The Three Walls
A deep dive into Shroud of the Avatar's resistance systems: magic wards, armor, and above all the hidden rules of stun resistance, measured against thirty years of the genre, with thirty concrete proposals.
Ask three players what "resistance" means in New Britannia and you will get three answers, and all three will be wrong in different ways. This page walks the three separate systems that actually decide what a fireball, a war maul and a stun each do to you, traces how the genre answered the same questions from Ultima Online onward, and lays out thirty concrete proposals. The center of gravity is crowd control: the stun resistance you cannot see, the mesmerize resistance that does not exist, and the diminishing-returns ladder the game has been running silently for years. Every card is open for comments and votes below.
Every long-lived RPG has asked what a resistance is, and nearly all of them changed their answer. Ultima Online began with a resist skill so mandatory it taxed one seventh of every build, and ended in 2003 with five capped percentages on the paperdoll. EverQuest kept the coin flip and spent two decades being reverse-engineered by its own players. Dark Age of Camelot invented the immunity timer. World of Warcraft invented the diminishing-returns category, then deleted its resistance stat outright in 2010 because nobody could say what it did. Diablo 3 wrote down the rule the whole genre now follows: control should always land, and never lock. Elder Scrolls Online made escaping a stun something you pay for in stamina rather than wait for on a cooldown.
Shroud runs three resistance systems, and they do not know each other exist. Spells are answered by an attunement contest that armor never touches. Physical hits are answered by armor math that magic never touches. Crowd control is rolled against a third store of numbers with its own hidden diminishing-returns ladder: every landed stun, knockdown, knockback, root or mesmerize quietly grants the victim twenty percent blanket control resistance for thirty seconds, five such hits reach total immunity, and neither the attacker nor the defender is ever shown any of it. Meanwhile the character sheet prints thirteen resistance rows and hides everything above; a resisted stun produces no text, no floaty and no log line; mesmerize resistance is a stat almost nothing in the game writes, so a wisp's mesmerize lands one hundred percent of the time forever; and thirteen boss templates author their crowd-control immunity into a stat that no code has ever read, so it does nothing at all.
Three gap clusters organize everything that follows:
Thirty proposals follow, in four tiers.
| Tier | Theme | Cards | Cost band | Who asked |
|---|---|---|---|---|
| 1 · The Honest Ledger | The defect wave and the missing chapter: mesmerize resistance made real, three break-free bugs, the armor gate that turns mitigation off, the dead boss stat wired, dead surfaces retired, and the documentation the system never had | S1–S12 | all small | everyone who ever wondered why the wisp's mesmerize always lands |
| 2 · The Answered Roll | Legibility and core upgrades: a Defenses ledger on the character sheet, a floaty and log line for every resisted roll, attacker-visible immunity, the true magic-mitigation math published, control-resist gear beyond one helm, caps with a purpose, break-free you can compute | S13–S20 | mostly medium | Wishdragon's wishlist ask, and nine sibling pages' worth of parked questions |
| 3 · The Governor | The full crowd-control specification: per-type diminishing returns replacing the blanket stack, a hard cap and floor, escape and immunity as one package, one stacking rule, a declared boss posture | S21–S26 | large, flagged | five waiting cards on four sibling pages |
| 4 · The Owner's Desk | Rulings: the stun diminishing-returns shape, PvP scope, the creature values changeset, and deterministic magic mitigation go or no-go | S27–S30 | rulings first | the Bludgeon page's reserved owner calls, now with a written spec to rule on |
Open the character sheet of a level 80 defender in plate with a shield, trained in Hard Headed and Stand Your Ground, wearing a stun-resist masterworked helm. This is everything the game will show about resistance:
Resist: 41.6 (4) ... RESISTANCES (1) Fire Resistance 62.0 (3) Air Resistance 50.0 Death Resistance 20.0 Magic Resistance 15.0 Bludgeon Resistance 0.1 rows at zero are not drawn; a new character sees no Resistances section at all (2)
Three separate pipelines decide what an incoming attack does, and no rule is shared between them. A spell is rolled against dodge, glancing, block and parry, where your school ward and magic resistance buy percentage points of partial-resist chance, and a partial resist deals one third damage. The number that looks like direct spell mitigation in the code, the attunement multiplier, is computed after the damage is already final and used only to write a diagnostic log line. A physical hit runs the armor math: per-weapon-type resistance, the glancing roll, block and parry, then flat damage resistance capped by absorption. A crowd-control effect ignores both of those walls entirely and rolls its chance against a third store, the resistance array, where the hidden ladder lives.
flowchart TB
IN[Incoming effect] --> W{Which wall}
W -->|Spell| S1x[Hit roll: dodge, glancing, crit
ward points raise partial-resist chance
a partial resist deals 1/3]
S1x --> S2x[Block and parry apply to spells]
S2x --> S3x[Attunement multiplier computed
then written to a log and discarded]
S3x --> OUT[Damage applied]
W -->|Physical| P1[Weapon-type resistance divides base]
P1 --> P2[Hit roll: dodge, glancing, crit]
P2 --> P3[Block and parry]
P3 --> P4[Flat damage resistance,
capped by absorption share]
P4 --> OUT
W -->|Crowd control| C1[Chance expression evaluated]
C1 --> C2[Resistance array consulted:
stun, knock, root multiply chance down
mesmerize alone subtracts flat]
C2 --> C3[Roll. A failed roll shows nothing]
C3 --> C4[If it lands: victim silently gains
+20% blanket control resistance, 30s]
OUT --> R[Matching resistance percentages
trim the final damage per type]
style S3x stroke-width:3px
style C3 stroke-width:3px
Crowd control already has diminishing returns. Every landed stun, knockdown, knockback, root, mesmerize or confusion grants the victim a hidden buff worth twenty percent blanket crowd-control resistance for thirty seconds, stacking up to ten times. Five stacks reach one hundred percent: total immunity to every hard control type, earned by being controlled five times in half a minute, and decaying thirty seconds after the last hit. Breaking free with the jump key grants a stack too. None of it has ever been shown in the interface on either side of the fight. The Air Magic page published that this ladder exists; this page's rule, inherited from that one, is that every proposal here surfaces or replaces the ladder and never pretends to invent it.
| Ladder fact | Today's value | The catch |
|---|---|---|
| Gain per landed hard control | +20% for 30 s | blanket: a knockdown builds stun immunity |
| Stack limit | 10 | immunity arrives at 5; stacks 6 through 10 do nothing |
| Visible to the defender | no | no buff icon, no sheet row |
| Visible to the attacker | no | glyphs spent into immunity are simply eaten |
| Mesmerize reads it back | no | mez grants stacks and ignores them when landing |
| Slows and snares | outside the ladder | correct, and defended as a rule by card S19 |
Shroud's ancestor asked the question first and answered it twice. In 1997 Ultima Online shipped Magic Resistance as a skill: chance-based, trained by standing in fire fields, and so valuable against the game's dominant threat that it became a de facto mandatory hundred points in a seven-hundred-point classless build. One seventh of every character, spent on a coin flip. The 2003 Age of Shadows publish tore it out and replaced it with five intrinsic resistances, Physical, Fire, Cold, Poison and Energy, each a flat percentage capped at seventy, fed by armor, jewelry and spells. The old skill was rescoped to what resistance skills are actually good at: shortening curses and paralysis, and setting a minimum resist floor that does not stack with gear but stands in for it. Both halves of that story matter here. A resist skill in a classless game becomes a tax. A resist stat on items becomes a suit-building economy, which is content if you have crafters, and Shroud does. And the floor-versus-gear rule, take the higher instead of the sum, remains the cheapest anti-stacking device ever shipped.
EverQuest kept the coin flip and became the genre's cautionary tale: five resist scores, a hidden roll, full resists on twenty-minute raid attempts, and a spell meta that answered resist inflation with spells that ignore resistance. Its community reverse-engineered the formula for two decades and still argues about it. Dark Age of Camelot built the other pillar in 2001: the immunity timer. Break a mesmerize and you cannot be mesmerized again for a minute; a stun buys immunity worth five times its duration. Chain control stopped being a rotation and became a spend. The design had a famous flaw, immunity granted by application rather than time served, so a careless area spell gifted enemies their freedom, but the core idea, control as a finite resource, is now everywhere. World of Warcraft systematized it as diminishing returns: within a category, the second application runs half duration, the third a quarter, the fourth is immune, and the ledger clears seconds later. Then in 2010 WoW deleted its resistance gear outright. The developers' stated reason has aged into doctrine: a stat that is only meaningful in narrow situations, farmed as a second gear set for one boss, generates no decisions and much resentment.
The modern entries refine the control half. Diablo 3's 2012 developer journal is the strongest single citation in this genre: monsters gain control resistance in proportion to time controlled and shed it while free, capped below immunity, because a hard immune state "felt wrong" and made the order in which teammates pressed their stuns decide whose skills did anything. Elder Scrolls Online prices the escape: breaking free costs a flat bite of stamina, grants about six seconds of visible immunity, white rings under the character that the attacker can see, and deliberately exempts snares and roots so control builds still function against an immune target. League of Legends distilled the stat itself: tenacity cuts control duration, is capped at one hundred percent with a duration floor, multiplies within a source group and adds across groups so no single source can be spammed to immunity, and carries an explicit exclusion list so the biggest effects always matter. Guild Wars 2 turned boss control into a resource, the defiance bar that crowd control damages until it breaks into a vulnerability window, after its first system, a boss immunity that reapplied the moment you broke it, became the genre's named example of control skills that exist and can never be used. Final Fantasy XIV runs the quiet version, each stun on a monster halving the next until a brief immunity, with one deliberate refinement: a stun landing on an already stunned target neither refreshes the effect nor advances the ramp, so uncoordinated teammates cannot waste each other's control.
Rows are things a player might reasonably expect a resistance system to do. A filled dot is a first-class mechanic, a half dot partial or conditional, an empty dot absent. The two Shroud columns are the point of the table.
| Capability | UO | EQ | DAoC | WoW | ESO | GW2 | XIV | LoL | D3 | SotA today | SotA proposed |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Deterministic mitigation, not a resist coin flip | ● | ○ | ● | ◔ | ● | ● | ● | ● | ● | ◔ chance of 1/3 | ● S25 |
| Hard cap on all mitigation | ● 70 | ○ | ● | ● | ● 50% | ● | ● | ● | ● | ◔ some clamps | ● S18 |
| Control diminishing returns, per type | ◔ | ○ | ◔ | ● | ◔ | ◔ | ● | ● | ● | ◔ blanket, hidden | ● S22 |
| Post-control immunity window | ○ | ○ | ● | ● | ● | ◔ | ● | ○ | ○ | ○ | ● S23 |
| Escape and immunity as one package | ○ | ○ | ◔ | ● | ● | ◔ | ● | ◔ | ○ | ◔ escape, no window | ● S23 |
| Escape priced in a resource | ○ | ○ | ○ | ○ | ● | ◔ | ○ | ○ | ○ | ● focus jump | ● kept |
| Soft control exempt from immunity | ○ | ○ | ◔ | ◔ | ● | ● | ◔ | ◔ | ○ | ● accidental | ● S19, law |
| Attacker-visible immunity telegraph | ○ | ◔ | ◔ | ◔ | ● rings | ● | ◔ | ○ | ○ | ○ | ● S15 |
| Delta feedback on a shortened or resisted control | ○ | ○ | ○ | ◔ | ○ | ○ | ○ | ◔ | ○ | ○ silent | ● S14 |
| Boss control beyond blanket immunity | ○ | ○ | ○ | ○ | ◔ | ● defiance | ◔ | ○ | ● | ○ masks or nothing | ◔ S26, ruled |
| Published numbers in the game itself | ◔ | ○ | ◔ | ◔ | ◔ | ○ | ◔ | ● | ◔ | ○ | ● Tiers 1–2 |
| One stacking rule across all sources | ● max() | ○ | ● | ◔ | ◔ | ● | ● | ● groups | ● | ○ additive accident | ● S24 |
Shroud holds two filled dots on that board already, and they are good ones: the break-free attempt is priced in focus, the resource-cost design ESO is praised for, and slows sit outside the immunity ladder, the exemption ESO and Albion both had to learn deliberately. The proposed column fills the rest without inventing a new resource system: nearly every row lands on plumbing the game already runs, and the biggest three land on data files the game has carried, authored and unread, for years.
| Anti-pattern | Shipped example | Where Shroud stands today |
|---|---|---|
| Chance-to-resist on decisive effects | EverQuest's full-resist raid roll | every control roll; kept in Tier 3's chance-capped form, tamed by caps and floors |
| Uncapped additive resist stacking | the failure every cap exists to prevent | live: five ladder stacks equal total immunity, and stacked stun sources can pass one hundred percent |
| Resist gear as an encounter key | WoW Classic's fire-resist wardrobe, deleted in 2010 | avoided; card S17 keeps gear broad, capped and optional |
| Blanket boss immunity | Guild Wars 2's pre-2015 defiance, immunity that reapplies on break | live for some bosses via full resist masks; card S26 declares the posture |
| Immunity granted by application, not time served | DAoC's five-times-duration stun immunity even when purged instantly | avoided so far; Tier 3 adopts the overlap-suppression rule that prevents it |
| Interrupt lock with no grace window | New World's stagger, its most complained-about mechanic | live in principle: no immunity window exists after any control ends |
| Offense and defense from one investment | attunement's double dip, the mono-school pull | live; named here, softened by S16 and S25's defensive accounting, not redesigned |
| Hidden matrices | every game whose community had to datamine its defense math | live everywhere this page touches; ended by rule: every number here is published |
Numbered S1 through S30, in four tiers: the honest ledger first, then the answered roll, then the governor, then the owner's desk. Every card names its cost, its risk, and where it plugs into the shipping data. Tuning numbers are illustrations for discussion, not commitments; the balance program owns every final value, and every number that ships gets published in the game.
the team's combat reference documents damage resistance in full and mentions crowd-control resistance zero times
The internal combat documentation explains armor, absorption and elemental resistance carefully, and stops there. The crowd-control half, the landing roll, the six places the hidden ladder is granted, the break-free math, the fact that one store of resistances is clamped and the other is not, exists only in code. Write the missing chapter, correct the combat log line that credits partial resists on the armor path to "armor" when the roll is avoidance-driven, and add the resistance band to the balance targets file, a follow-up the Death Magic page committed to and nobody wrote. A system this load-bearing should not require an archaeologist.
Seen in: World of Warcraft, negatively. Its control rules went undocumented for twenty years and the entire competitive scene runs third-party tools to see them.
One new documentation chapter covering the control-resistance branch table, the ladder's six grant sites, the struggle math and both resistance stores; amendments to the status-effect and formula references where they are silent or wrong; a resistance and control band added to the balance targets file. Ships with this page.
mesmerize resistance: one writer in the entire game, and a flat subtraction where every other control is proportional
Stun, root, knockdown, knockback and confusion resistance all scale the landing chance down proportionally. Mesmerize alone subtracts its resistance flat from the roll, the exact shape that was already fixed for confusion after self-stacking pushed it to accidental full immunity. Worse, the stat behind it is written by almost nothing: one dragon template carries a value, and no skill, no gear, no buff. And mesmerize landings feed the hidden ladder without ever reading it back. The result is that a wisp's mesmerize lands one hundred percent of the time, forever, on everyone. Give mesmerize the same proportional shape and the same ladder term as stun, and rebase the dragon's value in the same changeset so its defense stays where it was authored.
Seen in: this game. The confusion repair earlier this year is the precedent, applied to the one control it missed.
One branch of the landing roll changed from flat subtraction to the proportional form, plus the blanket ladder term the other controls already read. The remote-player deferral list is updated in the same commit, and a new editor test pins the shape, the ladder term and the stat's zero-to-one scale. The dragon value rebase ships in card S29's reviewed changeset.
a divide by zero, a discount that evaporates, and an unbreakable root that breaks
Jamming the jump key while controlled is the game's struggle mechanic, and it has three verified bugs. A mesmerized player with no stun magnitude enters the stun-break path and divides by zero, which makes the escape chance infinite: mesmerize is instantly breakable by accident while the deliberate damage-break rule already covers it. Failed attempts are supposed to shorten the remaining effect by a fifth, but the discount is written to a value the next stat recalculation rebuilds, so it silently evaporates. And a successful break removes even roots explicitly defined as unbreakable, the one thing the guard at the top of the function exists to protect. Three guards, no design change.
Seen in: Elder Scrolls Online's Break Free, the reference for a struggle system that works exactly as printed.
Guard the zero-magnitude case and route mesmerize out of the stun-break branch; apply the failure discount to the controlling effect itself so it survives recalculation; exclude unbreakable roots from the success cleanup. An editor test pins each: no infinite chance at zero magnitude, the discount survives a recalculation, unbreakable roots survive a successful break.
any character with a nonzero focus-transfer effect takes zero armor mitigation on every physical hit
Deep in the physical damage path, the final subtraction of armor damage resistance is gated on the target having exactly zero focus transfer. Anyone running a focus-transfer effect, and several buffs grant one, silently loses all armor mitigation on every hit for the duration, with no tooltip, no documentation and no comment in the data explaining why. If the gate is intentional, redirecting damage before armor is a rule players deserve to read; if it is not, it is a one-line repair. Investigate, then either restore mitigation or publish the rule. Nothing about the current silent state is defensible.
Seen in: nowhere. No game in the survey ships an undocumented interaction that disables a character's entire armor.
First establish original intent from the data and history, then either remove the gate or document and tooltip it. An editor test pins the chosen behavior at zero and nonzero transfer values so the interaction can never go silent again.
a debuff that zeroes the scaling threshold makes the next big hit compute infinite resistance
Above a threshold, damage resistance scales itself up to keep pace with very large hits, a quiet anti-burst mechanic the Heavy Armor page documented. The scaling divides by a stat that skills already lower, and nothing stops it reaching zero, where the next oversized hit computes infinite resistance. One guard, one test, and the inverted-sounding mechanic itself gets a plain-language paragraph in S1's chapter so players stop discovering it by rumor.
Seen in: plain defect hygiene. Every shipped system in the survey clamps its divisors.
Floor the divisor at a small positive value and pin it with an editor test covering zero and negative stat states.
bludgeon, puncture and cut resistance divide incoming damage; a value of minus one divides by zero
Resistance to the three physical weapon families works differently from every other resistance in the game: it divides the weapon's base damage by one plus the value. That makes negative values vulnerabilities, which is used well, skeletons take almost half again bludgeon damage by design, but the divisor is unclamped, and a value of minus one is a division by zero waiting for the first creature or debuff authored past the line. Floor the divisor, keep the skeleton's weakness exactly where it is, and publish the divisive shape in S1's chapter since no player would guess it.
Seen in: defect hygiene again, and the skeleton's bludgeon weakness as the pattern done right.
Floor the divisor at a published positive value, pin the floor and the skeleton vulnerability with an editor test, and note the three armor-scoped twin stats this family carries that no code has ever read, which card S11 retires.
the per-type damage trim iterates every type bit on a hit, control bits included
After a hit resolves, matching resistance percentages trim the damage per damage type. The loop walks every type flag the effect carries, including the control flags, so a rune tagged as a stun has its damage reduced by the victim's stacked control resistance. Against creatures whose templates carry seventy percent in those slots, control-tagged damage is quietly shredded, and rune authors have learned to dodge the loop by hand-picking flags. Restrict the trim to damage-bearing types with a published mask. The mechanism change is fair game; because it lands as a damage buff against those creatures, it ships credited to the combat balance program and on its schedule.
Seen in: no genre analogue. This is type-system leakage, not a design.
A published damage-bearing mask excluding the control and mechanic flags, applied at the trim loop. An editor test pins that control-tagged rune damage ignores control resistance while elemental trims still apply per type. Timing coordinated with the balance program's second wave since creature-facing damage moves.
a plate-and-shield build with two ladder stacks reaches 127 percent combined stun resistance
The stun landing roll subtracts stun resistance plus blanket control resistance, and while each store clamps itself, nothing clamps the sum. Hard Headed at high level, Stand Your Ground, a masterworked helm and two ladder stacks total well past one hundred percent: permanent, accidental stun immunity, invisible like everything else here. Clamp the combined resist factor at one hundred percent now, and let Tier 3's published cap tighten it properly later. The Shield page's Bolster Balance repair, which re-authors that buff at its observed effective values, lands first so its measurements are taken against today's behavior.
Seen in: League of Legends, whose tenacity is hard-capped at one hundred percent precisely so stacked sources cannot produce silent immunity. Not one shipped system in the survey is uncapped.
Clamp the summed resist factor in every control branch of the landing roll, leaving the roll's raw-value return contract intact for the diagnostic overlay. The remote-player deferral list is untouched by a pure clamp. An editor test pins the worked example: a 127 percent build yields zero landing chance, never negative.
thirteen boss and creature templates author crowd-control resistance into a stat no code has ever read
There is a stat named crowd-control resistance, it has a display name, and thirteen creature templates write serious values into it: mechanicals at one hundred, the spider queen at one hundred, Krampus, the envy bosses, five elemental bosses, dragons, daemons, cabalists. No code reads it. Every boss whose control immunity was authored this way has none, and three creature abilities whose landing math subtracts the target's value are subtracting a permanent zero. Wire the stat into the landing roll, fulfilling a to-do note that has sat in the code for years, but activate the authored values only through a reviewed changeset: at face value, one hundred would mean full immunity, and the boss posture card S26 says bosses resist hard and are never blanket-immune.
Seen in: Final Fantasy XIV's per-monster control resistance, the pattern of bosses that resist harder, visibly, without a wall.
Add the stat's clamped term to each control branch of the landing roll behind a flag, update the remote-player deferral list, and pin scale and branch coverage with an editor test. The thirteen templates' live values ship in card S29's reviewed changeset, capped per card S26's posture. The three creature abilities that already read the stat come alive at the same moment, which the changeset review prices in.
two resistance keywords in the skill formula language are stubs: one always returns 1, the other returns its input unchanged
The formula language rune designers write in exposes two resistance keywords, and both are placeholders that silently do nothing. No shipped rune uses them yet, which is exactly why now is the moment to decide. Wire the value keyword to return the target's clamped resistance for the rune's type, and it becomes the deterministic duration-scaling primitive this whole page wants: the thing a stun rune's duration can be multiplied by, the shape one air spell already hand-builds through the attunement contest. It would also be the formula language's first pre-clamped primitive, in a language that has no clamp operator at all. Or retire both keywords from the tokenizer. Either way, stop shipping vocabulary that lies to its authors.
Seen in: Ultima Online after 2003, where resistance acts on curse and paralyze duration, per effect, deterministically.
Wiring means defining the keyword's semantics once, forever: the recommendation is one minus the clamped matching resistance, suitable for duration expressions. Any chance expression that consults target resistance through these keywords must also join the remote-player deferral list, which is a documented rule in the same change. An editor test evaluates both keywords against a seeded resistance state.
three authored resist buffs no code ever casts, a dead stat trio, a dead tweakable, and a typo'd enum
The resistance data model carries a museum of unwired intent. Three fully authored per-type resist buffs, stun, knockdown and knockback, sit in the data and are never cast by anything; only the blanket version is, which is why the stun formula collapses into the root formula today. Those three are reserved: Tier 3's per-type ladder is built out of them. The armor-scoped twins of the weapon-type resistances have no writers and no readers; retire them from display. A struggle-chance tuning value has no reader; card S20 makes it the real knob. A misspelled resistance enum has zero uses; deprecate it. Each disposition is public: retired, or reserved for a named card.
Seen in: World of Warcraft's 2010 stat cleanup, the precedent for deliberately retiring stats that confuse more than they serve.
Display-name and documentation removals for the retired surfaces, deprecation notes on the dead enum, and a new editor test that pins every disposition so nothing quietly resurrects. Sibling claims respected: the armor-penalty and protection stats belong to Heavy Armor's cards, the shield's dead testudo stat to the Shield page, the dead knockdown masterwork stat to the Bludgeon page.
poison against the undead already floats "Immune"; a stun with zero landing chance floats nothing
The game already owns an immune floaty: poison a skeleton and it says so. But throw a stun at a target whose resistance has reached the cap and the glyph is simply eaten in silence. Reuse the existing floaty whenever a control effect's landing chance is mathematically zero, so the attacker learns, mid-fight, that this target is not currently stunnable. This is the cheapest possible slice of the feedback problem: one condition, one existing string, no new art.
Seen in: Diablo 3's developer journal, which names silent immunity as the thing that made teammates' skills invisibly cancel each other.
Emit the existing immune text when a control effect's computed chance is zero or below, distinct from a failed roll, which card S14 handles. A per-target throttle mirrors the existing zero-damage tick suppression so area control does not spam.
requested on the wishlist forum by Wishdragon: surface the hidden defensive attributes in the interface
Add a Defenses section to the character sheet that prints every real number: armor resistance and the absorption cap under an explicit "physical hits only" label, the ward table with both fire numbers in separate columns, and a Control block showing stun, knockdown, knockback, root, mesmerize and confusion resistance with base and temporary stacks listed separately. Rows never hide at zero, because a zero is information: mesmerize's honest zero is this page's exhibit A. Every line carries a hover tooltip printing the operands, which is the no-hidden-matrices rule made concrete. The full mockup is in section 4.
Seen in: Ultima Online's 2003 paperdoll, the moment resistance became a build system, and Elder Scrolls Online's mitigation lines.
A new persistent sheet section sourced from both resistance stores and the armor totals the sheet already computes, with the ladder's temporary stacks shown as a separate active column. An editor test pins that every rendered row maps to a value the combat code actually reads, using card S11's audit as the source of truth, so a dead stat can never appear on the ledger.
a resisted control today: no floaty, no log line, no sound, nothing on either screen
When your stun is resisted, nothing happens. The effect never appears, and neither player can tell a resist from a miss from a bug. Add the missing feedback: a floaty on a failed control roll, and a combat log line carrying the actual arithmetic, base chance, resistance applied, final chance, so both sides learn the system by playing it. The genre's core readability lesson is that the delta teaches and the state does not: "Stun resisted" plus the log's "chance 40 percent after 55 percent resistance" beats twenty years of silent rolls. The diagnostic overlay already records every roll; this promotes what it knows to the screens players actually watch.
Seen in: the survey's readability finding end to end. EverQuest's bare "you resist" taught nothing for twenty-five years; the games players understand print the numbers.
The roll site already raises an event for the diagnostic overlay; route the same event to the floaty and combat log systems with a per-target throttle for area control. Against remote players the victim's client owns the roll, so the outcome rides the existing echo back to the attacker for display, with no new network operation. Editor tests pin exactly one emission per failed roll and outcome carriage on the echo path.
the player who most needs to see a target's immunity is the one about to spend a glyph on it
While a target's control resistance sits at the cap, show it: a quiet ring under the character or a pip on the nameplate, visible to everyone, not just the defender. Attacker-visible immunity is the single best-proven telegraph in the survey, and its absence is a coordination tax on every group fight: glyphs spent into a capped target are simply wasted, and nobody learns why. Version one triggers on today's blanket cap; after Tier 3, the ring becomes category-scoped. Defender-side surfaces stay with the Air page's cards, which own the stun icon, duration readout and struggle prompt.
Seen in: Elder Scrolls Online's white break-free rings and New World's white weapon glow, the two clearest attacker-facing telegraphs shipped.
A world-space ring or nameplate pip driven by the replicated resistance state both clients already hold, rendering when the blanket value is at cap. Whether the ring shows in PvP is bundled into card S28's scope ruling; the recommendation is yes, for the same coordination reasons. An editor test pins the on and off thresholds.
what a ward point buys today: partial-resist chance toward a one-third-damage glancing hit, and nothing else
Publish what magic resistance actually is. The ward tooltips say resistance and let players assume mitigation; the truth is a chance contest, ward points against the caster's attunement, paying out a one-third-damage partial resist, plus a separate per-type percentage trim most players have never heard of. The combat log's "partially resisted due to magic resistance" line fires on a roll that avoidance also feeds, so even the message is not quite honest. Correct the ten ward tooltips to name the real currency, fix the log line, publish the attunement curve's shape in S1's chapter, note that chaos alone ignores general magic resistance as the Chaos page already established, and print the quiet asymmetry that the shield school's stun-resist passive outscales the heavy armor school's by a quarter with no stated reason. Expectation management is the whole card: the mechanism does not move.
Seen in: World of Warcraft's deletion rationale, in reverse. A resist stat nobody can explain eventually gets deleted; explaining it is cheaper.
Ten tooltip strings, one log string, and documentation. The discarded attunement multiplier in the spell path gets an explanatory annotation so the next reader does not mistake it for mitigation; making it real is card S25's separate, gated question. Editor tests pin tooltip text against the shipping constants so a retune forces both to move together.
exactly one masterwork in the game grants stun resistance, on heavy helms only, at five percent
A player who wants to buy control resistance has one option in the entire itemization system: a heavy helm masterwork. Nothing in gear touches knockdown, knockback, mesmerize or blanket control resistance. Add a small, published, capped family: a shield masterwork for knockdown resistance, a necklace ward for mesmerize once S2 makes that stat real, on the standard masterwork scaling with the standard durability price. The gear pool's total contribution per category is capped and printed in a table on this page, which keeps the lesson of the genre's deleted resist wardrobes: broad and optional, never an encounter key. Light Armor's page keeps its root-and-mez innate; this card is gear only.
Seen in: Albion Online, where armor weight carries innate control resistance and specific items add targeted, bounded relief; and in the crafting economy argument that made Ultima's capped suits content rather than debt.
Two or three new masterwork and enchant data entries on the existing item-effect scaling shape, plumbed through the same per-item resistance walk the stun helm already uses. Ships only after S8's clamp and S24's stacking rule, in that order, so the new sources land inside a governed system. Data pins cover masks and coefficients, and the S8 worked example extends to include the new gear at cap.
not one shipped resistance system in the survey is uncapped; Shroud is capped in some places and open in others
Declare the caps, in one place, published. The per-type damage trim already clamps at one hundred percent; the control ladder's cap becomes eighty percent under Tier 3; the gear pool's per-category cap comes from S17; the attunement curve is the school-resist soft cap and gets printed as such. Then give overcap a purpose instead of letting it die as waste: resistance strips exist and are growing, the Death page's ward rend, the air school's resistance strips, so points past the cap are insurance against being stripped, and the tooltip says exactly that. The card also tables the genre's cheapest anti-stacking device, the skill-floor-versus-gear maximum, as a presented option for the ward passives, for the owner to take or leave.
Seen in: Ultima Online's seventy cap that made suit building a bounded puzzle, and Elder Scrolls Online's overcap that buffers enemy penetration instead of evaporating.
A published cap table in S1's chapter and on the ledger tooltips, the few clamps not already carded, and overcap-as-strip-insurance wording. No numbers move outside changesets. The maximum-versus-sum option for ward passives is presented with worked examples but explicitly not decided here; it reshapes what a ward point is worth and belongs beside card S30's ruling.
slows and snares do not build the immunity ladder today, and that is correct, and it is an accident
The best rule in Shroud's control system is one nobody wrote on purpose: slows and snares sit outside the diminishing-returns ladder, because the movement stats simply do not appear in the switch that grants stacks. That accident is the exemption Elder Scrolls Online and Albion both had to learn deliberately, and it is the reason control-oriented builds stay playable against a target whose hard-control immunity is up. Write it down as a contract, defend it in Tier 3's specification, and pin it with a test so no future card can break it by adding a stat to a list. The Water page's governed freeze deliberately rides the root type to enter the ladder; that stays the pattern for any slow that wants governance.
Seen in: Elder Scrolls Online, whose Break Free neither clears snares nor grants them immunity, by design.
Documentation in S1's chapter and one editor test: no ladder stack is granted by movement-stat effects, and break-free does not clear them. Zero behavior change.
three different stat weightings, a chance that can exceed certainty, and a knockdown branch that ignores the knockdown
The struggle math is three separate formulas wearing one jump key. Stun breaks weigh intelligence, root breaks strength, knockdown breaks dexterity, which is defensible flavor; but the stun chance is governed by the effect's magnitude rather than its duration and is unclamped, so weak stuns are guaranteed instant breaks, while the knockdown branch ignores magnitude entirely and carries no failure discount. Unify the three into one published, clamped formula: a chance band with a floor and ceiling, the magnitude term retained but bounded, the failure discount applied durably after S3's repair, and the long-dead struggle tuning value wired as the real knob. The Air page's struggle prompt then has something honest to surface.
Seen in: Elder Scrolls Online's Break Free, whose printed cost and printed result are the whole reason players trust it.
One formula, three stat weightings preserved, chance clamped to a published band, per-attempt cadence and the focus price unchanged. The dead tuning value becomes the scale knob, reserved by card S11. PvE and PvP bands split only if card S28 rules so. Editor tests pin the band across a stat-by-magnitude grid and the knockdown branch's new magnitude term.
five cards on four sibling pages are parked waiting for this document to exist
The Bludgeon page reserved the stun diminishing-returns direction as an owner call and four other pages parked cards behind it. What none of them had was a written specification to rule on. This is it. Categories: stun; knockdown and knockback together; mesmerize; root. Slows and snares exempt by S19's law, unbreakable roots outside the system entirely and named on the exclusion list, so no stat ever trivializes a signature effect. Per-category stacks replace the blanket stack, capped at eighty percent so accumulation never reaches immunity. A control landing on an active identical control neither refreshes it nor advances the ladder, the overlap rule that stops teammates wasting each other. Escape and immunity ship as one package. One stacking rule governs every source. Boss posture is declared, not accidental. Every parameter is published, in the game and on this page. When the owner rules, the Shield, Bard, Chaos and Air cards unblock against this text instead of a rumor.
Seen in: World of Warcraft's category ladder, Diablo 3's never-immune posture, Final Fantasy XIV's overlap suppression, League of Legends' cap and exclusion list. The specification is the survey's consensus, assembled.
This card is the document plus the flag plan; cards S22 and S23 are the implementation. Behind the flag, behavior is bit-for-bit today's. The specification ships with a parameter-table test scaffold the implementation cards fill in, and the category table, cap, window and floor are page content the moment they are code.
a knockdown currently builds immunity to stuns; the per-type buffs that would fix it have sat authored and uncast for years
Replace the blanket stack with per-category stacks, built from the three resist buffs already sitting complete in the data plus two new ones authored on the same pattern for mesmerize and root. Four stacks of twenty percent reach the published eighty percent cap, so at least a fifth of any control's authored chance always survives: never immune by accumulation, the Diablo rule in chance form. Cross-category chains genuinely get more dangerous, a knockdown no longer shields you from stuns, which is why this card never ships silently: S14's feedback and S15's rings land in the same release, and the values move by reviewed changeset. The negative-resistance guard that protects the Gustball minigame survives untouched.
Seen in: World of Warcraft's seven-category ladder, and Albion's capped decay. The per-category budget is what makes control a resource instead of a lottery.
The blanket grant at the six landing sites becomes a per-category grant using the reserved rune assets, retuned to the specification's stack count and window. The existing stacking and refresh machinery carries it unchanged. Duration-domain scaling through S10's wired keyword is the specification's stated second step, gated separately because it touches every control rune's expressions. Editor tests pin category isolation, the cap, the exclusions and that flag-off equals today's blanket behavior.
breaking free today rewards you with one fifth of a resistance stack; you can be re-stunned within the second
No game in the survey ships an escape without an immunity window, because an escape without one is worthless: you are re-controlled inside the same breath. Shroud ships exactly that. Make a successful break grant a short, visible, category-scoped immunity window, four to six seconds, applied by the victim's own client, which is already the authority for control, and rendered as S15's ring. The price is the focus the jump already costs, printed at last: a resource decision, not a cooldown lottery, which is the half of this design Shroud already had right. The window after a control that expires naturally is the Air page's Grounded card; one shared effect serves both triggers, credited there.
Seen in: Elder Scrolls Online's Break Free package, stamina in, six visible seconds out; and in the genre's unanimous pairing rule.
The break-free success path grants the category-immunity effect instead of a lone ladder stack; the effect renders through the S15 telegraph; soft control is untouched per S19. Window lengths are changeset parameters, split for PvP only by S28's ruling. Creatures do not break free; their posture is S26's. Editor tests pin grant-on-success-only, category scope and the remote-authority path.
innate skills, gear, buffs and the ladder all pour into one number today, and only accidental clamps stand between that and immunity
A classless game must decide its stacking rule before it ships the second resist source, not after. Adopt the surveyed best answer: sources bucketed into groups, innate skill, gear, active buff, ladder stack; additive across groups so different kinds of investment all matter, capped at the published limit so no combination reaches immunity, and snapshotted at application so a mid-control buff change cannot retroactively rewrite a duration. The exclusion list is published beside it. This is the card that makes S17's new gear safe to ship and the card S8's emergency clamp graduates into.
Seen in: League of Legends' tenacity groups, the cleanest written answer to flat-versus-multiplicative in the genre, implemented in about twenty lines.
A source-group tag on resistance contributions, one accumulation function applying the group rule and cap at the landing roll's read, and the snapshot rule for in-flight durations. The maximally stacked build in S8's worked example loses a little headroom, stated as such, by changeset. Editor tests pin the group table, the cap and the snapshot.
the attunement multiplier is computed on every player-versus-player spell hit and thrown away
The option this whole page circles: take the resistance-driven attunement multiplier the spell path already computes and discards, and apply it, floored at a published fifty percent so mitigation never exceeds half, with the partial-resist chance rebased down in the same changeset so median outcomes hold. Ward points would finally do what every player already believes they do, deterministically, using the curve and tuning values that exist today: no new stats, no new data. It is also the largest balance surface on this page, every spell against every resistant creature moves, which is why it ships only on S30's ruling, sequenced against the balance program's undead wave so the same felt numbers never move twice in one patch, and never in the same release as the Heavy Armor page's spell ward.
Seen in: Ultima Online's 2003 retrofit, the genre's original chance-to-percentage conversion, and Elder Scrolls Online's two capped pools. The anti-pattern being escaped is EverQuest's coin flip.
Apply the already-computed multiplier at the end of the spell path with a published floor, rebase the partial-resist contribution, and re-baseline the golden damage tables. The migration and re-baselining are the cost, not the math. An editor test pins the floor, the curve and rebase neutrality at the median across an attunement-delta grid.
boss control immunity today is an accident of which of two authoring methods a designer used, and one of them does nothing
Some bosses were authored immune through resistance masks and are; thirteen were authored through the dead stat and are not. Declare the posture instead of inheriting the accident. Default: bosses resist hard, at the same published eighty percent cap as players, with true immunity reserved for deliberate per-ability immunity rules, the mechanism the undead bleed immunity already uses. The optional upgrade, if the owner rules for it, is the stagger-bar posture: a visible second bar that control damages, authored in verbal tiers, weak to massive, rather than raw integers, breaking into a short vulnerability window where control lands normally. The named disaster to avoid is the immunity that reapplies the instant it breaks. Undead mask numbers move only with the balance program's second wave, credited, never here.
Seen in: Diablo 3's posture as the default, Guild Wars 2's defiance bar as the upgrade, and Lost Ark's verbal tiers as the affordable authoring surface.
Posture A is S9 plus the cap: data and review, little new code. Posture B adds a bar value per flagged boss archetype, per-ability tier values, bar and break rendering, and a vulnerability window in which landed stuns stay normal-duration and breakable, honoring the no-fully-binding-stuns rule. Flagged, player-versus-environment only. Editor tests pin capped-not-immune, tier table, and break-window behavior.
The Bludgeon page reserved this call; Tier 3 is the specification it activates. Three shapes, in ascending ambition. A: the per-category ladder alone, capped, never immune, the cheapest and first to ship. B: A plus the stagger bar for flagged bosses, in verbal tiers. C: additionally move the ladder from chance into duration through the wired formula keyword, the most modern and the most rune-touching. Recommendation: A now, B on its own schedule, C as the stated destination. The day this rules, the Shield page's stun desk, the Bard page's peace verb, the Chaos page's shapeshift and the Air page's Grounded all unblock against S21's text.
The Bludgeon page's knockdown scope call, generalized: every player-versus-player parameter on this page in one ruling. Window lengths in PvP versus the world, whether the break-free band splits, whether the immunity ring shows to enemy players, and whether the knockdown category runs a shorter chain window. No card above hard-codes a PvP number; all of them ship as changeset parameters this ruling fills in.
One reviewed changeset activates the creature-side numbers this page wires: the thirteen dead-stat templates at values the owner picks under S26's cap, where one hundred may not mean immunity anymore; the dragon's mesmerize value rebased for S2's proportional shape, or converted to a deliberate per-ability immunity if dragons should truly never be mesmerized; and the wisp's always-landing mesmerize, kept as an authored puzzle through an immunity rule or left diminishable under the new system. Undead mask numbers stay with the balance program's second wave.
Whether S25 ships at all, its floor value, and its rebase policy, ruled together with the ward-passive maximum-versus-sum option from S18, because both redefine what a point of school resistance is worth and they must be one coherent answer. The ruling also fixes sequencing against the balance program's undead wave so determinism and the wall trim never move the same numbers in the same patch.
Thirty cards, three surfaces. One sheet panel, one feedback layer, one governor model: no widget zoo. Everything below is honest to the numbers found in the code, and half of the third mockup is explicitly conditional on the owner's ruling.
A level 80 defender in plate with a shield, Hard Headed at 200, Stand Your Ground at 80, a stun-resist masterworked helm, a fire ward gem, and two fresh ladder stacks from a rough fight:
ARMOR (physical hits only) Damage resistance 41.6 flat armor 34.2 + buffs 7.4 Absorption cap 84% of any single hit Avoidance 12 raises glancing chance, glance = 1/3 Armor does not reduce spell damage. WARDS (incoming magic) Ward pts Damage cut Fire 96 -12% Water 25 0% Air 75 0% Death 20 0% Magic (all but Chaos) 15 - Ward points contest the caster's attunement and raise your partial resist chance. Damage cut trims matching damage after the hit. The two are separate numbers. CONTROL (chance the effect fails) Base Active Stun 81% +40% Hard Headed 50, Stand Your Ground 25, helm 6 Knockdown 50% +40% Knockback 50% +40% Root 0% +40% Mesmerize 0% none Confusion 6% none Active: Resist Crowd Control x2, 21s remaining IMMUNITIES None active
Three deliberate honesty rules. Rows never hide at zero, because mesmerize's zero is the finding. The temporary ladder stacks print once and apply to every control row they touch, which puts the additive double-dip in plain sight: this build's stun total passes one hundred, today's accidental immunity, until S8's clamp and S24's rule bound it. And the ward table's two columns refuse to pretend the two fire numbers are one number.
Obsidian Golem
[############# ] 68%
CC [ o o . . ] IMMUNE 6s
(( white ring at its feet, visible to everyone ))
floaties, attacker's view:
STUN 1.2s (-40%) landed, shortened by resistance
Stun resisted roll failed; today this is silence
Immune window or authored immunity
Partial Resist 14 existing magic glancing, unchanged
combat log:
Your Ground Pound stuns Obsidian Golem for 1.2s
(2.0s base, reduced 40% by crowd control resistance).
Obsidian Golem resists your stun
(their control resistance is 60%).
Obsidian Golem is immune to stuns for 6s.
The delta is the teacher. A floaty that says how much was shaved teaches the ladder in one hit; a bare "resisted" teaches nothing, which the genre proved across twenty-five years of silent rolls. The ring is for the attacker: the entire point of telegraphing immunity is that the player about to waste a glyph is not the one who owns the buff.
SURFACE (no behavior change) buff bar: [Resist Crowd Control x3] tooltip: Chance of stuns, knocks, roots and mesmerize against you reduced 60%. 18s remaining. Gained each time a control effect lands on you. REPLACE (after the S27 ruling) nameplate, only while active: CC [ # # . ] stun next stun at 25% duration CC [ # . . ] knock next knock at 50% duration at the cap: CC [ # # # ] stun IMMUNE 6s (( ring, all viewers )) the printed ladder, same numbers here and in the game: 1st application full duration 2nd 50% 3rd 25%, then immune 6s window 15s duration floor 0.5s slows and snares never touch the ladder breaking free opens your window early
The Air page owns the stun forecast line. This page contributes only the data source: the forecast's resistance term reads the same governed state the pips display, so the two can never disagree, and while a window is open the arithmetic gives way to the immunity notice.
Bandit Footpad [======= ] 43% CC [#..] Stun forecast: base 29% + Charged 12% - their resist 20% = 26% or, window open: Stun forecast: immune 4s
| Phase | Name | Ships | Gate |
|---|---|---|---|
| 1 | The ledger and the book | S1 docs, S13 Defenses ledger, S16 truth pack, S19 written law | none; zero behavior change |
| 2 | The answered roll | S12, S14, S15 feedback and rings; the ladder surfaced, co-shipped with the Air page's cards | none |
| 3 | The repairs | S2–S6, S8 guard batch; S9 wired dark; S10 keyword decision; S11 audit; S20 struggle rework | Shield's Bolster Balance re-author lands before S8's clamp |
| 4 | The governor | S21 spec, S22 ladder, S23 package, S24 stacking rule, S17 gear, S29 values | the owner rules S27, and S28 for PvP values |
| 5 | The far wall | S25 determinism, S26 posture B, S7's mask | S30's ruling; the balance program's second wave for all creature-facing numbers |
flowchart TD
P1["Phase 1: ledger + docs"] --> P2["Phase 2: feedback + rings"]
P2 --> P3["Phase 3: the repairs"]
P3 --> P4["Phase 4: THE GOVERNOR"]
OC{{"OWNER RULES S27 + S28"}} --> P4
P4 --> P5["Phase 5: determinism + boss posture"]
OC2{{"OWNER RULES S30"}} --> P5
P4 --> W1["Bludgeon s26"] & W2["Shield S32"] & W3["Bard S22"] & W4["Chaos S29"] & W5["Air A14"]
A34["Air A3/A4 stun surfacing"] -.co-ship.- P2
M17["Moon S17 mez unification"] -.reads S2's pin.- P3
W12["Water S12 governed freeze"] -.enters the categories.- P4
BE["Balance program wave 2"] -.creature numbers.- P5
Page law. At most two of S17, S22, S23, S25 and S26 per release. S22 and S25 never share a release: one felt-defense revolution at a time. S25 never ships beside the Heavy Armor page's spell ward. S24's rule lands before S17's gear, and the pure-guard batch S3 through S6 plus S8 may ship together as one changeset with tests and no balance intent. Every edit to the landing roll's resist branches updates the remote-player deferral list in the same commit. And nothing on this page moves an undead number.
Four research passes feed this page: a code walk of the three resolution pipelines and every resistance store, writer and reader; a data-model sweep across stats, gear, skills, creature templates, interface surfaces and tunables; a prior-art sweep across the nineteen sibling proposal pages, the roadmap and the wishlist so nothing here re-proposes claimed territory; and the genre survey cited throughout. Every factual claim was verified against the shipping data in August 2026, and the load-bearing ones, the mesmerize asymmetry, the dead boss stat, the discarded attunement multiplier, were re-verified by hand. Full technical citations live in the team's internal twin of this document.
Every code-touching card names its editor test. The new pins: the mesmerize parity and scale pin, the three struggle defects, the focus-transfer gate, both divisor guards, the damage-bearing mask, the sum clamp's worked example, the dead-stat reader and its scale, the keyword semantics, the dead-surface dispositions, the ledger's row-to-reader coverage, the feedback emission and echo, the telegraph thresholds, the cap table, the exemption law, the struggle band grid, and the governor's parameter scaffold with category isolation, overlap suppression, window authority and flag-off equivalence. The existing pins this page extends rather than duplicates: the resistance stat-map, the Hard Headed resists, the Bolster Balance mask, the authored-chance bounds, and the control-glyph visibility guard.
Ultima Online: the official skill wiki and publish notes for Resisting Spells and the Age of Shadows resistance system, with the Second Age community wiki for the pre-2003 formulas. EverQuest: the EQEmulator community's classic resist study and the Magician's Tower debuff guides. Dark Age of Camelot: the Camelot Herald armor and resist tables and long-standing community consensus on the immunity timers, which have no first-party document. World of Warcraft: the Warcraft Wiki on resistance and diminishing returns, Maxroll's current DR reference, and the 4.0.1 and 5.0.4 patch records for the deletion. Elder Scrolls Online: UESP and ESO Academy on combat, Break Free and resistance caps. Guild Wars 2: the official wiki on the defiance bar, its predecessor, and Stability. Final Fantasy XIV: the community wikis on stun resistance and the PvP Guard and Purify. League of Legends: the wiki's tenacity article. Diablo 3: Blizzard's own 1.0.5 crowd-control developer journal, the strongest first-party source in this survey. Elden Ring and Path of Exile: the community wikis on poise and stun thresholds. Albion Online: the official wiki and community guides on crowd-control resistance. Lost Ark: community mechanics guides on stagger. Community-sourced numbers are flagged as such in the internal twin; where sources disagree, the page says so.
Thirty cards, three walls, one ladder finally shown. Comments and votes below; the rulings desk is open.
No account needed. Every proposal above also has its own Discussion box for talking about that one alone; this thread is for the plan as a whole. Comments are plain text, held to basic decency, and may be trimmed.