Skill Speed Doesn't Touch Cooldown — Three 'Faster' Stats and Two Separate Cooldown Numbers

✍️ Alice・愛麗沙

Get the terms straight first, because three of them mean “faster”

Nexon’s official stat page (guide thread 3474907, fetched directly from the Taiwan help corpus) lists five primary stats — Strength, Skill, Intelligence, Luck, and Will — each feeding exactly three sub-stats. Two of those fifteen sub-stats both translate loosely to “speed,” and that overlap is where the confusion starts.

Attack Speed is fed by the primary stat the game itself officially calls “Skill” — an unfortunate naming collision, since that’s also the generic English word for an ability, which is exactly the kind of overlap this article exists to untangle. Attack Speed governs only the interval between basic attacks. Gear and enhancement lines labelled “attack speed %” multiply against this base value.

Skill Speed is fed by Intelligence. It governs actual abilities, and — this is the part the community had to test for itself, since Nexon’s guide states the function but not the number — it accelerates both halves of a skill’s animation, not just one.

Beyond those two official sub-stats, two more effects show up on specific runes and class passives: Cast/Charge Speed (from the Supernatural rune and one of the Greatsword passives) and Skill Execution Speed (from the Storm rune). Neither is a primary-stat sub-stat at all — they’re rune/passive-specific effects, and each one only speeds up half of a skill’s animation.

What Nexon confirmed, and what the community had to measure itself

The "Official" column only confirms the stat exists and what it's defined to do — the exact speed-up percentages, the Storm-rune conversion example, and the cooldown-vs-cooldown split all come from KR/TW community testing, not a Nexon-published number. Full explanation in the body text.

Nexon’s official guide states two things plainly: Attack Speed exists and is Skill-stat-fed, with a stated function (shortens basic-attack intervals); Skill Speed exists and is INT-fed, with a stated function (speeds up a skill’s windup and release). What Nexon never publishes is a formula — no percentage-per-point number for either stat appears anywhere in the official corpus. Everything below the “official” row in that table — the windup/release split, the two rune-sourced effects, and the whole cooldown side — comes from Taiwan and Korean community testing, not from Nexon.

A skill has two separate phases, and only one stat speeds up both

The community research behind this article — a Taiwan board post running frame-perfect 120fps captures across five different skills — established a mechanical fact Nexon’s guide doesn’t spell out: every skill in this game internally splits into a windup phase (charging, the pre-swing delay) and a release phase (the actual hit).

Skill Speed is the only stat that accelerates both. The two rune-sourced effects each cover exactly half: Cast/Charge Speed only shortens the windup — the moment the hit actually lands is untouched — while Skill Execution Speed does the opposite, shortening only the release while the windup stays exactly as slow as before. Practically, a player who only sockets one of the two rune effects is only getting half the “faster” they think they bought — almost certainly the mechanical reason behind the recurring community complaint of “I socketed a speed rune and barely felt a difference.” They bought half a skill’s worth of acceleration and expected the whole thing.

Channel-style skills (the kind that need to be held down or re-triggered continuously) behave differently again: speeding them up raises the tick frequency, not the total number of ticks — more repetitions in the same window, not a shorter total animation.

On magnitude: the Taiwan board’s 120fps measurement put Skill Speed at roughly +1000 stat points ≈ 8–10% faster, with diminishing returns — shorter skill animations have less room to compress, so the same +1000 buys less visible speed on a fast skill than a slow one. One community-derived reference point for scale: an Epic-rarity Storm rune’s +32% execution speed converts, by a formula the Taiwan post borrowed from a Korean wiki, to roughly 3,000–5,000 points of Skill Speed — a player-reverse-engineered estimate, not an official conversion, labelled as such here on purpose.

Skill Speed never touches cooldown — it is a fully separate system

The claim “stacking Skill Speed reduces cooldown” circulates on the community board, and it’s wrong in a specific, checkable way: cooldown itself splits into two independent numbers inside the game’s own systems — a flat cooldown reduction (a fixed percentage subtracted from the total cooldown) and a cooldown recovery speed (the remaining-time timer itself ticking down faster, an effect that matters more the longer a skill’s base cooldown already is). Korean players researching their own version of this exact question note that even those two get confused with each other — but on the point that matters here, every source agrees: Skill Speed affects neither. Stacking it will never make a skill available again sooner. It only makes the windup-to-release motion itself play out faster once the button is pressed.

The practical build implication: if the goal is “make my skill rotation feel faster,” the first question to answer is which problem is actually being solved — a slow animation (Skill Speed, or the half-effects) or a long wait before the button lights up again (cooldown). They’re different stats solving different problems, and resources spent on the wrong one buy nothing.

One line to close on

Three different “faster” stats each own their own lane — Attack Speed governs basic attacks, Skill Speed governs both halves of a skill’s animation, and the rune/passive-sourced speed effects usually cover only one half each. Cooldown is a fourth, entirely unrelated system that Skill Speed never reaches. This classification is falsifiable on its own terms: if Nexon ever publishes an official Skill Speed formula, or a case surfaces that contradicts this split, this article gets rewritten.