Frost Mage gem tags: Elemental covers almost all of it
✍️ Alice・愛麗沙
The Frost Mage has the cleanest tag problem in the game. Five of its six actives carry Elemental, and the meta rune rewrites the sixth — skill 2 — so the class effectively runs six for six. So line one is a formality. The real call is line two, and it is Summon, not Disrupt: the three buttons producing almost all your damage — skill 4, skill 5, the ultimate — are all tagged Summon, while Disrupt hangs off one low-output skill the meta rune has already rewritten. Line three is Survival, which lands on skill 4 and the ultimate again. Take damage on all three; cooldown loses by roughly threefold even when flattered. And the rule that saves the most: never spend a Gem Skill Engraver on a crude-tier gem, because promotion re-randomises the tag set.
Coverage decides the order, not opinion
A tag line pays out only on skills carrying that tag, so “+2.0% Elemental damage” is not worth 2.0% to you — it is worth 2.0% times the share of your damage carrying Elemental. Write that multiplication down and the priority order stops being a matter of taste.
| Tag | Rotation skills carrying it | Share of output | One damage line, on total output | One cooldown line, best case |
|---|---|---|---|---|
| Elemental | 5 / 6 | 95% | +1.90% | −0.62% |
| Summon | 3 / 6 | 78% | +1.56% | −0.51% |
| Survival | 3 / 6 | 45% | +0.90% | −0.29% |
| Smash | 2 / 6 | 25% | +0.50% | −0.16% |
| Disrupt | 1 / 6 | 8% | +0.16% | −0.05% |
| Support | 1 / 6 | 6% | +0.12% | −0.04% |
Coverage is compiled from Korean skill-tag listings and rotation shares, not an official disclosure. The two right-hand columns scale one Elite Star Prism line (+2.0% damage / −0.65% cooldown) by that share; the cooldown column assumes the seconds saved convert fully into extra casts, which they do not — read it as a ceiling, not an expectation.
The two right-hand columns are the whole argument. Elemental sits at 95% of output, so one line converts to +1.90% on your total damage — almost no leakage. Summon converts to +1.56%. Then the floor drops out: Disrupt is 8% coverage and +0.16% per line, an order of magnitude below the top row.
Why the second line is Summon and not Disrupt
“Mages want Elemental first, then Disrupt” works for the base mage and fails here, for a reason worth understanding rather than memorising.
Disrupt appears on exactly one Frost Mage skill — skill 2 — and skill 2 is the one the meta rune rewrites. Run the meta build and Disrupt sits on a button contributing very little. Skip it and that same button doesn’t even carry Elemental. There is no configuration in which the Disrupt line looks good.
Summon is the class’s real damage tag, by an accident of categorisation. Skill 4 is a ground field, skill 5 a placed object, the ultimate a large placed zone; all three are judged summons, and all three are where both mainstream layouts put their output. Summon here is not a niche pick for a summon build — it is the main damage tag, wearing a misleading name.
Survival takes line three by the same logic: it lands on skill 4 and the ultimate, the two biggest hitters again. Not a defensive slot, a second coat of paint on what already matters.
One damage line against one cooldown line
Every tag line forces a binary choice. On an Elite Star Prism it is +2.0% damage or −0.65% cooldown per line; the ladder runs Crude S 1.7/0.55, Elite 2.0/0.65, S 2.1/0.7, Whole 2.2/0.75, and Triple Quartz 2.1/0.7 across two lines instead of three.
Give cooldown the most generous assumption available: pretend every saved second converts perfectly into an extra cast. That buys about +0.65% more casts, on the same tagged subset the damage line would have covered. Against +2.0%, damage wins by roughly three to one.
Reality is worse: the game has pronounced pre- and post-cast delays, and this rotation is gated by animation time, so the fraction you shave lands where the skill wasn’t ready anyway. Take damage on all three. A cooldown line is only ever a placeholder — socket an Elemental cooldown roll you already own, never carve one.
Three things people get wrong
The “separate multiplier” does not mean what you think. Gem tags do occupy their own bucket, multiplied against attack power, damage-increase and crit rather than folded into them. But inside that bucket the lines are additive with each other: every Elemental damage line sums into one figure and multiplies in once. So “each line compounds” is wrong; “one percent of Elemental damage is roughly one percent of total damage, scaled by coverage” is right. Twenty-two sockets, three lines each, all one tag, ceilings at +132% — a sum, not an exponential.
Skill-use speed is not cooldown reduction. This is the expensive misunderstanding, because it tempts people to carve cooldown lines as a substitute for it. Fast-skill effects act on the preparation phase — casting, charging — and the use phase — animation, channelling. Neither formula contains a cooldown term. Frost Mages fall for it more than most, because animation time genuinely is their bottleneck, so speed feels enormous and they reason backwards that cooldown must too. It does not. The comparison in detail.
Runes rewrite the tags your skills carry. The meta rune changes skill 2 from Smash·Disrupt to Elemental·Disrupt, which puts your gem plan downstream of your rune plan. Decide the rune first. And the “smash enhancement” printed on runes belongs to a different system from skill tags.
Why crude gems are a trap, and it isn’t because they’re weak
The reason is mechanical: promotion re-randomises the tag set. When three gems fuse into one, the result rolls fresh tags — unless you spend an additional 5 Gem Skill Engravers at promotion to carry the first-slotted gem’s tags onto the new one.
So the conclusion is hard-edged: any engraver spent on a gem that is not your final gem gets wiped at the next promotion. Leave crude and Crude S alone, fill all twenty-two sockets by volume, and release engravers only for tag carry-over. Taiwan’s reliable supply is the adventurer pass at ranks 23 and 24 — five per rank on the free track, ten on the paid. That is the whole budget.
Cheaper still: fresh Star Prisms roll random tags anyway, so fuse several and socket the best roll. And the re-roll odds are officially just “random” — no distribution has ever been published. We checked namu wiki, Inven, arca, dcinside and the Korean gem simulators. Treat any quoted percentage as invented.
Korean-server lookahead: the tag swap makes bad rolls valuable
Korea added a tag swap on 2026-04-23, and almost every summary of it in circulation is wrong.
It is not two gems trading tags with each other. You choose two tag types, and every gem on your account — stored and socketed alike — has those two exchanged. Pairs compose into a single permutation within one operation, so three-way cycles are reachable, but the primitive you are buying is still a pair. The price is 1,000,000 gold per permutation; if your damage lines and cooldown lines need different permutations, that is two operations and 2,000,000 gold. One hard limit: damage lines swap only with damage lines, cooldown only with cooldown.
Access comes from a one-time quest given by a bank NPC in one Korean-server town, and afterwards any town bank runs it. Korean-server only; not shipped on Taiwan, and no Taiwanese announcement has promised it. The consequence: a gem with three matched lines of the wrong tag stops being junk and becomes inventory. Do not break those down and do not re-roll them.
What to do now, what to ignore, what would change this
Keep, in order: any Elemental line, regardless of what sits beside it. Summon second — it is where skills 4, 5 and the ultimate live. Survival third, because it doubles up on skill 4 and the ultimate. Smash only rises if you are not running the meta rune, in which case skill 2 keeps its original Smash and joins skill 1. Ignore Disrupt and Support entirely.
That ordering assumes the two common engraving layouts. If your modification slots differ, count your own tags in the skill panel first — the arithmetic is general, the coverage numbers are not.
Three things would invalidate this page. If Taiwan’s meta rune does not convert skill 2 to Elemental, the coverage table needs recomputing. If cooldown conversion measures far higher on Taiwan than we assume — meaning the class is cooldown-gated rather than animation-gated — line three becomes Elemental cooldown. And if Taiwan opens a steady engraver supply, or ships the tag swap early, hoarding purely for promotion stops being right.
