By Open source contributorsUpdated 5 min read
A forest settlement in Blizzard's WoW Forever preview
On this page

Arcane, Fire and Frost use distinct talent and proc priorities. These level-60 single-target studies complement the beginner dungeon guide; they do not replace its control, movement and party-preparation advice.

This guide uses our own saved, reviewed simulation builds. Original references retain each study's saved support. KFC's Solo and Raid rankings are separate comparisons against the same five-minute level-63 boss, with different support packages. These remain estimates from pinned beta models, not player logs or a promise of live performance.

Tested builds, talents and equipment

Level 60 · pre-raid references

Mage build notebook

Frost

Original build reference572.2 DPS40,000 validation fights · original support
Talent allocation17 / 0 / 3451 points · level 60

The two DPS values use different support/encounter policies. Only the Solo and Raid comparisons belong in the tier list, using their own exact setups.

Save your current simulator character first. Loading replaces the saved setup; never silently substitute another engine.

Talents in the tested build

These are the allocated ranks in the recorded model's tree order. Descriptions are labelled client references, not independent tests proving that every point is optimal.

Arcane FocusArcane · 5/5

Improves your chance to hit with Arcane spells by 5%.

Confirmed client-reference rank
Improved ChannelingArcane · 5/5

Gives you a 100% chance to avoid interruption caused by damage while channeling Arcane Missiles and a 70% chance while casting Arcane Blast or Arcane Wrath.

Confirmed client-reference rank
Arcane ConcentrationArcane · 5/5

Gives you a 10% chance of entering a Clearcasting state after any damage spell hits a target. The Clearcasting state reduces the mana cost of your next damage spell by 100%.

Confirmed client-reference rank
Arcane MeditationArcane · 2/3

Allows 33% of your Mana regeneration to continue while casting.

Confirmed client-reference rank
Improved FrostboltFrost · 5/5

Reduces the casting time of your Frostbolt and Frost Blast spells by 0.5 sec.

Confirmed client-reference rank
Elemental PrecisionFrost · 5/5

Improves your chance to hit with Frost and Fire spells by 5%.

Confirmed client-reference rank
Ice ShardsFrost · 5/5

Increases the critical strike damage bonus of your Frost spells by 100%.

Confirmed client-reference rank
Piercing IceFrost · 3/3

Increases the damage done by your Frost spells by 6%.

Confirmed client-reference rank
Frost ChannelingFrost · 3/3

Reduces the mana cost of your Frost spells by 15% and reduces the threat caused by your Frost spells by 30%.

Confirmed client-reference rank
Ice LanceFrost · 1/1

Deals 28 to 32 Frost damage to an enemy target. Deals 200% increased damage to Frozen targets.

Confirmed client-reference rank
ShatterFrost · 3/3

Increases the critical strike chance of all your spells against Frozen targets by 50%.

Confirmed client-reference rank
Improved Cone of ColdFrost · 2/3

Increases the damage dealt by your Cone of Cold spell by 23%.

Confirmed client-reference rank
Fingers of FrostFrost · 2/2

Gives your Chill effects a 15% chance to grant you the Fingers of Frost effect, which treats your next 2 spells cast as if the target were Frozen. Lasts 15 sec.

Confirmed client-reference rank
Winter's ChillFrost · 5/5

Gives your Frost damage spells a 100% chance to apply the Winter's Chill effect, which increases the damage done by your Cone of Cold spell to the target by 40%, stacking up to 5 times.

Confirmed client-reference rank

The recorded rotation

Level 60 against one stationary level 63 Dragonkin for 300seconds, at 10yards behind the target, using the measured gear, buffs and consumes.

Read from the top. Use the first available action whose conditions hold, then start again. These are recorded model instructions, not a separately tested simplified rotation.

  1. Use Demonic Rune when available if you are missing at least 1200 mana.
  2. Use the enabled cooldowns, potion and carried-gem engine policies when available.
  3. Use Evocation when available if (your mana is less than 10% and the fight has more than 25 seconds remaining).
  4. Use Ice Lance when available if Fingers of Frost is active.
  5. Use Cone of Cold when available if Winter's Chill has at least 5 stacks.
  6. Use Frostbolt (Rank 11) when available.

The tested equipment

One complete loadout, with its recorded enchants. Not independent per-slot BiS rankings.

HeadSpellweaver's TurbanArcanum of Focus
+9 Intellect+36 Spell power+1 Spell hit input+73 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
NeckNacreous Shell NecklaceNo enchant recorded
+8 Stamina+8 Intellect+9 Spell power+1 Spell hit input

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
ShouldersRugged Mantle of the TimbermawChromatic Mantle of the Dawn
+15 Intellect+27 Spell power+1 Melee crit input+75 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
BackShroud of Arcane MasteryEnchant Cloak - Greater Resistance
+10 Stamina+11 Intellect+1 Spell hit input+43 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
ChestRobe of the ArchmageEnchant Chest - Greater Stats
+12 Intellect+40 Spell power+1 Melee crit input+96 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
WristsSublime WristguardsEnchant Bracer - Healing Power
+6 Stamina+10 Intellect+6 Spirit+12 Spell power+37 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
HandsGloves of Spell MasteryEnchant Gloves - Minor Haste
+10 Intellect+9 Spell power+2 Melee crit input+60 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
WaistKnowledge of the TimbermawNo enchant recorded
+22 Intellect+25 Spell power+1 Melee hit input+56 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
LegsMagister's LeggingsArcanum of Focus
+12 Stamina+15 Intellect+21 Spirit+7 Spell power+0.9 Melee hit input+76 Armor

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
FeetEarthen Silk SlippersEnchant Bracer - Greater Stamina
+7 Spirit+18 Spell power+33 Armor+6 Spell damage

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Ring 1Blessed Band of LightNo enchant recorded
+10 Stamina+8 Spirit+1 Melee hit input+31 Healing power+10 Spell damage

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Ring 2Rune Band of WizardryNo enchant recorded
+7 Stamina+16 Spell power+1 Spell hit input

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Trinket 1Briarwood ReedNo enchant recorded
+29 Spell power

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Trinket 2Weakness AnalyzerNo enchant recorded
+22 Spell power

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Main handPremier High Warlord's SpellbladeEnchant Weapon - Winter's Might
+8 Stamina+7 Intellect+94 Spell power+1 Melee crit input

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Off handScepter of Interminable FocusNo enchant recorded
+9 Spell power+1 Spell hit input+1 Spell crit input

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Ranged / relicBonecreeper StylusNo enchant recorded
+4 Intellect+11 Spell power

Pinned simulator inputs, not a live-game tooltip. Hit, crit and similar inputs retain engine units; current item data and access may differ.

Open item reference
Talents, race, consumes and exact options

Talent string, in the model's tree order:

05500050000002--055500033100032025
race: "RaceOrc"
consumes: {"flask":"FlaskOfSupremePower","food":"FoodRunnTumTuberSurprise","manaRegenElixir":"GreaterMagebloodElixir","spellPowerBuff":"GreaterArcaneElixir","firePowerBuff":"ElixirOfGreaterFirepower","frostPowerBuff":"ElixirOfFrostPower","mainHandImbue":"WizardOil","defaultPotion":"MajorManaPotion","defaultConjured":"ConjuredDemonicRune","zanzaBuff":"CerebralCortexCompound"}
buffs: {"blessingOfWisdom":"TristateEffectImproved"}
bonusStats: {}
itemSwap: {}
cooldowns: {}
healingModel: {}
mage: {"options":{"armor":"MageArmor"}}
Assumptions, limitations and source hashes
  • These instructions reproduce the measured beta engine APL, not a global optimum or guaranteed live-server DPS. Different movement, target count, encounter length, party buffs or consumes can change the ranking.
  • The model uses Forever Fingers of Frost, Ice Lance and Winter's Chill interactions. Do not substitute Classic crit-debuff rules or assume Frost Nova freezes a raid boss.
  • Maximum spell ranks and their launch trainer/book availability remain an explicit model assumption. Client or spellbook presence is not proof of a release unlock.
  • Automatic cooldown timing, carried-gem thresholds, potion policy and Cold Snap are delegated to the pinned engine. This first pass did not exhaustively optimize manual timing, weaving or every possible Mage rotation.
  • Demonic Rune causes self-damage. Its modeled mana trigger has no low-health protection: prioritize survival in real encounters.
  • Practical pre-raid first pass in the pinned Forever beta model, with item weapon-skill bonuses ignored. It is not a global BiS proof or live-server DPS guarantee.
  • The incumbent was fixed before two independent 20, 000-fight ranges. The separate 1% assessment is preserved alongside the owner's 2% retained-contender tolerance; nearby gains remain disclosed.
  • Level 60 stationary single-target Frost at 10yards behind a level 63 Dragonkin for 300seconds, with fixed buffs, Mage Armor, Wizard Oil and distinct Greater Mageblood 20MP5. Movement, party setup and encounter duration can change rankings.
  • Only eligible non-raid gear, enchants and crafting dependencies are admitted. Uncapped PvP is allowed; historical acquisition and launch unlocks still need in-game verification.
  • Historical no-Mageblood blue/gear/state studies are starting-point evidence only. Every current-model gear/state comparison and fresh confirmation was actually measured on the Greater 20MP5 engine, without importing old-engine scalar origins.
  • These instructions reproduce the measured beta engine APL, not a global optimum or guaranteed live-server DPS. Different movement, target count, encounter length, party buffs or consumes can change the ranking.
  • The model uses Forever Fingers of Frost, Ice Lance and Winter's Chill interactions. Do not substitute Classic crit-debuff rules or assume Frost Nova freezes a raid boss.
  • Maximum spell ranks and their launch trainer/book availability remain an explicit model assumption. Client or spellbook presence is not proof of a release unlock.
  • Automatic cooldown timing, carried-gem thresholds, potion policy and Cold Snap are delegated to the pinned engine. This first pass did not exhaustively optimize manual timing, weaving or every possible Mage rotation.
  • Demonic Rune causes self-damage. Its modeled mana trigger has no low-health protection: prioritize survival in real encounters.
Presentation release
a5f95c49c57e46f727b0267a
Engine SHA-256
1e5d84fd3d73d93ec47b18d8dcf05998ec21095d2431986bacc7a4fedc2b23e4
Catalog SHA-256
43fa27f482434a950cfcd755a1ccc560f85d5128d2e6f1e200ae1bb28e01a24b
Publication SHA-256
1a8e9754a1db096e93a2c4f3869cb1b91dd0376880cc5d6cdc8dc3f64703c171
Rotation SHA-256
0b8f7555e6963597ec8ad85863b12c2705dbf0ccb8a77eeab4aa017ad7e41826
Historical comparison request
faa9b1aa575260cc46f6c0b8615cca7b6ae3fe0972befcf8766f45363104a39d
Historical comparison response
c2d13e5d713b480d6104a876cbf5aa64fbd477a69db0c3e9790be9d3f4e88ec4

Client reference 1.60.1.70425. Data adapted from talentsforever.com (CC-BY-4.0). Model and gameplay validation remain distinct.

Decisions that matter for Mage

Read the conditions, not just the buttons

Read the actual proc and aura conditions before spending the next cast. A source-bound priority is more useful than a rigid sequence that assumes the same procs on every pull.

Keep the encounter attached to the result

Keep mana recovery and cooldown timing in the full reference. A short target may end before a planned recovery, while a five-minute comparison exposes the cost of unsustainable spending.

Compare complete setups

Compare staff and one-hand/off-hand equipment as complete packages. The benchmark retains the selected talents, equipment and consumes rather than isolating a universal value for one stat.

How to use a recorded priority

Check the list from the top and use the first available action whose conditions hold, then start again at the top. Availability still includes the model's resource, weapon, position, cooldown and spell requirements. An AND group requires every condition; an OR group is a choice within that group. A disabled action is not part of the active priority.

The instructions reproduce the saved rotation. They are not a separately tested simplified rotation or a claim that a player can reproduce perfect automated timing. Movement, target death, mechanics and party responsibilities can require a different decision. Do not change several abilities and continue quoting the original result as though the experiment were unchanged.

Avoid these comparison mistakes

  • Using a single-target position to claim the best dungeon AoE spec.
  • Importing proc behavior from a different WoW version.
  • Assuming a close numerical ordering will survive movement and target changes.

Dungeon control, level-30 balance, AoE farming and PvP need their own scenarios.

Gear is a complete reference, not a universal shopping list

The equipment and enchants above are the actual tested set. They are not an independently ranked list of best items in every slot. Weapons, paired rings, trinkets and set bonuses interact, and the studies do not establish a precise loss for every possible replacement. Intentionally empty slots remain visible instead of being filled with invented recommendations.

Pinned engine stats are labelled as such; hit, crit and similar model inputs must not be silently converted into client ratings. Open the linked item reference to inspect current data and evidence labels. Some references permit high-rank PvP, reputation or expensive crafted rewards. A non-raid item is not necessarily an easy launch-day acquisition, and historical sources do not prove current Forever availability.

Keep the selected race, professions, consumes and support when reproducing the original result. The Solo and Raid benchmarks change support and encounter by an explicit policy, rather than relabelling the original 40,000-fight reference. Different class profiles retain their pinned engine/catalog versions; version identity and model limitations remain part of the comparison.

What the DPS comparison means

Choose the comparison that matches your support:

The rankings cover 21 published builds across eight classes. Newer research builds in this guide may not be ranked until their matching simulator setup is published.

Our Solo vs Raid DPS comparison guide explains which result to use and how to reproduce it. A small difference can be less important than movement, uptime, utility or a model limitation. Tiers describe the selected reference characters; they are not a rule for excluding players from groups.

This is a level-60 single-target reference. It is not a level-30 beta dungeon ranking, an AoE ranking, a healing table or a PvP prediction. The recorded client source is build 1.60.1.70425; later changes require revalidation rather than a cosmetic date update.

Try your own controlled comparison

Save your current simulator character before loading a reference: loading can replace equipment, talents, race, professions, buffs, consumes and encounter settings. Change one decision at a time, retain the version, duration and support, and record both setups. A new result from a changed setup is a new experiment, not an update to this published reference.

The download contains the saved settings and release hashes. A public replay link is shown only when that exact release is available; a local reviewed build is not represented as already deployed. Never substitute another engine silently when reproducing a result.

Share a useful correction

Use the Mage talent reference to check the named choices. Bring a build link, the exact condition or item, source version and reproducible before/after result to the community Discord. Do not send account credentials or unrelated private messages.

Need a party rather than another spreadsheet? Find a group. You can use the guides and join the wider community without joining a particular guild.

Review history

  • October 23, 2026: first simulation-backed class edition, with exact-source rotation instructions and a matched-policy DPS benchmark. Gameplay/log validation remains separate.
Find your next group