Champion sheet · Support

Mel

Patches 16.19 to 16.21 · All regions

Tournament presence · pick + ban

10.0%#82 of 173 champions

Pick: 5.3%picked in 1,496 of 28,062 games
Ban: 4.7%1,074 bans over 22,823 games, against 5.2% in soloQ

Win rate as Support38.6% between 32.6% and 44.9% over 166 gamesSoloQ 42.9%: no clear gap

Median length29.5 minsoloQ 28 min

Amateur: 28,304 games from amateur tournaments, last 60 days, KR 65% · EUW 25% · NA 10%. SoloQ Master+: 919,797 games on the same patches. An amateur figure in gold clearly differs from soloQ.

At a glance

Four answers, each one opens its chapter.

What changes in tournaments3 clear gaps between soloQ and amateur tournaments as Support.DraftLength
How to play itArcane Comet, then Luden's Echo, Shadowflame, and Zhonya's Hourglass: the most played build in soloQ.
In a compTeam damage first: more damage than 86% of supports.DuelistDamagePoke
With / againstNo opponent with a clear gap to expected in amateur play: win rates remain ranges.
In draftPriority pick: taken in G1, burned afterwards.

01 — What changes in tournaments

What changes in tournaments?

3 clear gaps between soloQ and amateur tournaments as Support.

DraftPicked more often in amateur play
Amateur
0.2%
SoloQ
0.1%
0.2% of the role's picks in amateur play, against 0.1% in soloQ.
DraftBanned less often in amateur play
Amateur
4.4%
SoloQ
5.2%
4.4% ban rate in amateur play over 14,779 games, against 5.2% in soloQ over 387,203 games.
LengthLonger games in amateur play
Amateur
29.5 min
SoloQ
28 min
Median of 29.5 min in amateur play, against 28 min in soloQ.

Not enough games to judge: core.

02 — How to play it

How to play it?

Arcane Comet, then Luden's Echo, Shadowflame, and Zhonya's Hourglass: the most played build in soloQ.

SoloQ build, amateur differences under each block

RunesArcane Comet82% · 2,163

Same in amateur play

SpellsFlash + Heal78% · 2,038

Same in amateur play

SkillsQnot collected in soloQ

Amateur: 93% · 50 games

Core (3 legendaries)17% · 84

17%of soloQ games

Luden's Echo
Shadowflame
Zhonya's Hourglass

Same in amateur play

Not collected in soloQ: starting items, 4th item, and skill order.

03 — In a comp

What does it bring to a comp?

Team damage first: more damage than 86% of supports.

Team damagetop14%15% · supports: 9%
Takes its kills in duelstop20%23% · supports: 17%

Read on tournament games from 2026-04-15 to 2026-10-12 (44,638 games) · same reading as the draft review · compared with other supports

What marks it

Duelist
Damage
Poke

Like other supports

Absorbs damage
Plays teamfights
Crowd control (CC)

What it is not

Little map pressure
Carries its own kills

Damage / min by phase · at equal gold

Rank 87 in lane, 87 in mid game and 89 in late game, out of 100 supports.

398+110 vs 288

564+222 vs 342

678+290 vs 388

Lane · 5–14 minMid game · 15–22Late · 23–30

Melsupports

Gold vs. direct opponent

At 20 min, ahead of 83% of supports against the direct opponent.

supports

4.6takedowns outside its lane before 15 min; supports: 4.5.

Where it stands among supports · every measure

0 ← less · median · more → 100

Damage in lane · rank 87 of 100 among supports398 · median 288
Damage in mid game · rank 87 of 100 among supports564 · median 342
Damage in late game · rank 89 of 100 among supports678 · median 388
Gold at 15 min · rank 81 of 100 among supports+327 · median +105
Gold at 20 min · rank 83 of 100 among supports+497 · median +193
Map pressure · rank 17 of 100 among supports4.6 · median 4.5
Absorbs damage · rank 27 of 100 among supports278 · median 393
Plays teamfights · rank 29 of 100 among supports51% · median 54%
Takes its kills in duels · rank 80 of 100 among supports23% · median 17%
Lets allies carry kills · rank 14 of 100 among supportsrank 14 of 100
Crowd control (CC) · rank 53 of 100 among supports1.1 · median 1
Team damage · rank 86 of 100 among supports15% · median 9%

04 — With / against

With whom, against whom?

No opponent with a clear gap to expected in amateur play: win rates remain ranges.

Best results

Ranked by the gap to expected in amateur play, at the lower bound of its interval, at most 5 per role. The other source is a reference.