See what every champion actually contributes, where they fit, and how to build them for specific encounters. Every recommendation comes from the game's mechanics—not community rankings or subjective opinions.
Explore the System ↓Every result comes from a fixed, layered process. Each stage has a single responsibility — no step re-interprets logic from another. This is what keeps outputs consistent and predictable no matter how many times you run the same query.
Each tool answers a different question about your RAID roster. Use them individually for a quick answer, or together as a full build workflow.
Enter a champion and see their encounter coverage across the entire game at a glance. The radar shows every content category simultaneously — click any node to drill into specific bosses and see exactly which mechanics your champion brings to that fight.
Flip the question. Pick an encounter and see every champion in the roster ranked by what they actually contribute to that fight. Filter by faction, affinity, or rarity to find the best option among champions you own or are considering pulling.
Pick any encounter and see its full mechanical profile — what mechanics reward you, what helps but isn't essential, what actively hurts you, and what's completely wasted. Use it to understand a fight before you gear a team for it.
Enter a champion name and an encounter, and get back a concrete build recommendation. Stat priorities are labeled High, Medium, Low, or Off — so you know exactly which numbers to chase and which to ignore for that specific fight.
Upload your RSL Helper export and see every champion in your collection scored against every encounter. Coverage gaps surface immediately. No account required, no data uploaded — everything runs in your browser against the same engine that powers the Radar and Tier List.
A different kind of tool than the five above — this one isn't about scoring a champion against an encounter, it's about planning the resource spend for a fusion event itself. Readiness checklists, a spend/hold calendar, and cost-efficiency calculators for the specific sub-events that make up a fusion, built as each new fusion is announced.
The tools are designed to work as a simple workflow. Here's the recommended sequence for getting the most out of the system.
Decide which piece of content you're trying to clear or improve in. The system produces different results for each encounter, so the fight you target matters. Use the Encounter Guide to understand the mechanical profile of a fight before building a team for it.
Type your champion's name into the Champion Radar and click into the relevant encounter node. The tier rating and Role Identifier Chips show whether this champion is a strong fit before you invest in gearing them. Or use the Tier List to find the best options across your whole roster for that fight.
Open The Forge with that same champion and encounter. It outputs stat priorities, artifact set recommendations, main stat slots, and minimum floors — everything needed to guide your gearing decisions.
Every result page links directly to HellHades and AyumiLove for the same champion. This system's data-first output and their experienced player knowledge work well together — use both to make the most informed decision.
Most RAID tools change depending on who you ask, when you ask, and how the meta shifts. This one doesn't. Here's why that matters.
When a content creator rates a champion as "S-tier," that rating is implicitly based on a specific piece of content, a specific gear level, a specific team composition, and the current meta. If any of those change — or if your situation is different — the rating may not apply to you at all.
This system doesn't have opinions. It reads what each champion's skills actually do, compares those mechanics against what each encounter specifically demands, and produces a score. The score is the same every time you run the same query, because it's derived from the same fixed data.
Turn Meter manipulation is one of the most powerful mechanics in RAID — but it does nothing against the Demon Lord, who is immune to it. A champion whose primary value is Turn Meter control will score very differently for Fire Knight vs. Demon Lord in this system, because the encounter rules are baked into every calculation.
Similarly, poison damage is irrelevant against Ice Golem Hard, which is fully poison-immune. The Forge reflects this by recommending zero Accuracy for poison-focused champions in that fight — because building for debuffs that won't land is a waste of stats.
RAID champions are categorized as Attack, Defense, HP, or Support type — but that label doesn't always tell you how to build them. Alure is an Attack type whose real value is Turn Meter drain on critical hits, not raw damage output. The Forge recognizes this and prioritizes Critical Rate and Speed over Attack, because that's what actually makes her effective.
The same logic applies across the full roster. Coldheart's Heartseeker deals damage based on enemy max HP, so Critical Damage matters more for her than raw Attack. These nuances are handled consistently, for every champion.
This isn't another content creator's opinion tool. Here's what makes it different.
Enter the same champion and encounter once, ten times, or a hundred times — you should get the same result. There's no randomness, no hidden weighting, and nothing changing behind the scenes. The logic is fixed, and every result can be traced back to how it was calculated.
Tier lists rank champions in isolation — "S-tier" or "A-tier" without telling you what content, what gear, or what team they're based on. This system treats every evaluation differently: a champion's mechanical capabilities are the key, and an encounter's mechanical demands are the lock. How much of that lock the key can turn determines the tier — not a community consensus, not a content creator's assessment. The same champion might be S for Spider and D for Demon Lord, because those fights have entirely different locks. That context is built into every score this system produces.
The system detects whether a champion has a mechanic, not how strong that mechanic is. A champion with 2.5% poison and one with 5% poison both cover poison_single — the vector treats them the same. Skill depth (the count) does carry some weight in the scoring, but we don't model exact percentages, cooldowns, or conditional triggers. The scores tell you whether a champion's kit fits an encounter's demands, not how optimally built or booked they are. Use the scores as a directional signal, then cross-reference HellHades or AyumiLove for the finer details.
Some encounter mechanics are survival requirements — if your team cannot answer them, you will lose regardless of overall strength. For example, in the Fire Knight encounter, you must have enough multi-hit attacks to break his shield before you can deal meaningful damage. The system will track which mechanics directly solve these requirements and weight them accordingly.
Every score in this system measures one thing: how much of a specific encounter's mechanical demand a champion covers. The Radar shows that for one champion across all encounters. The Tier List shows it for all champions on one encounter. In both cases the comparison is always champion-versus-encounter, never champion-versus-champion. Kael and Duchess Lilitu aren't ranked against each other — they're each measured against Spider Hard independently, and the numbers happen to be comparable as a result.
The Refinery reads game data and converts it to tags. The Radar scores champion fit against encounter demand. The Tier List inverts that to rank champions per encounter. The Encounter Guide shows what a fight needs. The Forge outputs a concrete build. Nothing bleeds across — each tool has a clear scope.
The Radar scores a champion's fit against a specific encounter and places them into one of five tiers: S, A, B, C, or D. The bands are relative to the field, not a fixed cutoff table — S, A, B, and C are set from how every champion in the database actually scores against that encounter, so they reflect real standing, not an arbitrary number picked once and left alone.
Exceptional fit. The champion's kit covers the encounter's most critical mechanics and contributes strongly across the board. An S-rated champion for a given encounter is one of the best possible choices for that fight — their abilities directly answer what the encounter demands, with very little wasted.
Strong fit. Handles the encounter's key pressures reliably. An A-rated champion holds their slot effectively and covers most of what the fight asks for. Minor gaps exist but they're manageable — a solid pick that rarely lets the team down.
Moderate fit. The champion contributes meaningfully but leaves gaps that the rest of the team will need to cover. A B-rated champion is usable but not optimal — they work, but you may notice their limitations in harder content.
Limited fit. The champion addresses some mechanics but misses a significant portion of what the encounter demands. A C-rated champion can fill the slot if nothing better is available, but the team carries extra burden as a result.
Poor fit for this encounter. Unlike S through C, D isn't a relative band — it means the champion's net contribution to this specific fight is zero or negative, whether that's doing nothing useful or actively working against you. A D rating doesn't mean the champion is weak overall — it means there's a real mismatch between what they do and what this encounter asks for. They may be excellent elsewhere.
Because S through C are set from the live champion database, a tier can shift slightly as new champions are added — that's intentional. "How good" is meant to track "how good right now," not lock to a number decided once and never revisited.
Common questions about the system, its methodology, and how it differs from other RAID tools.
The full write-up on how the engine works, the rest of the FAQ, and the running changelog — expand any section below.
Derived facts, not opinions. Every output on this site is a mechanical fact pulled from the game's own skill text and encounter rules — no inference, no vibes, no "I think this feels strong."
Every champion in the system is mapped across roughly 357 mechanical axes. These aren't categories or vibes — they're facts derived by reading each champion's skill text. Does this champion place Poison? Does it require a critical hit to trigger? Does it reduce Turn Meter on enemies, or fill it on allies? Does the damage formula scale with HP, ATK, or DEF?
The same process runs on the other side: every encounter is mapped against the same set of axes. What does this fight reward? What does it punish? What buffs does it strip? What debuffs does it ignore?
Once both sides are mapped, the tools don't guess — they calculate. The encounter tells the system what it needs. The champion tells the system what it provides. Everything else follows from that.
The encounter is the lock. The champion is the key. The tier grade (S through D) measures how much of that lock the champion can turn.
A perfect key answers every mechanical demand the encounter places. A poor key fits the keyhole but barely moves the tumblers. Champion Radar tells you exactly how far the key turns — and for which locks it works best.
Team Matrix, inside the Roster Analyzer, extends this: your team is a keyring. The question becomes whether the ring contains keys that collectively open the lock — or five copies of a key that only fits the same tumbler.
The grade is always encounter-specific. A key that opens one lock may be useless against another. Same champion, different fight, different tier.
The system doesn't summarize champions. It reads their skill text and derives facts. Here's what that looks like in practice.
Champions are only half the picture. The system covers every major PvE encounter — from every Clan Boss difficulty, every dungeon stage, and Doom Tower bosses, to Hydra, Chimera, Faction Wars, Grim Forest, and Arena. Each encounter is mapped against the same axis set as the champions. Here's what that data looks like.
Same champion, same encounter, same result. It doesn't have good days or bad days. It doesn't drift. If you run a champion through the same fight today and again next month, you'll get the same priorities — unless the underlying data has been updated to match a game change.
This is the difference between a system built on derived facts and one built on opinion. Opinions shift. Facts don't — until the game changes them.
It doesn't rank champions against each other in a vacuum, and it doesn't tell you who is "best" in general. It answers specific questions: what does this champion contribute to this fight, and what do they need to do it well. Everything outside that scope is your call to make.
RAID has over a thousand champions, and new ones ship every few weeks. A system built on opinions requires someone to have an opinion about every champion for every encounter — and keep that current. A system built on data just needs the data to be accurate.
A few more questions that come up, beyond the ones already answered above.
A running list of things we fixed, things we broke, and things we broke while fixing other things.
NEW Fusion Tools launched — readiness checklists, a resource spend/hold calendar, and cost-efficiency calculators for the sub-events that make up a fusion (Champion Training, Champion Chase, gear enhancement events). Built fresh for each new fusion as it's announced, starting with the Cowardly Lion.
NEW Gear Enhancement's own tool ended up as a decision flow rather than a cost chart, after finding directly that the game never discloses upgrade success rates past +4 — only vague "High/Low" labels, never a real percentage. A precise cost curve built on top of that would have been presenting a guess as a fact, so the tool instead shows when a piece is cheap enough to keep pushing and when to walk away, without needing to know the hidden odds at all.
NEW Full Roster now supports searching and filtering by any effect a champion's kit actually produces — not a curated shortlist, but the complete set, including mechanics that don't have a polished display name. Select multiple effects and only champions with all of them show up, same logic as stacking any other filter.
NEW Each selected effect becomes its own column, showing that champion's real strength value for it, sitting right alongside the existing Rarity, Rank, Level, Faction, and Affinity columns. Selecting an effect sorts the table by it immediately, strongest first — from there it behaves like any other column already did: click it again to reverse, click a different one to sort by that instead.
CHANGE The Roster Analyzer split in two. The live version now ships with only Unmatched — the one sub-tool solid enough to stand on its own. Full Roster, Hidden Gems, Power Up Next, Power Up Priority, Best Per Encounter, Team Matrix, Team Foundations, and Tag Team Arena moved to a separate developmental build, to be promoted back individually as each one gets properly reworked rather than all shipping at once in an unfinished state.
FIX Faction ID 17 now maps to Argonites, same as 16. RSL Helper's own export inconsistently uses both IDs for the same faction — champions were silently losing their faction filter tag depending on which one their export happened to use.
FIX Simulate removed entirely from Full Roster — the max-rank/max-books/awaken toggles and the resulting projected-gain chart. The readiness formula behind it had a confirmed bug: it assumed every champion's level cap was a flat 60, when RAID's real per-rank level cap scales with rank (10/20/30/40/50/60) — so a champion sitting at their own actual maximum was scored as only half-leveled. On top of that, the formula blended rank, level, and books into one multiplier with no real derivation behind the weights. Removed outright rather than patched, since there was no way to leave it running quietly with either problem still in it. Simulation is staying on the roadmap, not gone for good — it needs real research into how RAID's stat curve actually works before it comes back.
CHANGE The Roster Impact Chart is reframed around a simpler question: how does your whole roster stack up against each encounter, right now. No simulated deltas, no projected gains — just the real, current total.
FIX Chart bars narrowed and labels switched from a diagonal angle to fully vertical. The old width meant anything past a couple dozen encounters needed horizontal scrolling just to read the labels.
NEW Chart now sorts highest-to-lowest by score instead of alphabetically.
NEW Area filter tabs added above the chart — Doom Tower, Amius, Hydra, Demon Lord, and the rest — reusing the exact same area groupings Champion Radar already uses, so "Doom Tower" means the identical set of fights everywhere on the site, not a second, silently different definition of the same word.
NEW Chart bars link directly to the new User Tier List tab, pre-filtered to that encounter.
NEW User Tier List added — same idea as the site-wide Encounter Tier List, scoped to just your own uploaded roster instead of every champion in the database.
CHANGE Hidden Gems no longer factors rank, level, or books into the ranking at all. The previous version multiplied a champion's value score by how unfinished they were — meaning two champions with identical mechanical capability could rank differently purely because one happened to be more built. That's exactly the blend this tool isn't supposed to make.
CHANGE Ranking metric switched from summed tier points to a count of real S/A-tier fits, excluding Faction Wars and General PvE's outsized reward lists. Summing every encounter's tier points rewards broad-but-mediocre coverage over a narrow, genuinely excellent kit — confirmed directly on a real roster, where a Rare with three merely-good reward axes outscored a Legendary with one very strong one, on the same fight.
FIX The champion-existence gate was wrong in both directions on the first pass. Excluding rank-6 champions missed the real case — players often have rank-6 champions built years ago and completely forgotten. But testing against a real roster then showed the opposite failure: several rank-6, level-60 champions were still surfacing as "hidden," and they clearly weren't — reaching max rank and level takes substantial, deliberate investment that a genuinely overlooked champion wouldn't have. Gate corrected to exclude champions at rank 6 and level 60 together, while leaving books and blessing grade out of the gate entirely, since neither of the two rank-6 examples that prompted the fix were anywhere close to fully booked or blessed — rank and level alone were doing the real signaling.
FIX Continuous Heal was flagged internally as non-stacking. It stacks — same as Poison, same 10-stack cap. Any champion or build relying on layering multiple Continuous Heal instances was being undervalued. Corrected.
FIX Three internal effect definitions existed as accidental duplicates — one clean version the engine could actually use, and a second orphaned copy nothing pointed to correctly. Consolidated all three so scoring pulls from a single source of truth instead of a fallback guess.
FIX Separately, found and closed a 6-axis gap in the reference position map used by the offline verification script — two immunity effects (HP Burn immunity, Decrease DEF immunity) had been correctly defined in the live engine's data all along, but were missing from this one downstream reference copy. Confirmed this never affected any live score on the site itself; it only affected the standalone script that reads that reference file directly.
CHANGE The scoring engine now emits a straightforward S/A/B/C/D letter grade instead of a raw 0–7 number that didn't mean anything to anyone who wasn't us. Higher ceiling, wider spread, easier to read at a glance.
FIX Champion Radar's color coding and shape were still built for the old 0–7 scale. Every score now landing above 7 meant the radar shape maxed out and every bar looked the same color. Radar now colors and scales directly off the S/A/B/C/D letter, one unified palette across dots, bands, and rows.
FIX Roster Analyzer's Hidden Gems, Rank Up Next, and Book Next categories were quietly returning empty for everyone after the letter-grade switch — same root cause as the radar bug, different page, all fixed now.
CHANGE Hidden Gems is being rebuilt with a clearer job description: it surfaces mechanically capable champions on your roster you're probably not using — full stop. It doesn't care how booked or ranked they are. That's what Rank Up Next and Book Next are for. Three tools, three questions, no overlap.
NEW Chimera Normal and Hard — reflect damage plus a wall of crowd control. Bring cleanse or bring regret.
NEW Demon Lord split into all 5 real difficulty tiers instead of one blanket "Clan Boss" entry. Each difficulty now scores champions against what that specific difficulty actually demands.
NEW Dragon Hard added as its own encounter, separate from Dragon Normal.
NEW All 5 Faction Wars color bosses re-mined from boss skill text for sharper mechanic coverage.
CHANGE 10 existing encounters patched with mechanics the engine had been missing — Spider and Spider Hard now correctly value Heal Reduction and Cleanse, Hydra and Sand Devil value Stun, Scarab King values HP Burn, Iron Twins values Cleanse, Celestial Griffin values Cleanse and Block Revive, Bommal values Revive. Champions that cover those mechanics score higher where it actually matters now.
NEW New encounter tag: buff-stripping is now specifically valued for fights where the boss buffs itself and taking those buffs away is half the strategy.
NEW Site-wide visual overhaul — dark theme, sharper typography, consistent styling across every page instead of whatever each tool happened to look like when we built it.
CHANGE The sub-tools have real names now: The Refinery reads a champion's kit and extracts what it actually does. The Crucible scores that champion against a specific fight. The Forge turns the score into a build. The Casting Hall (coming soon) builds the whole team, not just one champion.
CHANGE New homepage. The Forge — the tool most of you actually use — moved off the root URL to make room for it, so bookmarks may need updating.
FIX Scoring no longer double, triple, or quadruple-counts effects from the same mechanic family. A champion who lands Stun, Freeze, and Sleep was getting credit for three separate CC contributions when an encounter only needed one. Now the engine takes the best contribution per family and moves on.
FIX Mandatory-mechanic gates (the "you basically need Cleanse for this fight" rule) were applying as a hard, all-or-nothing cutoff regardless of how close a champion came to qualifying. Replaced with a proportional penalty, so a near-miss doesn't get treated the same as bringing nothing at all.
NEW CRUCIBLE — a new tool. Type any champion name and get a full breakdown of what they do, where they're useful, and what role they fill. No encounter selection needed. Just a name. (This tool later became Champion Radar.)
NEW Role Fit — every champion gets classified based on what their kit actually does. Debuffer, Nuker, Support, Controller, Tank — derived from real skill data, not vibes.
NEW Encounter scoring — see how well a champion fits every encounter we track. Higher score means their kit lines up with what the fight needs. Lower means bring someone else.
CRUCIBLE and The Forge share the same champion and encounter data. One database, two tools.
NEW ACC and SPD floors are now calculated from actual boss stats. Boss has 225 RES? ACC floor is 250. Boss runs at 190 SPD? Speed tune target adjusts to match. No more vague "High" when we mean a number.
CHANGE HP and DEF floors removed from encounter recommendations. Those depend on team comp and gear quality, not the encounter itself, so a single number here would have implied more precision than the data actually supports. Removed rather than left as a misleading floor.
CHANGE Adding new encounters to the site is now a matter of minutes instead of hours. When new content drops, expect faster coverage.
NEW Spider, Fire Knight, and Ice Golem each split into Normal and Hard encounters. Different difficulties have different boss immunities, which means different champions are viable. Pick the one you're actually running.
NEW Spider Hard — HP Burn on spiderlings heals the boss. If your strategy is "burn everything," that doesn't work here. AoE nuke the spiderlings instead.
NEW Fire Knight Hard — TM reduction is useless, but Freeze drops 15% of the boss's Turn Meter. Completely different gearing from Normal.
NEW Ice Golem Hard — Poison doesn't land. HP Burn gets a 10% damage bonus. Normal mode is the opposite.
FIX Dragon's poison-immunity tag was incorrect — poison is a viable strategy there. Corrected.
NEW Clan Boss now properly rewards Unkillable and Counterattack team comps — the two mechanics that matter most for that fight.
NEW Debuff/buff magnitude now factors into scoring directly — 5% Poison scores higher than 2.5% Poison, 60% Decrease DEF higher than 30%, instead of every instance of a given effect being weighted the same regardless of size.
NEW A1 skills valued higher for debuff placement. If a champion has Poison on their A1, it's available every single turn with no cooldown. That's worth more than Poison buried on a 5-turn cooldown skill, and the scoring reflects it.
CHANGE Major accuracy pass across all 1,000+ champions. Every champion re-analyzed from their actual skill descriptions. More accurate results across the board.
NEW All 5 Faction Wars color bosses added with full mechanics — each has unique skills that change which champions you should bring. Red boss ignores DEF and has irresistible Stun. Green boss retaliates with HP Burn. Yellow boss retaliates with irresistible Freeze.
NEW All 5 Keeps added with real boss data — Force Keep (Thorn Da'vil), Magic Keep (Princess Krila), Spirit Keep (Sylvania), Void Keep (Malek Kavar), Arcane Keep (Sir Gallaroth).
NEW Cursed City Normal and Hard added — Amius the Lunar Archon.
CHANGE Total encounters: 76. Up from 69. Nice +7.
NEW FORGE now runs on our own champion database — 1,006 champions with full skill descriptions, base stats, factions, affinities, and auras. Every result comes from parsed skill data, not community opinion.
NEW 22 newly released champions added, including the Argonite faction batch.
CHANGE Database updates happen faster now. New champions and balance changes get picked up without delay.
NEW FORGE only loads game mechanic rules that are relevant to your champion. Kael gets poison rules. Martyr gets counterattack rules. Nobody gets rules they don't need. Faster, more focused results.
NEW Encounter-specific mechanics apply automatically — Clan Boss TM immunity is always factored in when you select Demon Lord, regardless of your champion.
FIX Rathalos Blademaster no longer gets ACC High on Dragon. His A1 Decrease DEF can't be resisted on bosses, so ACC does nothing for it there — the ACC-priority logic now accounts for irresistible debuffs specifically.
NEW Tauraze — HP Burn retaliation. Counterattacks, shields, and everything is on fire. Bring cleanse and Block Debuffs.
NEW Draugnell — poison everywhere. Think Nether Spider's younger, angrier sibling.
NEW Isheth — Shield or die. If you thought Scarab King was annoying, wait until you meet this guy.
NEW Maximoz — TM manipulation and AoE. Classic "bring Decrease ATK or get deleted" boss.
NEW Leshun — irresistible CC. RES doesn't save you. Block Debuffs does. Bring a cleanser or bring revive.
All 5 searchable by boss name — just type "tauraze" or "leshun" into the encounter field.
CHANGE Encounter difficulty now scales survivability targets directly — Clan Boss Normal calls for HP 20K / DEF 1,500, UNM for HP 50K / DEF 3,500, with speed targets staying roughly flat between difficulties since survival stats are what actually double.
CHANGE Cleansers get faster in cleanse-heavy fights. Shielders get beefier against Scarab King. HP Burn placers get ACC boosts where HP Burn matters. The engine gets smarter about what each encounter actually needs from your champions.
NEW Gear main stats (Gloves, Chest, Boots) visible for everyone. No locks. No paywall. Just answers.
NEW Recommended artifact sets shown on every build, correctly accounting for mixed piece counts — Savage (4-piece) and Cruel (2-piece) can stack on the same build, and the recommendation logic reflects that.
NEW HellHades Raid Optimiser priority column. Copy the number straight into HH. You're welcome, HH users.
NEW Encounter stat floors in the Min column. Now you can see "ACC 250+" instead of just "High" and wondering what "High" means.
NEW Bug report link on every result. It pre-fills your champion and encounter because we're considerate like that.
NEW "Arena - Go First" — for when you want to outspeed and nuke. SPD floor 250.
NEW "Arena - Go Second" — for when you want to sit there in Stoneskin and make your opponent regret queueing. HP/DEF/RES floors. Speed actively suppressed.
CHANGE Generic "Arena" remains as a balanced option for people who haven't picked a side yet.
FIX Ultimate Deathknight in "Arena - Go Second" now correctly says: be a wall. Previous behavior: told UDK to go fast. UDK has never gone fast in his life.
FIX Bomb champions (Gnishak, Lord Shazar, etc.) no longer told to build C.Rate and C.Dam. Bombs deal fixed ATK-based damage and can't crit at all, which the engine's generic crit-scaling rule didn't account for. Excluded specifically for these 20 champions.
FIX Guaranteed crit champions (Titus, Ithos, etc.) no longer told to build C.Rate. If every hit crits automatically, why would you stack C.Rate? C.Dam boosted instead because every hit CRITS. 17 champions fixed.
FIX Trunda's ACC now depends on where you're taking her. Arena? Off — her extra hits proc on targets NOT under stun, so landing the stun literally reduces her damage. Dragon? High — stunning wave enemies is helpful. Context matters.
OOPS Marichka was classified as a bomber. She places Block Damage when allies RECEIVE bombs. She protects from bombs. She is not a bomber. We're sorry, Marichka.
NEW Crit-dependent healing is now a thing the engine understands. Apothecary's A2 heals 35% normally, 52.5% on a crit. C.Rate isn't a damage stat for him — it's a healing stat. 34 champions affected.
OOPS A pattern meant to detect "increases enemy cooldowns" was matching "Increase Speed" + "Cooldown -1" from book notes. This silently told ~400 champions they needed ACC when they absolutely did not. Four hundred. We found it because Apothecary — a champion with zero debuffs — was showing ACC Low. Sometimes the smallest champions cause the biggest fires.
FIX Enemy MAX HP champions (Acrizia, Seer, Royal Guard, Coldheart, Armiger) no longer told to build ATK. Their damage comes from the enemy's health bar, not their own attack power. ATK suppressed. C.Dam boosted. 12 champions.
FIX 111 HP-scaling champions now correctly prioritize HP for damage instead of ATK.
FIX 152 DEF-scaling champions now correctly prioritize DEF for damage instead of ATK.
CHANGE Michelangelo's ATK is now High. His passive places a Shield equal to 300% of his ATK every time he's hit. He also taunts. So he gets hit a lot. ATK isn't a damage stat for him — it's a "please don't die" stat. 6 ATK-scaling shield champions affected.
NEW The engine now reads what your champion's skills actually do and applies the game's mechanics consistently. No opinions, no guesswork.
NEW The engine understands encounters. Demon Lord is immune to Turn Meter manipulation. Fire Knight needs multi-hit. Dark Fae copies your buffs and uses them against you.
NEW 111 HP-scaling and 152 DEF-scaling champions now correctly deprioritize ATK.
NEW About page. FAQ page. Written in plain English for humans, not SEO bots.
NEW 69 encounters. Nice.