Mechanics · Spell efficiency

Mana efficiency and spell ranks

Compare useful output per mana, account for overhealing and understand the trade-offs of lower spell ranks.

Reviewed 4 October 2026 · Arithmetic and official beta notes

A spell can deliver more healing per cast and less healing per mana. Both statements can be true. The useful choice depends on how much output you need, how quickly you need it and what happens to the mana you spend.

Three numbers to keep separate

Output per cast measures the damage or healing delivered by one cast. Output per mana measures how much you receive for its resource cost. Output per second also includes the time taken. A rank with the best ratio can still be too slow or too small for the situation.

For an ordinary comparison, start with average results from the same character, target and conditions. Include critical hits and unsuccessful outcomes consistently across both samples. Comparing one unusually large critical heal against an average normal heal tells you little about the ranks themselves.

The basic accounting is:

useful output = gross output × (1 − unused fraction)
output per mana = useful output ÷ mana spent

Use one unit throughout: damage per mana or healing per mana. Comparing damage with healing does not establish which spell is better.

Overhealing can reverse the result

Consider two hypothetical healing options. These are arithmetic examples, not Forever spell values.

Option B supplies 60 more useful healing per cast. Option A supplies 60% more useful healing per mana: 4.8 divided by 3.0 is 1.6. Neither result tells you whether the next hit will kill the target before a heal arrives.

If your log already gives effective healing, enter that as the output and set unused output to zero. Subtracting overhealing a second time would understate efficiency. The same rule applies to damage: do not remove missed attacks or lost periodic damage twice when your average already includes them.

What downranking changes in Forever

Downranking means deliberately casting a lower spell rank. You might value a smaller heal, a shorter cast or a utility effect, but check the actual rank rather than assuming that every lower rank has the same properties.

Blizzard's 1 October 2026 beta notes say ranks far below the character's level receive less benefit from Spell Damage and Spell Healing, with reduced activation chances for class abilities and talents. Their example is rank 1 Frostbolt at level 60: it cannot activate Frostbite in that example. Official beta notes

The notes do not provide a complete penalty formula or a table for every rank and talent. A Classic coefficient or a raw client coefficient therefore cannot settle the best Forever rank at your current level. The example at level 60 also describes a future-level case relative to the current level-30 beta cap. Official beta level cap

A useful test compares ranks at the same character level. Keep gear, talents and target conditions fixed; record which spell ID and rank you used. Repeat after gaining levels if a character-level penalty could have changed the result.

Choose an appropriate mana denominator

For a simple cast comparison, use the mana actually charged for the cast under the stated conditions. This can differ from a base client field. A mana return is a separate event: record it separately unless you can attribute an average return to this spell and explain the accounting.

A free cast has no finite output-per-mana ratio. Do not divide by zero or replace its cost with an invented small number. Compare its output, timing and availability directly. The calculator displays a free-cost result separately.

Regeneration changes how long you can keep casting

Cast efficiency and sustained mana balance answer different questions. A spell can have a strong output-per-mana ratio while a rapid sequence still spends mana faster than your character recovers it.

For a defined interval with unchanged mana-pool rules, include all expenditures and gains:

ending mana = starting mana − mana spent + mana received

Count recovery that actually reached the mana bar. Recovery lost at the mana cap does not increase the mana available afterward. Compare full-regeneration and reduced-regeneration time separately when your observed rates differ. Talents, temporary buffs and active mana gains belong in that interval's conditions.

The comparison below does not model regeneration, cast times, proc chances, threats or survival. It isolates the output-per-mana question so you can make that part of the decision explicit.

Compare two options

Enter observed values or an explicitly hypothetical scenario. Leave no ambiguity about whether the output is gross or already useful. An unused percentage of zero is a stated assumption, not evidence that every cast lands efficiently.

For periodic spells, use the same observation window and target count for both options. A full-duration tooltip total and a short-lived target's actual result describe different situations.

Reading the result

The higher ratio buys more useful output per mana under your inputs. Also inspect useful output per cast: a smaller efficient heal may still require another cast, and a larger spell may prevent an emergency. Cast timing and utility can change the choice even when the ratios are correct.

Keep a comparison with its build, character level, spell ranks, talents, buffs and measurement method. Those details are what turn a calculation into advice another player can reproduce.