What is the best build for Cassiopeia?

We do not publish a best build for Cassiopeia, and the reason is worth thirty seconds. The most recorded Cassiopeia mid build on patch 16.14 is Rylai's Crystal Scepter, Liandry's Torment and Rod of Ages, in 30 games. A finished item set is read at the end of the game, so a build only appears once the game ran long enough to buy it, and a game that runs long is usually a game you were winning. Any share of wins taken from those games leans high for that reason on its own, which is why there is not one in this answer. The page lists every recorded build with the games behind it.

What are the best runes for Cassiopeia?

The most recorded Cassiopeia mid rune page on patch 16.14 takes Deathfire Touch out of Sorcery, with Manaflow Band, Celerity and Scorch beside it and Cut Down and Legend: Haste out of Precision. Players who ran it won 54.9% of 546 games, range 50.8% to 59.1%. A rune page is locked in at champion select, so that share is not inflated by long games the way a build is, but it counts who picked the page rather than what the page did.

Cassiopeia build, patch 16.14

Q: What skill order should I use on Cassiopeia?

Recorded Cassiopeia mid players max E, then Q, then W. 232 recorded games on patch 16.14. Win comparisons only count games that ran long enough to finish the max: 100% of 230 E-first games and 100% of 2 Q-first games. Short games are missing from both sides, and short games are usually decided ones.

1,560 games of Cassiopeia recorded on patch 16.14, across mid, bot, top, support and jungle.

Win rate

54.4% of mid games this patch

Pick rate

1.1% of mid games

Games

835 mid, patch 16.14

Rank in role

8 of 109 mid champions

Win impact against the average mid pick on patch 16.14: +1.3 wins per 100 games (range -0.6 to 3.1) 835 games Pooled

What most Cassiopeia mid players build, patch 16.14

Build Rylai's Crystal Scepter, Liandry's Torment and Rod of Ages, and run Deathfire Touch

The rune page they run

Sorcery

Deathfire Touch

Precision

Shards

546 games ran this page exactly as it is drawn. We never stitch the most picked keystone onto the most picked second row and call it a page: nobody ran that one.

The skill path Each filled cell is the skill most players took at that level, from 232 games.
skill	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
Q		Q						Q		Q		Q	Q
W			W	W										W	W		W	W
E	E				E		E		E
R						R					R					R

Levels 10 to 18

10 Q, 11 R, 12 Q, 13 Q, 14 W, 15 W, 16 R, 17 W, 18 W

Max E, then Q, then W. The level by level count.

30 games finished this exact set. It is the set recorded most often, which is a fact about players rather than about the items.

Share of those games that were wins, in percent 70% (range 52-83) · 30 games · Measured

Games that finish a build are games that lasted, and lasting usually means winning, so this figure leans high before the items do anything. It is how those games went, not what the build does to your chances. How far that can bend a number.

Cassiopeia mid is the role with the most recorded games on patch 16.14: 835 games, of which 454 were wins. That is the role this page leads with. Measured on this page: mid, bot and top, the roles that cleared 30 recorded games.

wins +1.3 per 100 vs the mid average · 835 games on 16.14 · first-time players measured (792 games) · +15.0 vs 16.13

The build above is the one recorded most often, nothing more. Every recorded build, with its own games and its own share, is in the build list, in the order players built them.

Why this answer

A champion and role pair needs 30 recorded games before we put a score or a win impact on it: under that, a handful of lucky games moves a win rate by ten points. The same floor applies to a single build and to a single rune page, and anything under it prints its games and no number.

The share on that figure says how the games went, not what the build does to you. Nothing on this page measures that; the method note at the end says why.

What do most Cassiopeia players build?

Item sets recorded on patch 16.14, most built first, never sorted by how the games went. The share is how many of a build's own games were wins; a build under 30 games shows its games and no share.

Cassiopeia, mid

3 item sets were recorded in this role and 1 of them cleared 30 games.

Rylai's Crystal Scepter, Liandry's Torment and Rod of Ages
70.0% 30 games

A finished set is read at the end of the game, so every share here leans high: games that last long enough to finish a build are usually games that were already going well. How far that can bend a number.

Cassiopeia, bot

1 item sets were recorded in this role and 0 of them cleared 30 games.

Rylai's Crystal Scepter, Zhonya's Hourglass and Rod of Ages
not enough games yet. 14 games so far, this needs 30. 14 games

Cassiopeia, top

3 item sets were recorded in this role and 0 of them cleared 30 games.

Rabadon's Deathcap, Rylai's Crystal Scepter and Rod of Ages
not enough games yet. 8 games so far, this needs 30. 8 games
Rylai's Crystal Scepter, Liandry's Torment and Rod of Ages
not enough games yet. 8 games so far, this needs 30. 8 games
Rylai's Crystal Scepter, Liandry's Torment and Bloodletter's Curse
not enough games yet. 8 games so far, this needs 30. 8 games

What runes do most Cassiopeia players run?

Whole pages that players actually ran, most recorded first. A page is locked in at champion select, so long games do not inflate its share the way they inflate a build's; it still counts who picked the page rather than what the page did.

Cassiopeia, mid

8 complete pages were recorded in this role and 4 of them cleared 30 games.

Deathfire Touch Sorcery with Precision

546 games

Sorcery

Deathfire Touch

Precision

Shards

546 games ran this page exactly as it is drawn. We never stitch the most picked keystone onto the most picked second row and call it a page: nobody ran that one.

Other pages players ran

Conqueror Precision with Sorcery

111 games

Deathfire Touch Sorcery with Resolve

110 games

Stormraider's Surge Sorcery with Precision

46 games

Keystones picked, one slot on its own

Deathfire Touch: 54.3% 665 games
Conqueror: 59.8% 122 games
Stormraider's Surge: 46.0% 50 games

3 keystones were recorded in this role: one slot on its own, counted across every page that used it, so the share carries the rest of each page with it.

Cassiopeia, bot

5 complete pages were recorded in this role and 2 of them cleared 30 games.

Deathfire Touch Sorcery with Precision

211 games

Sorcery

Deathfire Touch

Precision

Shards

211 games ran this page exactly as it is drawn. We never stitch the most picked keystone onto the most picked second row and call it a page: nobody ran that one.

Other pages players ran

Conqueror Precision with Sorcery

126 games

Keystones picked, one slot on its own

Deathfire Touch: 49.6% 234 games
Conqueror: 42.9% 126 games

3 keystones were recorded in this role: one slot on its own, counted across every page that used it, so the share carries the rest of each page with it.

Cassiopeia, top

6 complete pages were recorded in this role and 2 of them cleared 30 games.

Deathfire Touch Sorcery with Precision

172 games

Sorcery

Deathfire Touch

Precision

Shards

172 games ran this page exactly as it is drawn. We never stitch the most picked keystone onto the most picked second row and call it a page: nobody ran that one.

Other pages players ran

Conqueror Precision with Sorcery

125 games

Keystones picked, one slot on its own

Deathfire Touch: 51.6% 192 games
Conqueror: 53.0% 134 games

3 keystones were recorded in this role: one slot on its own, counted across every page that used it, so the share carries the rest of each page with it.

What skill order should I use?

Counted from recorded level-ups on patch 16.14, one player per game.

Max E, then Q, then W

232 recorded games on patch 16.14. Win comparisons only count games that ran long enough to finish the max: 100% of 230 E-first games and 100% of 2 Q-first games. Short games are missing from both sides, and short games are usually decided ones.

not enough games

Cassiopeia mid skill path Each filled cell is the skill most players took at that level, from 232 games.
skill	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
Q		Q						Q		Q		Q	Q
W			W	W										W	W		W	W
E	E				E		E		E
R						R					R					R

Levels 10 to 18

10 Q, 11 R, 12 Q, 13 Q, 14 W, 15 W, 16 R, 17 W, 18 W

Almost nobody maxed Q first, so there is nothing to compare. -1 wins per 100 games (range -12 to 10) 232 games Measured 230 games maxed E first, 2 games maxed Q first.

Max E, then Q, then W

103 recorded games on patch 16.14. Win comparisons only count games that ran long enough to finish the max: 100% of 102 E-first games and 100% of 1 Q-first games. Short games are missing from both sides, and short games are usually decided ones.

not enough games

Cassiopeia bot skill path Each filled cell is the skill most players took at that level, from 103 games.
skill	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
Q		Q						Q		Q		Q	Q
W			W											W	W		W	W
E	E			E	E		E		E
R						R					R					R

Levels 10 to 18

10 Q, 11 R, 12 Q, 13 Q, 14 W, 15 W, 16 R, 17 W, 18 W

Almost nobody maxed Q first, so there is nothing to compare. +1 wins per 100 games (range -11 to 12) 103 games Measured 102 games maxed E first, 1 game maxed Q first.

Max E, then Q, then W

101 recorded games on patch 16.14.

Cassiopeia top skill path Each filled cell is the skill most players took at that level, from 101 games.
skill	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
Q		Q						Q		Q		Q	Q
W				W										W	W		W	W
E	E		E		E		E		E
R						R					R					R

Levels 10 to 18

10 Q, 11 R, 12 Q, 13 Q, 14 W, 15 W, 16 R, 17 W, 18 W

When do players buy each piece?

The pieces of the most recorded Cassiopeia mid build, in the order players buy them. The time is the middle of the spread across every recorded game on patch 16.14, measured at the buy rather than at the end of the game: half the recorded buys of the item happen before it, half after.

Buy times, earliest first

Rod of Ages: 11:13 403 games
Rylai's Crystal Scepter: 18:05 705 games
Liandry's Torment: 18:46 906 games

The bar is scaled to the latest piece of the set. Each item links to its own page, where the whole spread sits beside the games behind it.

Who counters Cassiopeia?

The opponents Cassiopeia mid loses to most and the teammates Cassiopeia gains the most beside on patch 16.14. The chip is the edge in wins per 100 games, and the games behind it sit under the name.

Hardest opponents

Not enough games yet to rank Cassiopeia's opponents. Cassiopeiamid has 863 games recorded on patch 16.14, and ranking opponents needs 1,000 before the baseline is steady; every opponent so far is counted, not ranked.

Best teammates

Not enough games yet to rank the champions Cassiopeia wins beside.

These numbers come from 108,616 ranked games we recorded on EUW, Korea and NA, found by following Diamond players and the people they play with. That is not every rank and it is not every region. Everything on this page is patch 16.14. Counted up to 24 July 2026.

How does Cassiopeia rank?

Every role we recorded Cassiopeia in on patch 16.14, scored or not. Win impact is measured against the average pick in that role on the same patch, and a role needs 30 recorded games before it is scored.

Cassiopeia by role

mid B +1.3 wins per 100 games (range -0.6 to 3.1) 835 games Pooled: 56 835 games
bot C -0.4 wins per 100 games (range -2.4 to 1.6) 376 games Pooled: 44 376 games
top B +0.4 wins per 100 games (range -1.6 to 2.4) 344 games Pooled: 48 344 games
support not enough games yet. 4 games so far, this needs 30.: not scored 4 games
jungle not enough games yet. 1 game so far, this needs 30.: not scored 1 game

Why no share of wins on this page says a build is good for Cassiopeia

A finished item set is read at the end of the game. A player who bought the first two items and lost at twenty minutes never finished the set, so they are not in its row. Everybody in the row lasted long enough to buy it, and lasting is most of what winning looks like. The share those games won is pushed up by that on its own, before the items do anything at all, and the push is largest on the most expensive builds and backwards on the ones you buy when a game is going badly.

We know the size of it. In simulation, with every item's real effect set to exactly zero, an item read at the end of the game shows about 92% as a sixth completed item. A defensive item with a real 5 point benefit shows about 20 points the wrong way, because it gets bought when a player is behind. The measurement is here.

So this page counts and labels rather than ranks. The estimate that answers the real question is taken at the moment of the buy, from the state the purchase happened in, and it runs in the desktop app because it needs the game in front of it.