[Bug]: AoE spell coefficients in SpellData inconsistently authored vs TBC empirical values

[Xerrion]

Problem

PhDamage's Data/SpellData_*.lua files store spell-power coefficients for AoE/multi-tick spells using inconsistent conventions. Some entries pre-bake the TBC AoE penalty (× 1/2), others store linear unscaled values, producing systematic damage errors of up to ~50% on affected spells.

Background: The TBC AoE rule

Per WoWWiki Spell_power archive (oldid=1549180, July 2008, patch 2.4.3 era):

Area of Effect spells receive only 1/2 of the total bonus, divided among each tick (for damage-over-time effects). Formula: C = CastTime / 7 (instead of CastTime / 3.5).

This is the TBC 2.0+ rule. Vanilla 1.x used × 1/3; patch 2.0 relaxed it to × 1/2. Patch 2.2.0 patch notes raised Hellfire / Blizzard / Rain of Fire coefficients further, consistent with this relaxation.

Per-spell secondary penalties (slow/snare/daze) stack on top of × 1/2 and don't follow a clean rule - they must be hard-coded per spell from the empirical table, not recomputed.

cMaNGOS-TBC SpellMgr.cpp::CalculateDefaultCoefficient does NOT apply a runtime AoE multiplier. Emulators consume per-spell coefficients from spell tables (post-penalty already baked in). PhDamage should follow the same convention.

Empirical TBC values (canonical)

From the WoWWiki July 2008 table:

Spell	TBC empirical (total)	Per-tick	Currently in PhDamage	Status
Arcane Explosion (instant)	21.43%	n/a	0.213	OK
Holy Nova damage (instant)	16% (split)	n/a	0.161	OK
Cone of Cold (instant)	13.57% (× 0.633 snare)	n/a	TBD	Needs audit
Consecration 8s ground	95.24%	~11.9%/tick	TBD	Needs audit
Blizzard (Mage) 8s/8 ticks	76.19%	~9.52%/tick	0.952	WRONG
Hurricane (Druid) 10s/10 ticks	128% (= 142.86% × 0.9 slow)	12.8%/tick	1.07	WRONG
Rain of Fire 8s/4 ticks	95.24%	23.81%/tick	0.5727	WRONG
Hellfire (enemy) 15s/15 ticks	214.29%	14.29%/tick	0.4286/tick	WRONG
Hellfire (self damage) 15s	142.86%	9.52%/tick	(same path?)	Needs split
Volley (Hunter) 6s/6 ticks	No SP scaling - ranged AP only	n/a	0.50	WRONG

Proposed solution

Pure data fix - NO engine changes:

1. Audit every *_AoE-flagged entry across all Data/SpellData_*.lua files.
2. Replace coefficients with canonical TBC empirical values from the table above.
3. Split Hellfire into two SpellData entries (or a hybrid encoding) to handle the 142.86% self-channel vs 214.29% enemy-AoE split.
4. Remove SP scaling from Volley entirely (ranged AP only).
5. Add a header comment to each SpellData file documenting the policy: "Coefficients store TBC canonical empirical values (post-AoE penalty, post-secondary penalty). DO NOT divide by 3 or 2 at runtime."
6. Add citation comments inline for non-obvious values (e.g. Hurricane's 0.9 slow penalty, Hellfire's split paths).

Out of scope

• AoE damage cap mechanic (patch 2.2.0+ separate system) - separate issue.
• Partial resist modeling - deferred per [Bug]: Expected damage overestimates actual; add diagnostic for missing-overlay spells #44.
• Engine downranking penalty - separate BLOCKER issue.

References

• WoWWiki Spell_power archive: https://wowwiki-archive.fandom.com/wiki/Spell_power
• WoWWiki Spell_power_coefficient archive (oldid=1549180): https://wowwiki-archive.fandom.com/wiki/Spell_power_coefficient?oldid=1549180
• Patch 2.2.0 patch notes (AoE coefficient adjustments)
• cMaNGOS-TBC SpellMgr.cpp:316-334 CalculateDefaultCoefficient

Acceptance criteria

• All AoE-flagged entries in SpellData match the canonical TBC table.
• SpellData file headers document the post-penalty storage convention.
• Hellfire split into self-channel and enemy-AoE paths.
• Volley no longer scales from SP.
• Cone of Cold and Consecration audited and corrected if needed.
• No engine code changes.