Fighter build: the whole loadout exists to feed one skill

✍️ Alice・愛麗沙

One skill, seven-tenths of the damage — and the sheet explains itself backwards

The Fighter is not a class where everything hurts. It is a class where one thing hurts enormously and everything else exists to set it up. Skill 5 holds seventy to eighty per cent of total output (Korean guide consensus), which means the loadout can be read in reverse — from the payload back to the delivery system.

• Hot Blood+ rewrites skill 5 — the somersault finisher, which our rune dictionary renders Summer Salt — into a frontal area attack. Giving an area shape to the skill that already owns most of your damage is why this is the first of the three accessories, not one of three equals.
• Advance+ rewrites skill 2, Thrust Kick, into a dash and raises the damage of the final combo skill and combo skills generally. It is not bought for skill 2’s own sake. It is bought to feed the combo that skill 5 finishes.
• Demolition+ rewrites skill 4, Burst Punch, so that a hit with no other enemy nearby deals a bonus hit. That is a boss-fight rune with the condition written on its face — in a room full of trash it is half a dead rune. Half of the Fighter’s “great at bosses, mediocre at packs” personality is literally printed on this one accessory.

The armour follows the same logic. Plains Wanderer (damage +12% and movement speed +30% for ten seconds after a movement skill) meshes exactly with the Fighter’s movement tag — once skill 2 becomes a dash, every combo triggers it without any management at all. And Rock Giant is gated on landing smash-judged hits, stacking attack to three and then converting the stack into a large short burst. Note what that means: it is not a generic armour rune here, it is a rune whose trigger condition happens to describe the Fighter’s damage profile precisely.

The sheet, slot by slot

This table is generated from the site database at build time. Full loadout: Fighter

Free-to-play and mid-budget columns are one click away in the loadout builder; the class’s positioning and ranking sit on the class page.

Two substitution rules. The weapon’s Duel+ (attack +18%, multi-hit damage −6%) is a genuinely useful bridge before Thousand Swords has stacked up — the Fighter’s damage is concentrated in single large hits, so it pays that multi-hit penalty more lightly than most classes do. And the emblem’s fallback is Goddess’s Grace (+25% damage for thirty seconds at the start of combat and after your ultimate): note that it is an abyss/raid rune and cannot be ascended. If it drops, good. Do not build a plan around getting it.

The combo is a manual transmission

Korea’s 2026-08-12 balance stream filed Monk, Fighter and Electric Mage into the same category: execution-dependent (Korean, community-relayed). Meaning there is a pair of hands between these classes’ paper numbers and their real output. For the Fighter, that gap has a specific shape. The Korean sheet’s rotation is 1·1 → 5 → 2 (or 3) → 5·3·5 → 1 → 4×3. The point is not to memorise the string, it is to see its structure: skill 5 appears three times. Every segment of the combo is engineered to put a cooled-down skill 5 back into the sequence, with skill 2’s dash and skill 3 filling the gaps between.

Auto-battle distributes skills roughly by cooldown order. It will not save skill 5 for the end of a combo, which is the whole reason the Fighter’s unattended output falls off so visibly. If you want a character to farm resources while you are away, do not force the Fighter into that role — switch to a class built for it (the Greatsword Warrior is the argument), farm there, and switch back when there is a boss. The levels you pick up along the way are feeding Thousand Swords anyway.

Thousand Swords on a class that punches

The recurring hesitation: “the Fighter wears knuckles — is Thousand Swords even relevant?” It is, and its situation here is simpler than it is for the sword classes. The rune grants +4% attack for each sword-using class you take to class level 40, stacking up to six times, and the bonus sits on the character, not on the weapon in your hands. The Fighter is not itself on the sword list (and that list is community inference, not an official roster), so it never has to work out whether it counts as one of its own stacks. Just level them.

Two practical notes. Taiwan has not opened the Warrior class yet, so the real current ceiling is five stacks, twenty per cent. And the accumulated levels you spend getting several classes to 40 are simultaneously feeding the versatility title’s 675-level threshold — every step on this path pays twice. The full argument, including the stack list, is in the Thousand Swords article.

One board observation that keeps getting lost: this bonus does not show up in your displayed combat power at all (player-measured). Do not conclude you levelled for nothing because the panel did not move.

Cards: WILL against LUCK, six hours of dummy each, a draw

The Fighter has a well-known community argument about its cards. The first two places are uncontested: the attack card first, the main-stat card second — and the main stat here is DEX, recorded with high confidence and not in dispute.

The third place is the fight. The sub-stat slot: WILL (which converts to raw attack) or LUCK (which converts to crit)? Korean players genuinely ran six hours on a training dummy for each and reported back that the results were near-identical — take whichever you like (Korean, player-measured; the original source URL was never pinned down, so we hold this at low-to-medium confidence).

We are keeping that as a draw, because it is one. Our advice is only that this is the least worthwhile slot in the entire build to agonise over — max your attack cards first and revisit it when there is nothing left to do. Defence, maximum HP and movement speed cards get skipped per the general damage-dealer rule.

Carving and sub-stats: smash all the way down

Gem tags run the movement and smash pair (two independent Korean sources agree, from launch-era sheets through to late ones; a third line goes to support or doubles down on smash). The carving priority is written in unusually extreme terms: smash (damage increase) far above movement (damage increase), far above hinder (cooldown) and smash (cooldown). Skill 5 is a heavy producer of smash judgements, which makes this ordering and “one skill carries seventy per cent” the same sentence written twice. Prism re-rolls follow the general rule — crit > multi-hit-enhancement and smash-enhancement > skill power — with crit taken past 8,000.

Now look at what the damage model actually says about those stats, because it makes the ordering above less obvious than it sounds:

Multi-hit, smash, AoE and skill power all share the same divisor, so per point invested they are worth exactly the same before any condition is checked. Which means the ranking between them is not decided by the numbers on that chart — it is decided entirely by which trigger gate your damage profile clears. The smash judgement wants one hit landing at roughly 40% of your panel attack power (a Korean-community threshold at medium confidence, not an official figure); the multi-hit judgement wants six hits inside three seconds. A class whose output is concentrated in one enormous finisher clears the first gate constantly and the second one rarely. That is the real reason smash outranks multi-hit here, and it is worth understanding rather than copying — the mechanics of those gates are their own article. Combo, meanwhile, sits on a divisor roughly twice as large as the others and pays in tiers on top of that: it is the trap line on the prism.

The transitional phase, and what to fuse first

Before graduation the armour principle inverts. Prefer runes that are unconditional or cheap to satisfy — Stability+ (attack +11% above 75% HP), Enlightenment (+6% at class level 40, another +4% at 65) — because a line that is always lit beats a higher-ceiling rune whose condition rarely fires. Mana Torrent is itself a low-bar rune (stacks attack simply for using skills), so equip it the moment you have it.

Fusion resources go in order of which rune changes how the class functions: Hot Blood+ first (it gives an area shape to the skill that owns most of your damage), Advance+ and Demolition+ next, then the armour core, and Thousand Swords last. Thousand Swords is last for a structural reason, not a value one — its stacks are bound to other classes’ levels, so no amount of fusion material can move it. Only levelling can.

Lookahead: Korea replaced skill 4 outright

Korea’s Season 2 change to the Fighter is a rebuild rather than a tuning pass: the new core is an Ascension-family combination, skill 4 is swapped for a new skill, and the rotation is rewritten end to end (Korean, medium-to-high confidence). The weapon and emblem runes have also moved off the Thousand Swords and Brilliance pattern that defines S0.

Two consequences for a Taiwan player. First, this page’s shelf life is S0 through S1 — copying Korea’s current sheet today will just produce a list of runes that do not exist in your client. Second, the direction is worth knowing in advance: Korea’s rebuilt Fighter kept the “one skill carries seventy per cent” structure. The skill changed names; the shape did not.

The execution-dependent label survives into Season 2 as well, so if hand-driven combos are what you enjoy, this class’s personality is not going to be patched away from you.

What we do not know

The card argument is the honest gap: the WILL-versus-LUCK test we are citing has no locatable source, and we hold it accordingly — if that bothers you, note that the finding it supports is “it does not matter much”, which is the least costly possible thing to be wrong about. The smash gate figure is a Korean-community measurement, not an official disclosure, and Taiwan has never published it. And the tier placements above are dated community opinion — note that the abyss data itself contains a direct contradiction, one rating calling the class the worst in the game for area damage and survivability, another re-rating it to the top tier because its DPS climbs as kill times shorten. Both are recorded. Neither is averaged away.

Our verdict

The Fighter is Taiwan S0’s clearest case of a manual-play premium: the build is cheap (all three accessory runes are the unique correct answer, so there is no build-branch tax), the entry bar is low, and the identical gear is two different classes in the hands of someone who plays the combo versus left on auto. Our recommendation: pick it as your main hand-driven class, not as your idle class. Lock mods 4 and 5 and stop deliberating over the third. Resource priority is level-30 on skill 5, then Thousand Swords, then everything else.

A claim you can overturn: on a Taiwan S0 single-target dummy, skill 5 accounts for at least half of the Fighter’s damage. Below half, the entire “feed skill 5” priority order on this page has to be re-sorted. Bring a measured damage-share screenshot and, if it holds, the article changes.