Flat +450 attack or +1.5% attack? One division decides it
✍️ Alice・愛麗沙
One division settles it. Divide a flat value by your current attack; divide a percentage by your attack% bucket total (1 plus whatever percentage you already carry). The larger quotient wins. On a character with 6,000 attack and +80% attack already stacked, the legendary fashion 4-piece set effect of +450 flat attack is worth 450 ÷ 6000 = +7.50%, while an accessory scroll’s +1.5% attack is worth 1.5 ÷ 180 = +0.83%. That is nine to one. The crossover sits at 54,000 attack — and even swapping in a +170 armour scroll, it is still 20,400. Below that, flat wins every time. Not because flat values are inherently better. Because the percentage bucket is already crowded and the flat bucket is not.
Updated 9/14: Taiwan’s enchant system went live on 9/9 (update log thread 3541097). The published extra-effect values are all flat: STR/DEX/INT/LUK/WIL 17 each; break, damage reduction, smash/quick/chain/multi-hit enhancement, skill power, skill speed, AoE enhancement, vital-hit evasion, crit, pursuit, recovery and ultimate 22 each; bonus HP 67 — every one of them lands on the A-bucket side of this article. The scrolls’ base effects (the +170 / +1.5% used here) were not listed in the log and remain Korean figures; Taiwan’s first scrolls come from the new Mas deep dungeon (瑪斯深層地下城) drops and, from 9/14, Molly’s 附魔卷軸：狐狸(★6) for 400 赤紅深淵的化石. Read the scroll tooltip, then run the division.
Corrected 9/28: This page used to take +175 as Taiwan’s legendary fashion 4-piece attack, based on our own 8/14 capture of the Cosmic Starchild explainer. Nexon Taiwan’s three current notices (that explainer, the 8/26 end-of-sale notice and the 9/9 Elderwood Sovereign introduction) all say +450 (with all stats +220 for the 3-piece). The current notice no longer shows +175 and no update log announced a change, so we cannot tell whether Taiwan changed the value or the capture was wrong. Every calculation below is redone with +450, with +175 kept for comparison; the conclusion does not flip — it gets stronger.
Two buckets, not one scale
“+450 attack” and “+1.5% attack” read like two spellings of the same thing. Inside the formula they are nowhere near each other.
- Flat attack falls into bucket A, where it adds together with character attack, weapon base, engraving, necklace, pets, the fashion set option and runeword flat attack. All of it, summed.
- Attack% falls into bucket B, where it adds together with the percentages from gear, skills, enchants and artifacts — and then that whole sum multiplies the finished A.
final attack contribution = A × B
Which means the two are multiplied together, not competing. They only need comparing when they are fighting over the same slot: one scroll, one fashion piece, one enchant line. And the comparison is a marginal one:
- adding flat x:
x ÷ A - adding percentage p:
p ÷ (1 + the percentage you already have)
That is the entire method. There is no ranking to memorise — two quotients, three seconds.
Four sets of numbers
A character one lap in: A = 6,000, bucket B already at +80%
| Option | Arithmetic | Real gain |
|---|---|---|
| Fashion set, attack +450 (Taiwan notices, current) | 450 ÷ 6000 | +7.50% |
| Fashion set, attack +175 (our 8/14 capture) | 175 ÷ 6000 | +2.92% |
| Armour scroll, attack +170 | 170 ÷ 6000 | +2.83% |
| Accessory scroll, attack +1.5% | 1.5 ÷ 180 | +0.83% |
| Weapon scroll, attack +1% | 1.0 ÷ 180 | +0.56% |
A character just starting: A = 3,000, bucket B already at +30%
| Option | Arithmetic | Real gain |
|---|---|---|
| Fashion set, attack +450 | 450 ÷ 3000 | +15.00% |
| Fashion set, attack +175 (for comparison) | 175 ÷ 3000 | +5.83% |
| Accessory scroll, attack +1.5% | 1.5 ÷ 130 | +1.15% |
Flat wins on both characters, and it wins by more on the weaker one — the opposite of the usual instinct that percentages are the “late-game” option. The conclusion does not lean on the fashion number either: even the +170 armour scroll beats the accessory scroll three and a half to one at A = 6,000.
Where the crossover is
Set the two sides equal: x ÷ A = p ÷ (1 + B), then solve for A:
A = x × (1 + B) ÷ p
With B = 0.80 and p = 1.5%, try three values of x:
- x = 450 (fashion 4-piece, Taiwan notices, current):
450 × 1.8 ÷ 0.015 = 54,000 - x = 175 (our 8/14 capture):
175 × 1.8 ÷ 0.015 = 21,000 - x = 170 (armour scroll):
170 × 1.8 ÷ 0.015 = 20,400
Attack has to climb to 54,000 before that 1.5% starts beating the fashion set’s 450. The “21,000” this page used to quote is now, near enough, the armour scroll’s crossover (20,400). Taiwan’s current gear depth is nowhere near either number.
Read the equation the other way and it tells you something more useful still: the higher you stack attack%, the less a percentage entry is worth and the more a flat entry is worth. Most players’ intuition runs exactly backwards on this.
+175 or +450: Taiwan’s notices now say +450
This is the most-argued number in the whole fashion system, and this site got it wrong for a while too:
| Rarity | Pieces | Period | Set Effect | Confidence |
|---|---|---|---|---|
| Legendary | 3-piece | S0 (KR launch)（2025-03-27） | All Stats +90 | Confident |
| Legendary | 4-piece | S0 (KR launch)（2025-03-27） | Attack +175, Defence +175 | Confident |
| Legendary | 3-piece | S1 Buff #1 (KR)（2025-09-25） | All Stats +155 | Confident |
| Legendary | 4-piece | S1 Buff #1 (KR)（2025-09-25） | Attack +310, Defence +310 | Confident |
| Legendary | 3-piece | S2 Buff #2 (KR-current)（2026-06-25） | All Stats +220 | Confident |
| Legendary | 4-piece | S2 Buff #2 (KR-current)（2026-06-25） | Attack +450, Defence +450；Final Damage +3%（Generation 6 only，not retroactive to existing sets） | Confident |
| Legendary | 3-piece | TW 3504848 as captured 08-14（2026-07-22） | All Stats +90 | Unconfirmed |
| Legendary | 4-piece | TW 3504848 as captured 08-14（2026-07-22） | Attack +175, Defence +175 | Unconfirmed |
| Legendary | 3-piece | TW official, current（2026-07-22） | All Stats +220 | Confident |
| Legendary | 4-piece | TW official, current（2026-07-22） | Attack +450, Defence +450 | Confident |
| Epic | 3-piece | S0 (KR launch)（2025-03-27） | All Stats +20 | Confident |
| Epic | 4-piece | S0 (KR launch)（2025-03-27） | Attack +35, Defence +35 | Confident |
| Epic | 3-piece | S1 Buff #1 (KR)（2025-09-25） | All Stats +20；presumed unchanged — not independently confirmed for this window | Unconfirmed |
| Epic | 4-piece | S1 Buff #1 (KR)（2025-09-25） | Attack +35, Defence +35；presumed unchanged — not independently confirmed for this window | Unconfirmed |
| Epic | 3-piece | S2 Buff #2 (KR-current)（2026-06-25） | All Stats +45 | Confident |
| Epic | 4-piece | S2 Buff #2 (KR-current)（2026-06-25） | Attack +90, Defence +90 | Confident |
The full timeline of legendary and epic fashion set effects: each row is the value for one (rarity, piece-count, version) combination, with its effective date and confidence. Checked 2026-09-28. The TW rows: three Nexon TW notices — the Cosmic Starchild page 3504848 (current text), the Cosmic Starchild end-of-sale notice 3531886 and the Elderwood Sovereign introduction 3541066 — all state the legendary 3-piece at all stats +220 and the 4-piece at attack and defence +450 each. Our 08-14 capture of 3504848 recorded +90 / +175; the current text no longer shows those numbers and no update log announced a change, so both rows are shown, each with its basis. The S0 / Buff #1 / Buff #2 rows are Korea's history. No TW notice states epic set values or the generation-6 final damage +3%.Cash Catalogue
Korea’s history runs +175 at launch → +310 on 2025-09-25 → +450 on 2026-06-25, both buffs retroactive. On Taiwan, all three current Nexon Taiwan notices say +450; the “Taiwan runs +175” this page used to state came from a single 8/14 capture that the current notice no longer matches. Both rows sit in the table with their own confidence, and the set-effect ladder explains every row’s source.
Run the same division across the correction. On a character at A = 6,000, +175 is +2.92% and +450 is +7.50% — 4.6 percentage points in one slot. The old version put “if Taiwan catches up to Korea, this slot becomes the most valuable flat source” in the future tense; on Taiwan’s own notices it is the present tense. If you are one piece short of a 4-piece set, the +450 that fourth piece unlocks is more than two and a half times a +170 armour scroll.
To confirm which number your client runs: wear all four pieces, then take one off. A single fashion piece carries no attack of its own, so the panel drop is the 4-piece attack bonus. The two candidates are two and a half times apart — you will see at a glance whether it is the 450 tier or the 175 tier.
Three judgement errors to avoid
One: do not read the size of the printed number. “+1.5%” sounds more advanced than “+170” and loses by more than three to one in most configurations — and nine to one against the fashion set’s +450. A percentage entry is worth what the crowding of its bucket says it is worth, and nothing about the digits on the tooltip tells you that.
Two: do not put an emblem’s attack% into this comparison. Emblem attack% is not in bucket B. It multiplies inside the weapon sub-bracket within A — a genuinely different calculation, and the division above does not apply to it.
Three: do not price the weapon scroll’s damage-over-time. Korea’s Season 0 weapon scroll carries, alongside its +1% attack, a damage-over-time proc at 35% chance on a 12-second cooldown — and its coefficient has never been published. No coefficient, no valuation; we are not inventing one. It might carry that scroll past the flat options and it might not. We do not know, and that is what the page says.
What to do now
- Run the division at every fork. No ranking to memorise, just two quotients.
- Load up on flat attack early. Pets, the fashion set option, necklaces, flat armour enchants — every point is expensive while A is small.
- Do not rush percentage entries before A has grown. They are multipliers, and multiplying a small base is not worth much.
- Do not treat the two as enemies. A and B multiply, so you want both in the end. This division is only for the moment when one slot can hold one of them.
What would overturn this
The A/B bucket boundary, and the claim that flat sources all add together, come from the Korean community’s long-maintained damage formula — Nexon has never published a combat calculation on any server. We label that segment high confidence. Taiwan’s +450 is the figure in three current Nexon Taiwan notices; +175 survives only in our own 8/14 capture and is labelled low confidence. If Taiwan’s panel turns out to show only +175 for the 4-piece, swap every calculation above back to the +175 row — the crossover returns to 21,000, and “flat keeps winning” still holds.
If somebody measures on Taiwan that a fashion set’s flat attack does not enter the attack figure on the stat panel, or that attack% sources multiply each other rather than adding, the division on this page has to be replaced. The cheapest test is the most direct one: take off a single flat-attack source and check whether the stat panel drops by exactly that number.