The Steel Between

Nearly every fighter in New Britannia straps on chain or plate before walking into trouble. Far fewer ever cast what the school actually teaches. Over the last thirty days on the live server, 462 players held an investment in Heavy Armor, 204 cast even one of its runes, and 21 used the skills that define its role in a group. This page walks all twelve skills and the mitigation math beneath them, traces the genre's armor and tanking designs back to Ultima Online, and lays out thirty concrete proposals to give the wall a will of its own. Every card is open for comments and votes below.

Executive summary

Heavy armor's price tag is the oldest design decision in the genre. Ultima Online decided in 1997 that a mage in plate keeps his walking speed and loses his mana engine: the Meditation skill simply refuses to work through metal. EverQuest invented Taunt two years later and with it the whole grammar of tanking, a number nobody could see called hate, and twenty years of addons trying to display it. Dark Age of Camelot gave the tank four different ways to stand in front of somebody. World of Warcraft formalized threat, then spent a decade learning that a tank whose survival lives entirely in the character sheet is bored, and moved mitigation into timed decisions. Elder Scrolls Online made heavy armor itself a skill line that levels by wearing it and answers crowd control with immovability. Final Fantasy XIV deleted the last damage penalty from its tank stance in 2019, and the genre effectively voted: the tank pays in something, but never in the ability to do their job.

Shroud's Heavy Armor school sits on the historically literate side of most of those arguments. Its costs are spell fizzle and triple durability wear, set deliberately in the June 2020 update that also shipped Provoking Shout, and never a step of movement speed, which both Ultima Online's precedent and the historical record of men vaulting onto horses in full harness would endorse. Its specialization is the rare one that visibly transforms the kit, doubling the armor value of every worn piece of chain and plate. Players notice: 53 percent of everyone who learns the specialization pushes it to 80, the strongest conversion signal of any armor school, against Light Armor's 19. But thirty days of live telemetry and the school's own data files tell three uncomfortable stories:

1. The role nobody can play. The school's group identity is aggro, and the aggro kit reaches 21 of 1,237 active players. Taunt adds a one-time flat bump to a threat table that damage feeds continuously, so any committed damage dealer buries it within a swing or two; its level scaling spends itself on cooldown and radius rather than magnitude. Threat itself is invisible, so a tank cannot tell whether the button did anything. The complaint has a nine-year paper trail, and the newest entry was filed this month.
2. The dishonest ledger. Verified defects stand between players and the school's real rules: the glancing-blow tooltip promised one third damage while the game dealt one quarter, fixed August 15 by raising the number to match the tooltip; Hard Headed lists three resists and grants one; Knight's Grace advertises an instant slow cleanse while a vestigial second effect fires about one percent of the time; Provoking Shout bills its focus cost to the Shield school's cost stat and wears Taunt's icon. And the largest single fact about armor appears nowhere a player can read it: damage resistance and absorption apply to physical damage only. The magic path subtracts neither.
3. The missing decisions. The school is maintained buffs and passives. There is no short-window mitigation to time against a big hit, no poise to hold a cast steady while the blows land, no per-piece scaling for the half-plate hybrid, and the protect verb is one genre generation behind. The code even carries finished, unused hooks: reaction timestamps for glancing blows that no rune reads, a protection stat with no reader, an armor penalty channel with no writer.

Thirty proposals follow, in four tiers.

The school we have

"Their ideas are good, but maybe there's a math issue or corner case code trap that renders them useless." — Akandriel, on Taunt and Provoking Shout, wishlist forum, August 1, 2026. This page is the investigation that post asked for.

Who actually plays it

Heavy Armor ranks 12th of 21 adventuring trees by specialization: 462 players learned the mastery, 243 pushed it to 80. That 53 percent conversion is the school's quiet endorsement. The specialization is the only one in the game wired directly into the damage pipeline, doubling the base armor value of every worn chain and plate piece at skill 80, and the players who try it keep it. Light Armor, the mirror school, converts at 19 percent. The pairings sketch the archetypes: Blades with Heavy Armor is the 11th most common specialization pair in the game at 97 players, the classic sword-and-board silhouette minus the shield school's own low numbers, and Bludgeons, Earth, Polearms and Tactics pairings follow.

The cast telemetry tells the other half. In the thirty days ending August 14, players cast 8,898,439 skills; the eight Heavy Armor runes account for 140,991 of them, 1.58 percent, by 204 distinct casters. The funnel from wearing to playing narrows hard:

Thirty days on the live server

[Body Slam]        60,464 casts    86 players   (1)
[Inner Strength]   29,248 casts   161 players   (2)
[Glancing Blow]    28,678 casts   138 players
[Knight's Grace]    8,294 casts    51 players   (3)
[Absorb Impacts]    6,692 casts    52 players
[Taunt]             3,015 casts    16 players   (4)
[Valiant Warden]    2,359 casts    22 players   (5)
[Provoking Shout]   2,241 casts    12 players   (4)

school total ..... 140,991 casts   204 players
combo: Titan Smash     746 casts     7 players

Concentration runs high: the top five casters produce 42 percent of the school's entire cast volume, and the single busiest produces 17. The school's public life is a handful of committed tanks and a wide population borrowing two buffs. The passives never appear in cast telemetry at all; their story is the 462-learner census above.

The tree as the skills window shows it

Glancing Blow      L1    Chance for blows to glance
                         for partial damage
Body Slam          L10   Charge and stun. Cannot be
                         used while rooted
Inner Strength     L10   Raises Strength. Usable
                         outside combat
Deflection         L10   Passive. Raises damage
                         absorption
Absorb Impacts     L40   Raises damage resistance.
                         Wears armor faster        (1)
Hard Headed        L40   Passive. Raises stun
                         resistance                (2)
Endurance          L40   Passive. Raises Strength
Knight's Grace     L40   Breaks stuns, clears slows
                         and bleeds                (3)
Valiant Warden     L40   Carries a share of one
                         ally's harm for 20 seconds
Taunt              L40   Draws the attention of up
                         to 15 nearby foes
Provoking Shout    L80   Raises all threat you
                         generate for a time       (4)
Specialization     L80   Doubles the armor value of
                         worn chain and plate

The mitigation math, in order

Every proposal on this page plugs into one pipeline, and none of its order is stated anywhere in the game. When a physical hit lands on an armored character, the game rolls whether the blow glances (armor's damage avoidance feeds that chance, and a glancing hit deals one third damage), then applies block and parry, then subtracts flat damage resistance, capped so that resistance can remove at most a fixed share of the hit. Specialization intervenes three times: it doubles each worn chain or plate piece's base resistance, it raises the absorption ceiling, and it lowers the big-hit threshold above which resistance scales itself up, which quietly makes specialists better against exactly the crits that kill tanks.

flowchart TD
  H[Incoming hit] --> P{Physical or magic}
  P -->|Physical| G[Glancing roll
chance rises with Damage Avoidance
a glancing hit deals 1/3]
  G --> B[Block and parry]
  B --> DR[Flat damage resistance
per piece, doubled per chain or plate piece
at Specialization 80]
  DR --> CAP[Absorption ceiling
resistance removes at most a capped share
of the hit. Specialization raises the cap]
  CAP --> SC[Very large hits
damage above the scaling threshold
scales resistance up to meet it]
  SC --> OUT[Damage applied]
  P -->|Magic| MG[Partial resist roll only, 0.25x] --> MB[Block and parry] --> MOUT[Damage applied
no armor resistance, no absorption]
  style MOUT stroke-width:3px

What the armor costs

Heavy armor charges three currencies and famously refuses a fourth. A full plate set adds roughly 64 points of spell fizzle chance before skills reduce it, which is the wall between plate and serious spellcasting; only the Focus school sells relief, and the school itself contains none, which is deliberate. Plate takes durability damage at three times cloth's rate, a multiplier set alongside Provoking Shout in the June 2020 update. Every rune in the school bills focus through its own cost stat, which masterworked heavy helms can discount. And there is no movement or dexterity penalty anywhere in the code: the slow knight is a myth the game correctly declined to implement, and modern experiments in genuine 15th century harness agree. Heavy Armor's own runes never fizzle; they are combat skills, not spells. What this page fixes is not the tax but the receipts: the fizzle number a player actually pays is assembled from five sources and displayed in none of them (card S17).

Lessons: Ultima Online to the modern tank

Shroud's ancestor invented this school's central bargain. In 1997 Ultima Online let anyone wear anything, and charged the mage who chose plate in the currency that mattered to him: the Meditation skill, the engine of mana recovery, simply refuses to function in metal armor. Movement was never touched. That is precisely Shroud's fizzle tax, twenty years early, and UO also demonstrated how the bargain dies: the later Mage Armor item property deleted the meditation penalty piece by piece, and once every caster could buy the exemption, armor choice collapsed into a gold cost. The 2003 Age of Shadows publish then replaced armor ratings with rolled item resists, and the community still argues that armor identity never recovered. Both halves of that story are load-bearing here: the tax belongs on the character, and no item may ever sell the way around it.

EverQuest built the other pillar in 1999. Taunt set your hate to one point above the current leader, hate itself was invisible, and two decades of threat-meter addons were born from that opacity. World of Warcraft formalized threat with multipliers and stances, capped armor's physical reduction at 75 percent on a curve of explicit diminishing returns, and then delivered the genre's most important tanking lesson in 2012: once threat generation stopped being the moment-to-moment problem, passive-mitigation tanks were bored, so Mists of Pandaria moved survival into short, resource-priced defensive windows. The tank's skill expression became timing, and it never went back. Dark Age of Camelot had already shown how rich the protective grammar could be, with four distinct verbs for standing in front of somebody: Guard, Protect, Intercept and Engage. Warhammer Online then perfected the form in 2008: its Guard split half a groupmate's damage and a third of their hate onto the tank, but only within thirty feet, which made bodyguarding a positioning game and is still cited as the best tank mechanic ever shipped in PvP.

The modern entries frame Shroud's choices from both sides. Elder Scrolls Online is the closest structural cousin: heavy armor there is a skill line that levels by being worn, scales its passives per piece so a five-heavy hybrid is a real build, and spends its one active on immovability, six seconds of immunity to knockback and disable. It also carries the cautionary scar: the passive that returned resources when struck made heavy armor mandatory in PvP for years. Final Fantasy XIV solved tank scarcity by deleting every penalty from tank stance and moving all skill expression into a layered rotation of mitigation cooldowns. Guild Wars 2 ran the control experiment, shipping no taunt and no threat table, and had to bolt tanking back on for raids by making bosses fixate the highest-toughness player: proof that some deterministic answer to "hit me instead" is load-bearing in group content. Dark Souls contributed poise, the best non-numeric armor fantasy in games: heavy armor buys the right to not flinch, to finish your swing while being hit. And the tabletop tradition supplies the oldest ideas still worth importing: D&D's Heavy Armor Master shaves a flat amount off every physical hit, chip-damage immunity that percentage reduction cannot express, and Pathfinder's Bulwark trait lets the armor itself stand in for agility, which is exactly the argument Glancing Blow has been making quietly all along.

The capability matrix

Rows are things a heavy-armor player might reasonably expect the school to offer. A filled dot is a first-class mechanic, a half dot is partial or conditional, an empty dot is absent. The two Shroud columns are the point of the table.

"SotA today" holds two filled dots on that board, and one of them is the cost column. The proposed column fills the rest without inventing a new resource system: every row lands on plumbing the game already runs, and three of the biggest land on hooks the code has carried, finished and unread, for years.

The thirty proposals

Numbered S1 through S30, in four tiers: the wrong numbers first, then the kit, then the new verbs, 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.

Tier 1 · Straightened Rivets · eight cards of true numbers

Tier 2 · The Refitted Harness · the kit repaired

Tier 3 · New-Forged Steel · verbs the genre has proven

Tier 4 · The Council Table · rulings before numbers

Considered and rejected

• A third armor tree. Asked again this month; ruled out in the team's armor deep dive, and the ruling stands here. The parity ask underneath it belongs to the Light Armor page.
• Reviving medium armor. The equipment category exists in code with zero items and stays a reserved slot; a third tier of the same trade is a spreadsheet, not an identity.
• Any mobility or dexterity tax. The school has never charged legs, the code contains no such penalty, and the historical record of men running obstacle courses in full harness backs the game's choice. Imports that arrived with speed penalties attached, including the 2018 stance proposal, were taken without them.
• An item property that deletes the fizzle tax. Ultima Online ran this experiment as Mage Armor and it dissolved the school's entire cost structure into a gold price. The tax stays on the character.
• Removing the durability trade. Plate's triple wear was set deliberately in June 2020. S14 reshapes its harshest cliff; nothing here deletes the cost.
• Taunt as an absolute lock. Reverses the recorded 2017 ruling; S9 targets contention and S28 asks the owner to set the ceiling explicitly.
• A new aggro skill. Provoking Shout was 2020's answer to exactly that request. The recorded ask is to make the existing two scale, and that is S9.
• An engage-style total-defense channel. Collides with the one avoidance buff channel that Glancing Blow, Tumble and Dig In already compete for; the guard fantasy ships through S12 without the collision.
• Continuous weight-fraction costs. Elden Ring's equip-load, smoothed, is elegant and opaque; it is also a whole-game cost rebalance, which is the balance program's desk, not this page's.
• Toughness-driven aggro. Guild Wars 2's raids taunt by gear stat; elegant, and players tank by accident. S9 and S10 keep aggro deliberate and visible.
• Pet aggro parity. Real ask, wrong page; creature target selection and pet taunting belong with the taming page's territory, and S9's shared threat-table repair likely helps pets on the way through.
• Attribute-based crowd-control resists. Akandriel's dexterity-versus-roots mirror is a Light Armor and attributes proposal, noted for that page.

The composed kit

The test of thirty cards is whether they compose into one playable half minute. Below is a tank's thirty seconds with the proposed kit, every element riding a UI surface the game already has: the buff bar, the nameplate, the combat log, the hotbar. No new window is proposed anywhere in this catalog.

[ BUFF BAR ]    Glancing Blow 22s | Absorb Impacts 8s | Inner Strength 4m
                Provoking Shout 38s | Brace 4s | Poise steady          (1)

[ TARGET ]      Troll             fixated on YOU                       (2)
[ WARD ]        Warden bond: Elowen   9m of 10m                        (3)

[ COMBAT LOG ]
  Your armor turned the blow. Glancing hit for 61.                    (4)
  Brace caught the slam. 214 damage prevented.                        (5)
  Your plate held your footing. Casting continued.                    (6)
  Elowen took 180. You carried 90 of it.                              (3)
  Focus returned: struck while in full plate.                         (7)
  The troll turns to face you.                                        (2)

[ HOTBAR ]      Q Taunt (ready) | E Brace (2s) | R Warden swap | F Slam

• 1The bar carries two new entries at most: Brace (S19) and the poise state (S18). Everything else is today's kit with honest numbers.
• 2The fixation line and the log's last sentence are S10: threat, finally visible, driven by S9's taunt that scales.
• 3The ward line is S12: a leash readout that makes bodyguarding a positioning promise.
• 4Sixty-one is a third of the roll, the fraction the tooltip always promised: S2's honest glancing, shipped, opening S25's riposte window.
• 5The timed window doing its one job: S19.
• 6Poise holding a cast through a hit: S18.
• 7The struck tank's focus trickle: S23, capped by design.

Sequencing

flowchart LR
  S28[S28 ruling] --> S9[S9 taunt repair]
  S9 --> S13[S13 aura]
  S10[S10 threat visible] -.pairs.-> S9
  S4[S4 resists] --> S18[S18 poise]
  S5[S5 cleanup] --> S11[S11 uptime]
  S5 --> S20[S20 immovable]
  S2[S2 honest glancing] --> S25[S25 riposte]
  S21[S21 shave] --> S27[S27 materials]
  S25 --> S27
  S1[S1 book, built] --> S24[S24 per piece]
  S9 --> S29[S29 PvP, parked]

The power budget. Six cards raise tank power in earnest: poise, Brace, the shave, immovable, the focus trickle and the spell ward. No release should ship more than two of them, and the trickle and the ward never together, because sustain plus magic mitigation in one patch is the exact recipe that made heavy armor mandatory in another game's PvP for half a decade. And one optics rule, owed to the Light Armor thread that watched this page being researched: every wave that ships a heavy power raiser ships heavy repairs and legibility beside it. The Light Armor page carries the mirror catalog, so the parity ledger has two open books.

Appendix

Where the numbers come from

Constraints honored

• Combat resolves on player machines; the server holds no threat table and no positions. Every card computes identically on every client from shared data, and threat displays are labeled as the local view.
• The stat, skill and equipment enums are append-only and serialized by index. The catalog adds at most one unconditional skill identifier (Brace), two creature rows, and up to three stats, all appended, all mirrored to both sides of the shared tree.
• Armor mitigation is physical-only, and every mitigation card states which path it touches. Exactly one card (S26) touches the magic path, on purpose, priced.
• The avoidance buff channel is shared with Tumble and Dig In; nothing in this catalog adds a competitor to it.
• The school's costs, fizzle and durability and focus, are design and stay. No card sells an exemption.
• All final numbers ship through the balance program. Itemization belongs to the armor crafting page, stances and reactions to Tactics, the Shield and Light Armor kits to their own pages.

The test surface

Cards whose changes are pure formula carry EditMode pins as part of their cost, not as an afterthought: the piece-count table (S8), the glancing multiplier (S2), the shave ordering (S21), the per-piece fraction tables (S24), the relative-threat math and its boost interaction (S9), the fizzle assembly (S17), the durability curve (S14), the Knight's Grace regression (S5), and the resist-row parse (S4). The school currently has no dedicated test class; wave 1 creates it.

Sources and credit

Game systems consulted for the lessons section: Ultima Online's armor rating, Meditation rule, Mage Armor property, Parrying, and the Age of Shadows resist retrofit; EverQuest's Taunt and the birth of the trinity; World of Warcraft's threat formalization, armor curves, and the Mists of Pandaria active mitigation rework; Dark Age of Camelot's Guard, Protect, Intercept and Engage; Warhammer Online's Guard and the Ironbreaker; Elder Scrolls Online's heavy armor skill line, Unstoppable, Constitution and its PvP history; Final Fantasy XIV's tank stance history and mitigation rotation doctrine; Guild Wars 2's trinity experiment and toughness fixation; Dark Souls and Elden Ring's poise and equip load; Monster Hunter's lance; D&D fifth edition's heavy armor rules and Heavy Armor Master; Pathfinder second edition's Bulwark trait and armor specialization effects; and the historical record, including modern mobility experiments in genuine fifteenth-century harness. This game's own record: the skills window, the corrected school reference, the armor crafting page of this series, the specialization research notes, and the telemetry and census described above.

Heavy Armor's costs were set with care: fizzle for the caster, durability for the purse, and never a step of speed, a choice both Ultima Online's meditation rule and the historical record would endorse. What the school owes its players now is smaller and harder: numbers that mean what the tooltips say, a taunt that repays ten years of investment, and a handful of real decisions to make while the blows land. Nothing in this catalog is final. The defect fixes are filed, the rest are arguments, and the balance work ships elsewhere. If you carry plate for a living, the comment threads under each card are where your dents count. Credit where it is due: Akandriel, Osunn, Punkbeach, Segalleon2, Ysold, DalTXColtsFan, Cora Cuz'avich, Hemswal The Descended, Numa, Nikko, Antrax Artek, FBohler, Bahamut Arc, and the posters of nine years of taunt threads whose reports shaped this page.

General comments

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.