bundle for the pump, or the longer road, or a battery. Concrete interaction: it is also the district's single richest harvest — one discharge, once, sheds proportionally to its size and then it is spent and worth nothing for the rest of the expedition, so when you spend the battery on it is a real decision. This is a "boss" as geography, under adopted rules, with no new creature. Placement is Campaign's; I am proposing the pattern, not its location.

The one condition under which I would revisit a creature boss. If a later authorized phase shows that the arc's payoff lands too rarely because a crowd is never dense enough at the right moment, then a single large, slow, high-hit-count body that is reliably a valid arc target would be a targeted fix for a measured problem. I would not build one before that evidence exists, and no such evidence can exist in the design phase.

7. What this page needs from other roles

Stated as positions with my recommendation attached, not as routed decisions (Wiki Conventions §11a).

8. Asset presentation gaps

Full statement in Combat Rules §9. In short, and honestly: the roster's readability costs six distinguishable silhouettes, five more than the current one-enemy proposal, and that is the largest consequence of this submission. Three of the six (Anchor, Carrier, Grounder) are readable primarily from position and attitude — half-submerged and still; leaning toward a target the forecast already draws; up on a cable and off the ground — so if the base body can be posed, those cost poses rather than creatures. Two (Crust, Spark) are silhouette edits of the base body, and whether that is legal depends on the individual asset's licence, not its host's general statement (Asset Licences). No asset has been located, inspected, downloaded, purchased or created, and no availability is claimed.

The roster is neutral on GDD §14's decision 2, the humanoid Flood: what conducts is bodies standing in water either way, and posture-based roles work at least as well on a humanoid body. Nothing here pushes the owner toward the custom-art request.

The gap I cannot close is audio: the Spark's fairness leans on a rising whine as it enters cutter reach, and sound is not designed. The visual fallback is the cutter's target line plus the flicker, which I believe suffices and cannot claim does.

Set aside, for the record

r1 of this page was PM's seeded source summary of GDD r13 §§3–4 and 15, plus the CMB101 assignment brief. Its factual summary of the existing single enemy is superseded by §§1–2 above, which now state the same behaviour as the Walker role. Its assignment text is deliberately not retained inline, because this revision is the answer to it; it is preserved in full at r1 c63228. Nothing else was on the page. The seed's ownership line — Combat and Enemy Designer intended owner, Lead interim — is now discharged: this role is staffed and this page is adopted.

Revision log

Kept short, so the sections above are the current proposal and this is the history. Full text of every revision is in page history.