Fieldwake · Build-Readiness Assessment
Fieldwake is a strikingly coherent piece of design — a world built outward from one honest premise (a planet where emotion is physics and humans aren't the apex predator) rather than retrofitted onto a combat engine. The bones are excellent. This review exists to find the small number of load-bearing pieces that aren't pinned down yet, so we fix them on paper before they get expensive to fix in code.
Almost nothing here is a design problem. The 50 flags from document ingestion are overwhelmingly our tooling (schema/linkage), not yours. The real review is about running the game and playing it — and there it nets out to one true blocker (the Pendulum has no resolution procedure or numbers yet — confirmed by the Combat Codex itself), two math/balance items that need a playtest, and a few GM-facing rubrics that the digital tools will mostly carry for you. The aesthetic, the world, the economy, the character system, the secret — all solid.
Harvey worked all 11 decision cards. The headline: the P0 Pendulum blocker is resolved — he delivered a complete, end-to-end resolution flow (a five-position momentum track at ±10, commitment as the reward/risk dial, and a cost model where the emotional band is the Essence→Vitality exchange rate rather than a separate multiplier). That single answer also closes the Essence-cost question (it's the formula inside the flow) and most of the success-curve one. What now moves from "blocker" to "build it": prototype the Pendulum as a playable scene. The only piece he kept open on purpose is the full damage spectrum, because it's entangled with armour — he wants to work that out with us in a session rather than on paper. Everything else is decided; details fold into the sections below.
Several things that look like gaps in other systems are deliberate, working choices here — confirmed on the call. Calling them out so they're protected, not flattened, during the build.
Talking to a shopkeeper and fighting for your life run on the same loop (notice → understand → declare → commit → roll → consequence). Players learn one system, not five. Keep this sacred — the UI should never branch into a separate "combat mode."
No HP/mana/stamina split — effort and damage spend the same life. Elegant and thematically total. Attrition ("never at full health") is intended, not a bug.
By design the player doesn't stare at a number; the GM narrates state at the turn of the tide. This is a feature — the interface must honour it (state as feeling, not a bar).
Hiding crit/unlucky at Volatile, and the Great Unspoken, are intentional and powerful. The build will carry the secrecy structurally so it can't leak.
A weapon defined by the person, useless to anyone else, GM-set ranges rather than fixed stats — intentional. We model items as GM-authored, not catalogue loot.
Natural and a little fantastical — a weapon can live on the wrist, the back, wherever the player builds it. To keep it grounded, cap the number of fast slots (easy-access positions the player describes); anything in a fast slot swaps at no time cost in combat. Swapping otherwise costs a simple one-position loss on the Pendulum. Build action: render a freeform layout with a fast-slot limit; wire the off-slot swap to a −1 Pendulum step.
Bands (Steady/Heightened/Volatile), intensity (Light/Standard/Full), Vitality states (Functional/Compromised/Broken), bond tiers. The triune rhythm is a real structural signature — we'll mirror it in the UI.
You asked whether the 50 flags from breaking your codexes into structured notes are real design gaps or documentation/linkage artifacts. ~90% are ours to absorb.
| Flag family | What it really is | Owner |
|---|---|---|
| Missing note "types" (flora, item, lore, condition) | Our schema was built from only the two guides, before it met mega-flora, trade goods, fiction, and the Depleted. The world is richer than the starter schema. We widen the schema; no content is missing. | Build side |
| "Attribute deltas missing" on the peoples | A routing artifact — the race codices are lore; the numeric deltas live in the guides (which you have, and confirmed on the call). We reconcile across documents. | Build side |
| No range / unit fields ("2–8 shards", "60–120 Essence") | Our value shape held single numbers. Vireth runs on ranges. We add a {min, max, unit} shape. | Build side |
| Aliases & forward-links (velis vs free-city-of-velis; 47 cross-links) | Different codexes named the same thing two ways, and concepts reference each other across domains. Normal for a 43-document corpus; we canonicalise. | Build side |
| Intelligence Category, per-haven legality, characteristic bonuses | Your data is finer-grained than our fields — a good problem. We add the fields. | Build side |
The genuinely creator-facing items are tiny, and the call suggests you already hold the answers ("everything is really knitted together… I know why you can't be a vampire, why travel is affected…"). Two worth a 30-second confirmation:
Two codex passes name the Scars differently (e.g. crown-of-glass / drowned-spiral / quiet-basin / ashen-lift / mirror-march vs glass-wound / sighing-basin / ember-vein / pale-orchard / hollow-crown). Are these the same wounds under different names (aliases) or distinct places? Only you know.
Not aliases. They're separately listed and described as distinct permanent Scars. Canonical authority: the World Ecology codex is the bible for all such conflicts — it's the most up-to-date document and the final resolution of the earlier canon notes. Build action: model all seven as distinct geography nodes; when source docs disagree, World Ecology wins.
It appears as a Field Province, a permanent Scar, and a false-lane corridor. Intended (it's all three) or should one be primary? Affects how it's modelled on the map.
By its nature it genuinely fulfils all three roles at once. Build action: model it as a province and a Scar and a corridor-behaviour; no single classification is primary.
This is the single most important thing to pin down, and it's the one place the design genuinely stops short of itself.
Because every conflict in Fieldwake routes through the resolution loop, and combat is the loop's most mechanical instance, the Pendulum is the heart the digital engine has to beat. Right now it's described as a feeling, not a procedure. To build the scene/combat layer we need answers to:
It's the one piece we cannot fake or defer: the engine has to compute exchange outcomes deterministically and the GM/player interfaces render from them. Good news — you've already built the philosophy and the structure (2–4 exchanges, press/manage, subdue-vs-kill, commitment widens the crit margin). What's needed is a single page that turns that into numbers. Recommendation: a one-page Pendulum spec (even rough), which we immediately prototype as a playable scene so you can feel the numbers and iterate — rather than balancing it in the abstract.
Harvey answered this as a complete, single-roll resolution system, not just a choice of model. Three things work together and never overlap: the emotional band sets cost, the commitment level sets reward and risk, and success vs. the opposition moves the Pendulum. Calibrated so a starting character (~70–80 in core attributes like Body / Composure) can weather three big scenarios before a proper rest, but be badly affected. The flow, in his words:
Step one — the GM sets the Pendulum, a five-position advantage scale: • Greatly disadvantaged −10 · Disadvantaged −5 · Neutral 0 · Advantaged +5 · Greatly advantaged +10. The modifier applies to whoever the advantage favours, and it moves as the scene plays out. Step two — the player picks an action and a commitment level (the real decision; sets reward and risk): • Controlled — no bonus; standard effect; a success vs. a failing opponent swings the Pendulum 1 position. • Committed — +5 to the roll; on a failure, −5 to the next roll; effect ×1.5; clean success swings 2 positions. • Desperate — +10 to the roll; on a failure, −10 to the next roll; effect ×2; clean success swings 3 positions. (A controlled failure is a stumble to adapt around; a desperate failure is the cost landing AND the Pendulum turning against you at once.) Step three — the band sets the cost. Field/Glyphcraft/Vire/force-beyond-natural costs Essence; purely mundane or social actions take the bonus and swing with no Essence cost. Essence by commitment, with a die of variance: • Controlled 5 + d10 · Committed 15 + d10 · Desperate 30 + d10. That Essence is drawn from the pool, and spending it costs Vitality at the band's exchange rate — the band's ONLY job, no separate multiplier: • Steady 12 Essence : 1 Vitality · Heightened 8 : 1 · Volatile 4 : 1. (Volatile is already 3× as expensive as Steady because the same Essence burns 3× the Vitality. And because the Essence pool is derived from Vitality, every point lost — to spend or to wounds — shrinks the pool. The system tightens its own grip as a fight wears on.) Step four — the crit. Always swings the Pendulum one further position AND waives the Essence cost entirely. The crit margin widens with the band: base at Steady, +1 at Heightened, +2 at Volatile — the Volatile gamble is the most expensive place to act and the most likely to hit the cost-waiving crit. One pass: the player makes ONE roll for the narrated action. Apply Pendulum position + commitment bonus to the Characteristic's target; compare. On a success beating a failing opponent, move the Pendulum by the commitment's swing. If Essence was used, (commitment cost + d10) ÷ band rate = Vitality lost — unless it was a crit, which costs nothing and swings one further. Then narrate what it cost and what it changed.
What this unblocks: momentum is the ±10 five-position track; an exchange is one roll against a Pendulum-and-commitment-modified target; "five stages" = set / choose / cost / crit / resolve; a Transfer/decisive moment is a Pendulum-position swing. Build action: this is directly encodable — we prototype it as a playable scene immediately so Harvey can feel the numbers and iterate. Open thread: the exact damage values a hit deals (see MATH-1) are deliberately left for a working session because they're entangled with armour.
GapThe thresholds exist (Compromised < 50%, Broken < 25%) but how much Vitality a hit removes is unstated. A deterministic engine can't render "how you're doing" without a damage model.
Why it mattersPairs with the Pendulum: an exchange has to output a Vitality cost. Without it there's no attrition curve, no "two-to-four exchanges," no Broken moment.
RecommendationEven a simple model: cost = (outcome margin + weapon weight + band modifier), scaled so a clean hit is meaningful and a Volatile/Full exchange is brutal. We already do exactly this shape in the existing engine (Feng Shui's "smackdown"); we just need your intended feel for lethality (you said 2–4 exchanges — that implies fairly heavy hits relative to a ~50–60 pool).
Between two equals, 2–3 solid hits end it (with the option to subdue or finish). Vireth is dangerous by design. When matched-ness breaks, the curve stretches — 4–6+ — and a man vs. an elephant needs a crit or the right weapon. Outcome is shaped by surprise, preparedness (offensive/defensive stance), and relevant knowledge/learning, and the model must handle multiple attackers (packs, crews) — on Vireth a prolonged fight raises emotion, makes noise, and draws worse trouble (stronger animals, sudden weather). A core principle: you take damage whether you attack or defend — that's the cost of any fight; this is why there are no turns, only the Pendulum (an advantaged actor may even choose to wait and defend).
Deliberately open: Harvey hasn't cleanly worked the full damage spectrum yet because it's bound up with how armour works — he wants this resolved in a conversation, not on paper. Build action: book a damage-and-armour working session; everything upstream of it (the Pendulum flow) is ready to prototype now with placeholder damage values.
GapRoll-under, target = Characteristic ÷ 2 + Experience (+5/+10/+15) + GM bonus (1–7). Your own worked example: Vitality 55 → target 27; a Familiar-trained character is at ~32%. So a competent character succeeds at a trained task roughly 1 in 3 at Steady.
Why it mattersThat can feel grim in the wrong way — players failing things they're "good at." It interacts with the single pool (you spend Vitality even on misses at Full intensity) and could discourage the very Essence-spending that makes the game sing.
RecommendationA 30-minute math pass: decide intended hit-rates (e.g. trained-at-Steady ≈ 55–65%?) and back-solve the divisor / Experience bonuses. This is pure foundation, independent of the fiction — and cheap to get right now, expensive to retune after content is built on it.
Key reframe from Harvey: a competent person doing a known task under normal conditions simply succeeds — no roll. The GM just narrates the success. Dice only come out when something is abnormal: untrained, under time pressure, or under heightened stress. So the "1-in-3" never bites on the things a character is actually good at.
When a roll is called, the worked target lands near 55% for a starting character: ~35% base, +5 for first-level Experience at Steady, +10 for full commitment, plus the GM's situational −7…+7 for approach/narration. He explicitly does not want the world made easy for the inexperienced, nor a character ballooning to ~85% on hard tasks after 5–6 sessions of inspiration-pool growth. Build action: back-solve the divisor/bonuses to this curve; surface a clear "routine → auto-succeed / abnormal → roll" gate in the GM tool. Note this is also one roll inside the Pendulum flow (commitment is the same +5/+10 dial), so MATH-2 and the Pendulum share a calibration.
You said you don't want it "too onerous." These are the places where, as written, the GM is carrying a lot in their head — and where a rubric (or our tooling) removes the load.
Tension"You don't know the cost until you act" is great player-side immersion (confirmed on the call). But someone has to set that cost consistently — or the same act costs 6 Essence on Tuesday and 14 next week, and two GMs run wildly different games.
FixA GM-only cost rubric: default Essence by intensity × action-type × circumstance, hidden from players. Preserves the mystery; guarantees consistency; lets the digital GM tool suggest a cost you can override. This keeps the magic without the math landing on you mid-scene.
Harvey chose a repeatable formula over a hand-tuned table, and pointed straight at his Pendulum answer: Essence cost = commitment base + d10 (Controlled 5 / Committed 15 / Desperate 30, each + d10), then ÷ the band's exchange rate (Steady 12:1 · Heightened 8:1 · Volatile 4:1) to get Vitality. Build action: the GM tool computes and proposes this every time; the variance die keeps it from feeling like a fixed price, and a crit waives it. Consistency without a lookup table — and the player still doesn't see it coming.
Control rolls are "triggered by the fiction" — but with no cadence guidance, bands either never move (too few) or play stutters (too many). Needs a short trigger list ("call a Control roll when: witnessing harm to someone you care for; entering grief-saturated ground; …") so the emotional engine actually turns over.
Primarily GM-driven, fired by the narrative moment — post-combat, bad or good news, sudden danger, and the obvious ones from our list. But Harvey wants it light on the GM: a GM may change triggers or skip them when it suits the scene — "nothing too rigid; the world is punishing enough." Build action: ship the trigger list as an editable suggestion/prompt, never an enforced rule.
One GM is asked to secretly roll Volatile crit/unlucky, grow each player's hidden Adventure Experience + Journey Insight, run Metin betrayal tests, and track Bondmate truths. That's heavy for a person — but it's exactly what software is good at (see §07).
"GM watches and assembles a third Experience" is lovely, but: what if it never coheres? When is it "done"? A light rule (e.g. reveal at a dramatic beat once it has 3 Learnings + a Knowledge) keeps it from becoming open-ended GM homework.
Harvey doesn't want it held off forever: as soon as the package is complete it's revealed to the player — initially 3 Learnings + a Knowledge. Then two further Learnings can be added by the GM over time, granted when the player levels the package up, exactly like their other Experience packages. Build action: auto-surface the Adventure Experience at completion (3+1), then treat the last two Learnings as level-up rewards.
Windows, Anchors, corridor cost "by difficulty + field-state + party band" — rich, but no procedure. How many rolls is a crossing? Who rolls? This is a second mini-Pendulum and deserves the same one-page treatment (after combat).
Parked until combat is locked, by design. The intent (see the Tender race doc + player codex): travel is not teleport-with-a-price. You find a team to take you, a workable route must exist, it takes time, and the window can close or the corridor can shift (what the Tender reads); you still need a skiff or beast — like flying, but the currents are the Field and they move fast. A leg can be a full adventure (raided by beasts or outlaws on a cargo skiff) or a one-line narration — GM's call. The one fixed rule: a crossing always costs Essence, and Vitality if the Essence is high enough to trigger it. Open: the cost formula — Harvey wants it driven by distance × weather, reduced by the Tender team's skill (experience/learning). Build action: defer the loop; reserve the cost = f(distance, weather, ÷ tender-skill) shape.
No initiative is the right call — but a player can feel passive when "the GM decides who acts." The cure is interface, not rules: the screen must clearly cue "this is your moment — press or manage?" vs "you're defending." Flagging it so the build treats it as a first-class UX requirement, not an afterthought.
The game asks players to choose intensity with weight — but if the cost is hidden until after, the gamble isn't fully informed. That's fine if the player has a felt sense of risk. Worth watching in playtest: do players hoard Essence (never spend, no fun) or over-spend into death spirals? Tune recovery (1/hr, Steady-only) against drain accordingly.
Harvey ran the numbers and likes the balance: at Vitality 80 / Steady a character holds ~105 Essence, so burning the whole pool costs only ~10 Vitality — but pushing 100 Essence at Volatile costs ~25 Vitality before any wound, "and that feels good to me." Keep the punishing rate for now, revisit after playtest. (Note: this confirms the band exchange-rate model from the Pendulum answer is the live cost system, not a separate dial.)
Elegant distinction, but players will fumble "which applies," and "the GM volunteers info from your Knowledge" asks the GM to remember every player's Knowledge list. The app can surface relevant Knowledge automatically when a scene matches — turning a memory burden into a prompt.
The dreamstorm-amnesia opening (from the call) is a great teaching device. Pair it with progressive UI disclosure: start with Vitality + band only; reveal Essence costs, Insight pools, and crafting as the fiction reintroduces them. Don't show a full sheet on session one.
A lot of the "GM load" above isn't a design flaw to fix — it's work the platform is built to carry. You worried about it being "too onerous"; here's what the software quietly handles so you can stay in narration.
| The load | Carried by the platform |
|---|---|
| Tracking each character's band, Vitality state, Essence pool | Live on every sheet; the GM sees all, the player sees their own. |
| Secret Volatile crit/unlucky rolls | Rolled server-side, shown to the GM, hidden from the player — structurally impossible to leak. |
| The Metin Great Unspoken & betrayal tests; Bondmate true stats | GM-only secret store; the owning player is redacted even from their own companion's numbers (descriptors only). |
| Journey Insight + the hidden Adventure Experience | A GM-only pool that accrues from what the character actually does (the event log is the evidence). |
| Suggesting a consistent Essence/Vitality cost | The GM cost rubric, encoded — the tool proposes, the GM overrides. |
| The ever-changing world (residue, Field-state, closing Windows) | The dive-map (globe → region → scene) with the world reacting to what players feel and do. |
Net: the tabletop version asks one person to hold a great deal; the digital version hands most of it to the machine, leaving the GM to do the part only a human can — read the table and respond.
Status as of 15 Jun 2026. Of the eight lines, six are now decided, one stays open by choice (damage spectrum, tied to armour), and one was always ours.
| Priority | Item | Status | What's next |
|---|---|---|---|
| P0 | Pendulum resolution spec — momentum state, exchange resolution, the five stages, Transfer triggers | decided | Full flow delivered (§03). Prototype it as a playable scene. |
| P0 | Damage → Vitality model — what a blow costs | feel set · values open | "Heavy" confirmed; exact spectrum waits on the armour session. Prototype with placeholder values. |
| P1 | Validate the success curve — confirm hit-rates, back-solve divisor/bonuses | decided | ~55% trained-committed; routine tasks auto-succeed. Back-solve to that curve. |
| P1 | Essence cost rubric + band-shift triggers | decided | Cost = commitment+d10 ÷ band rate (§05). Triggers = editable GM prompts, light touch. |
| P1 | Attrition tuning — drain vs 1/hr Steady-only recovery | decided | Keep as-is; revisit after playtest trials. |
| P2 | Travel/crossing loop; Adventure-Experience guardrails; weapon equip slots | decided | Travel drafted later (cost = distance×weather ÷ tender-skill); Adv-Exp reveals at completion; freeform slots + fast-slot cap. |
| P2 | Canon confirms — seven-Scars naming; Mirror March classification | decided | Scars are distinct (World Ecology codex is the bible); Mirror March is all three. |
| ours | Schema/linkage flags (types, ranges, aliases, fields) | build side | The compile pipeline — not the creator's concern. |
MVP order: character builder first (the race → 6–7 questions → name → spend 150 points flow, choices irrevocable, character developing live on screen) → minimalist player interface (six characteristics · Essence pool · emotional band · body/equipment outline · companion + link strength) → the dive-map. Aesthetic: dark fantasy à la Abercrombie / Name of the Wind / Game of Thrones, with Voynich-manuscript-strange flora & fauna — emphatically not D&D elves-and-dwarves, and not sci-fi. Concept directions to follow.