Mechanics · Weapon and ability scaling

Weapon speed, Attack Power and ability scaling

Compare ordinary weapon swings and understand how weapon abilities, scripted attacks and class mechanics change the calculation.

Reviewed 4 October 2026 · Client sources and dated beta changes

WoW Forever beta · Client snapshot 1.60.1.70205 · Official changes dated 1 October 2026 · Checked 4 October 2026

A slower weapon can produce a larger hit without producing more damage per second. An ability can depend on Attack Power without using your weapon’s swing time. Those distinctions matter when comparing two weapons: first identify the attack you want to improve, then compare the parts that attack actually uses.

The comparison below is a limited arithmetic model for ordinary weapon swings. Client data supports its displayed AP basis; it does not establish measured combat DPS or a complete formula for every ability.

Ordinary swings

The matching Forever character-sheet module divides effective Attack Power by 14 when displaying its nominal damage contribution. The pinned addon load sequence places that module after an earlier file containing a different value, so the later assignment supplies the basis within that declared sequence. This is evidence about the client tooltip, rather than a live server damage test. Matching client module, addon load sequence.

Using that basis, with AP held constant and the weapon’s listed, unhasted speed:

• Nominal AP contribution per second = AP ÷ 14.
• Average base weapon hit = base item DPS × listed speed.
• Nominal average ordinary hit = (base item DPS + AP ÷ 14) × listed speed.
• Nominal ordinary DPS = base item DPS + AP ÷ 14.

These equations exclude misses, critical strikes, glancing blows, mitigation, off-hand adjustments, haste, procs, abilities and lost attacking time. They are a conditional comparison, not a prediction of damage against a target. The distinction between a weapon’s average hit and its damage per second also appears in the matching character-sheet code. Client damage display.

For a hypothetical example, give two weapons 40 base DPS each and keep 280 AP:

Both receive 20 nominal DPS from AP. The slower weapon concentrates more damage into each swing; the faster weapon delivers smaller swings more often. All numbers in this example follow the stated model and hypothetical inputs.

Weapon abilities

An ability’s weapon calculation can differ from an ordinary swing. Normalization is one reason.

As a Classic reference, the original WoWSims implementation uses a separate reference speed for the AP component of a normalized attack, while keeping a base weapon damage component. Normalization therefore does not automatically make two weapons with equal item DPS equivalent for an ability. This describes that simulator’s implementation; it does not establish Forever’s reference speeds. Original Classic simulation code.

Holy Strike is a useful Forever example. The level-6 spell record contains normalized weapon damage, a weapon-percentage effect and a script effect. Its effect rows are not a complete server formula. A zero in a direct AP coefficient field does not rule out AP entering through the weapon calculation; a displayed spell coefficient does not prove that the entire hit uses that coefficient. Holy Strike’s current database entry, build-stamped effect table.

For a Holy Strike comparison, the useful questions are the base weapon damage, the ability’s weapon fraction, its AP treatment and any separate spell or scripted component. The ordinary-swing calculator cannot answer all four. Calling an attack “Holy” identifies its damage school; that label alone does not determine its AP or Spell Damage scaling.

Scripted attacks

Published changes can be more informative than an isolated coefficient field:

The ratios come from Blizzard’s 1 October notes. The display differences are visible in the current Bloodthirst entry, Swipe entry and 1.60.1.70205 effect rows.

An AP ratio describes a contribution to an ability, not its entire damage or DPS. Base damage, rank, modifiers, target defenses and usage frequency still need their own evidence. A raw spell-effect field is especially easy to misread when the ability also contains dummy or scripted effects.

Stat conversions and damage modifiers

An attribute bonus, an AP bonus and a damage bonus act on different quantities. Treat their interaction as a specific talent or buff rule; their names do not establish a universal stacking order or rounding rule.

For example, Blizzard’s 1 October notes give Champion of the Light an Intellect-to-Spell-Damage ratio of 20/40/60%, and remove its incorrect healing description. Official change.

The current generic entry still displays a larger value, while its rendered talent ranks show the newer progression. The native base record also differs from the rank-specific description. Compare the rank you actually have, and keep Spell Damage and healing separate. These records do not justify adding the older generic value to the newer talent value. Current Champion of the Light entry, native effects.

The matching character sheet also has conditional AP-from-SP and SP-from-AP display paths. Their existence does not establish that either conversion applies to every class, build or ability. A Paladin’s ordinary swing and Holy Strike need separate scaling evidence even when the same gear improves both. Client conversion paths.

Timing, forms and other exceptions

Next-swing attacks and off-hand attacks. Heroic Strike is marked as a next-melee attack. Blizzard’s 1 October notes say queueing it no longer improves off-hand hit chance. An older queue-based hit bonus should therefore not be imported into a Forever comparison. This does not supply a complete off-hand hit or damage formula. Heroic Strike entry, dated fix.

Ranged attacks. The client reads ranged AP separately, and Aimed Shot has a normalized weapon effect. Use the relevant ranged attack’s rules when comparing a Hunter’s weapons. The manual model below does not include ammunition or shot-specific calculations. Client AP branches, Aimed Shot.

Druid forms. The matching client treats Cat Form’s Agility-to-AP display separately, and Swipe’s current description requires Bear Form or Dire Bear Form. The equipped weapon’s listed swing is not a verified substitute for a form’s attack calculation. No form damage formula is assumed below. Client form handling, Swipe’s form requirements.

Haste and procs. Listed weapon speed and the time between current attacks are different inputs. In the reference simulator, attack timing changes separately from the weapon’s AP damage calculation, and some proc models use weapon-speed-based chances. That is a reason to check each named effect, rather than award every faster weapon a generic proc advantage. Forever’s effect-specific rules and their DPS values remain unverified here. Original simulation implementation.

Compare two weapons

Enter each weapon’s base item DPS and listed, unhasted speed, then one shared AP value. Use the item tooltip’s weapon DPS, rather than the character-sheet DPS that may already include AP and other changes.

The shared AP value holds your character constant to compare the weapon damage lines. If a weapon swap changes your AP, talent effects or buffs, that is a further comparison. This model also excludes off-hand attacks, forms, normalized specials, hit outcomes, mitigation, haste and procs.

Reading the comparison

Nominal average hit describes the model’s damage concentrated into one ordinary swing. Nominal DPS describes that damage divided by the listed swing interval. A larger hit is not automatically a larger DPS result.

Because AP is shared, its nominal contribution cancels from the DPS difference: that difference equals the change in base item DPS. Swing damage can still differ because the weapons have different listed speeds.

An equal result means the weapons are equal within these inputs and assumptions. It does not settle Holy Strike damage, resource generation, proc value, burst damage, ranged shots or form attacks. For an ability-specific decision, record the ability and rank, talent ranks, form, gear and beta build alongside the result. The unresolved parts then have a clear question to test.