Nightreign Optimizer

Guides · All · patch 1.03.5 (balance 1.03.2) · updated 2026-11-01

Sparring Grounds: what to measure

Each test below settles a number the model now has to estimate. Do them on the training dummy, solo, with no relics unless a step says otherwise, and note the damage numbers the game shows. A photo or a short clip of the numbers is enough. Ordered by how much they move the rankings.

1. Defense on the dummy (all builds)

Why: every damage number in the model goes through one factor for the dummy’s (and the Nightlords’) defense 100. It comes from a 2025 Recluse test whose staff showed sorcery scaling 159, while the current decode gives 177 for the same character and level.

  1. Pick Recluse, level 15, with her starting staff. Open the status screen and write down the sorcery scaling the staff shows.
  2. Cast Glintstone Pebble on the dummy five times. Write down each number.
  3. Pick any hero, level 15, starting weapon. Write down the attack power the status screen shows, then hit the dummy with one light attack (the first of a chain) five times. Write down each number.

2. Halo Scythe Duchess (skill spam)

Why: the model makes this the strongest Duchess build, faster than any reported solo kill. If it is too high, the likely places are the share of time spent pressing the skill and whether each ring’s burst always hits.

  1. Duchess, level 15, the Halo Scythe at the rarity you found it, no relics. Write down its attack power.
  2. Use Miquella’s Ring of Light three times in a row on the dummy. For each press, write down every damage number that appears (the slash, the ring, and the ring’s burst if it is a separate number).
  3. Repeat with both Balancer relics worn (red and blue), then with only one. That tells whether two Improved Skill Attack Power lines multiply, add, or count once.

3. Restage

Why: the model replays 50% of the last 3 s, buildup included, but not a status proc’s own damage. Testers disagree on the proc part.

  1. Duchess with a frost weapon (Chilling Mist relic, or any frost weapon). Hit the dummy until frostbite procs, then Restage at once. Does a second frostbite damage number appear from the replay?
  2. Hit the dummy for exactly 2 s, wait 2 s, then Restage. Does the replay include the hits from 4 s ago? (Settles the 3 s window.)

4. Stamina (dagger Duchess)

Why: “Stamina Recovery upon Landing Attacks” is core for dagger Duchess, but the model has no stamina budget, so it is worth 0 there.

  1. Duchess, level 15, her starting dagger, no relics. From full stamina, do one-handed R1 chains without stopping. Count how many full chains before the bar is empty.
  2. Time how long the bar takes to refill from empty when you stand still.
  3. Repeat step 1 with Stamina Recovery upon Landing Attacks worn.

5. Executor’s Cursed Sword

Why: the model assumes the sword is out 70% of the fight and the unlocked skill comes twice a minute.

  1. Executor, level 15, no relics. Draw the Cursed Sword. Does it stay out until you sheathe it, or does it time out? If it times out, how long does it last, and how long until you can draw it again?
  2. Count how many deflects it takes before the unlocked skill is available (the decode says 50 points: a deflect is 15, a knockback deflect 20).
  3. Hit the dummy with the unlocked skill: write down the number. Then the same with [Executor] Character Skill Boosts Attack worn (the model says ×1.35).

6. Crits on a katana

Why: with four crit lines (Improved Critical Hits three times, +1 once) the model gives ×2.02 on every crit. Players measured ×2.02 on a dagger but ×1.83–1.86 on katanas, so only part of a katana crit may take the boost.

  1. Executor, a katana, no relics: riposte the dummy and write down both numbers of the crit if two appear.
  2. Repeat with one Improved Critical Hits relic.

7. Blood Blade on Duchess

Why: the model times the three presses at 1.3, 0.73 and 0.8 s and gives the blood wave Arcane scaling only.

  1. Duchess, level 15, Blood Blade on her starting dagger, no relics. Press it three times in a row on the dummy and write down each slash and each wave number.

8. Nullify Attacks

Why: “Occasionally Nullify Attacks When Damage Negation is Lowered” (the red Balancer) has no chance in the game data, so the model gives it 0.

  1. Wear the red Balancer and anything that lowers your damage negation (the [Executor] Cursed Sword relic does when you attack). Let a camp enemy hit you 20 times. Count how many hits were nullified (no damage).

9. Wylder: Art gauge from Claw Shot

Why: “[Wylder] Art gauge greatly filled when ability is activated” is the model’s top Wylder line (+14%): decoded, each Claw Shot adds 18 gauge points (SpEffect 7032402), about six Claw Shots for a full gauge. One player called it not useful.

  1. Wylder, level 15, wear only that relic. Empty the gauge (use the ult), then Claw Shot the dummy, waiting out the cooldown, without any other attack. Count the Claw Shots until the gauge is full.
  2. Repeat once without the relic: count light attacks until full (that is the normal gain).

10. Wylder: Claw Shot damage

Why: Claw Shot is half of Wylder’s damage in the model: the throw (100% motion value) and, with a greatsword, the follow-up slam (155%). The dummy is small enough to pull, so there you get the throw and the pull (80%) instead.

  1. Wylder, level 15, his starting greatsword, no relics. Write down its attack power.
  2. Claw Shot the dummy five times. Write down every number (the throw and the pull).
  3. If you can, Claw Shot a large enemy (a field boss or a Nightlord) once and note the numbers: throw and the slam.