Invocation planner

This is a deterministic melee plan, not average DPS. Auto spends available energy on DDS specials when both hits can land, then uses your normal weapon. Per-hit choices override that plan. Normal attacks spend no special energy; an explicit DDS special spends its full cost even if its second hitsplat is lost. An unaffordable DDS special falls back to a normal dagger attack and is flagged. It assumes uninterrupted attacks, constant boosts and prayer, instantaneous gear switches and adrenaline lasting for every down when enabled. DDS uses aggressive style.

Exposure depends on HP when the core is ejected: at least 80% gives 21 ticks, 60–80% gives 29, and below 60% gives 37. The first hit occurs at your delay, with subsequent hits spaced by weapon speed. Changing to BGS for the last hit is supported: its six-tick cooldown delays only the following attack, not the current hit. The second DDS hitsplat arrives one tick after the first; if it reaches the retraction boundary it deals no damage.

Core HP scaling uses the raid level rounded down to ten-level steps, party scaling, and final HP rounding to the nearest ten. Path levels do not scale Warden core HP. Timer penalties affect loot after the raid, so the original selected level is used here.

Supported normal weapons use their standard strength bonus and attack speed. Fang applies its 85% maximum. Voidwaker is modeled as normal melee only: its special is magic damage and is not treated as a guaranteed maximum melee hit. Other weapon passives, special weapons, unusual sets and external effects require Custom melee weapon with a manually verified maximum. Soulreaper normal hits add 6% of boosted Strength per existing stack, separately from prayer; a stack is gained after damage, up to five. Behead removes the stack Strength bonus and instead multiplies the unstacked max hit by 1 + 6% per consumed stack. It costs zero special energy. Its accuracy/minimum-hit bonuses do not improve the core’s guaranteed maximum. Starting stacks are explicit per down because intervening attacks and downtime are unknown. Stacks decay after 50 idle ticks, longer than any single modeled exposure; assume the entered starting stacks are fresh. Team estimates assume identical attackers.