DPS rankings
Every specialisation simulated on its best-in-slot profile at 1, 2, 3, 5 targets. Switch the target count — the order changes, and that is the whole point.
1. Outlaw Rogue — 188,981 DPS
2. Subtlety Rogue — 187,191 DPS
3. Arms Warrior — 185,491 DPS
25 specs · top 188,981 DPS · median 172,631 · spread 189%
25 specs · top 268,459 DPS · median 217,280 · spread 153%
25 specs · top 330,571 DPS · median 264,202 · spread 145%
25 specs · top 492,277 DPS · median 350,885 · spread 175%
How to read this
These are five-minute patchwerk sims on each spec's best-in-slot loadout — the ceiling a spec reaches when everything goes right and nothing moves. They are not a tier list. Real fights have movement, priority targets, downtime and mechanics that punish some kits far more than others, and none of that is modelled here.
Augmentation Evoker is last here, and that means nothing
Augmentation is a support specialisation. Almost all of its contribution is damage it adds to other players, and a solo patchwerk sim measures exactly none of that. Its own damage being lowest on this chart is a fact about the measurement, not about the spec. It is greyed out and marked above for that reason. Do not read this chart as a reason to play or not play it.
Why some specs dip at 3 targets and recover at 5
Every spec here runs its single-target talent build at every target count — the same build the reference player logged. So the multi-target columns answer "what does this build do when more enemies show up", not "what does this spec do when specced for AoE".
That produces some genuinely non-monotonic curves. Arms Warrior scores higher at 2 targets than at 3, because its cleave doubles perfectly on two and then stops scaling while the rotation shifts to weaker filler. It is not a bug — we verified it by re-running the same target counts as separate sims with no profilesets involved, and the numbers agreed to 0.05%. A spec that would re-talent for AoE will do better than its column here suggests.
Treat gaps under about 3% as noise. The sims converge to a 0.05% error, but the underlying profiles come from one logged player per spec, and a different player's gear or talents would move a spec further than that. Use this to see the shape — who scales into multi-target, who falls off — not to pick between two neighbours.
Simulated on PTR data for Midnight Season 2. Per-spec detail is on the spec pages.
Gear is the part you control
A spec three percent behind on this chart is worth far less than three item levels you did not plan for. GearPath maps the whole season — crests, vault choices, sparks — so the gear side is not where you lose.