Head-to-head Dota 2 7.41f win rate from ranked matches.
In Dota 2 7.41f, Witch Doctor wins the Omniknight vs Witch Doctor matchup: Witch Doctor beats Omniknight in 53.03% of 17,561 ranked matches.
Trying to save an ally cursed by Witch Doctor is a frustrating experience for Omniknight. The core problem is Maledict; it can't be removed by Heavenly Grace, and its delayed damage burst completely ignores the physical immunity from Guardian Angel. This single spell undermines Omniknight's entire defensive game plan. While Guardian Angel is excellent for blocking a Death Ward, a smart Witch Doctor will ensure Maledict is already on the primary target first. All the healing and protection Omniknight provides often isn't enough to prevent the curse from securing a kill. It forces him into a position where his most powerful saves are made irrelevant before the real damage even lands.
How the matchup shifts across skill brackets — Omniknight's win rate against Witch Doctor in recent ranked matches, split by average match rank. Ranks with too few recent matches are omitted. These rows use the last 30 days of quality-filtered ranked matches, all patches in that span combined — a wider window than the patch-scoped count above, so the per-rank match totals can exceed it.
| Rank | Omniknight win rate | Matches |
|---|---|---|
| Herald & Guardian | 46.37% | 179 |
| Crusader & Archon | 46.56% | 610 |
| Legend & Ancient | 47.99% | 546 |
| Divine & Immortal | 49.53% | 107 |
Omniknight's win rate against Witch Doctor in each recent balance patch — whether this matchup is trending up or down. Patches without enough matches are omitted.
| Patch | Omniknight win rate | Change | Matches |
|---|---|---|---|
| 7.41f | 46.97% | ▲ +0.3pp | 17,561 |
| 7.41e | 46.68% | ▲ +0.1pp | 54,578 |
| 7.41d | 46.60% | ▲ +0.4pp | 70,213 |
| 7.41c | 46.22% | — | 29,369 |
Move this data into a draft
Omniknight starts on Radiant and Witch Doctor on Dire. Add the other eight picks to see whether this head-to-head edge survives the full draft.
Data updated Oct 5, 2026.