The Hong Kong Massacre

Both lean on execution about the same (8 vs 8). Biggest difference: optimization (5 vs 6) - Route planning and mask choice per floor

Both lean on execution about the same (8 vs 8). Biggest difference: variance (2 vs 1) - Rooms are identical on every attempt

Both lean on execution about the same (8 vs 8). Biggest difference: variance (2 vs 1) - Spawns are nearly deterministic once a stage is learned

Both lean on reads about the same (0 vs 0). Biggest difference: execution (8 vs 7) - first-person melee, kicks, and physics timing

Both lean on reads about the same (0 vs 0). Biggest difference: execution (8 vs 7) - cloudstep chains demand precise aerial control

Both lean on reads about the same (0 vs 0). Biggest difference: execution (8 vs 7) - Precise platforming jumps and sword-timing are demanding.

Both lean on execution about the same (8 vs 8). Biggest difference: variance (2 vs 1) - Everything is placed and scripted

Both lean on execution about the same (8 vs 8). Biggest difference: reads (0 vs 1) - Rival AI is non-adaptive

Both lean on reads about the same (0 vs 0). Biggest difference: execution (8 vs 7) - Chicken-form gauntlets push precision platforming further

Both lean on reads about the same (0 vs 0). Biggest difference: execution (8 vs 7) - Tank aim and dodge timing under pressure decides every stage

Both lean on execution about the same (8 vs 8). Biggest difference: optimization (5 vs 6) - Keeper rescue sequencing and phase timing are fully learnable

Both lean on execution about the same (8 vs 8). Biggest difference: variance (2 vs 1) - Nothing random once spawn patterns are memorised