What is the best build for Rumble?

We do not publish a best build for Rumble, and the reason is worth thirty seconds. The most recorded Rumble top build on patch 16.14 is Zhonya's Hourglass, Liandry's Torment and Bloodletter's Curse, in 51 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 Rumble?

The most recorded Rumble top rune page on patch 16.14 takes Deathfire Touch out of Sorcery, with Nimbus Cloak, Celerity and Scorch beside it and Unflinching and Bone Plating out of Resolve. Players who ran it won 52.1% of 357 games, range 46.9% to 57.2%. 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.

Rumble build, patch 16.14

Q: What skill order should I use on Rumble?

Recorded Rumble top players max Q, then E, then W. 200 recorded games on patch 16.14. Win comparisons only count games that ran long enough to finish the max: 100% of 199 Q-first games and 100% of 1 E-first games. Short games are missing from both sides, and short games are usually decided ones.

844 games of Rumble recorded on patch 16.14, across top, mid, jungle, support and bot.

Win rate

50.5% of top games this patch

Pick rate

0.9% of top games

Games

699 top, patch 16.14

Rank in role

50 of 122 top champions

Win impact against the average top pick on patch 16.14: +0.1 wins per 100 games (range -1.8 to 2.0) 699 games Pooled

What most Rumble top players build, patch 16.14

Build Zhonya's Hourglass, Liandry's Torment and Bloodletter's Curse, and run Deathfire Touch

The rune page they run

Sorcery

Deathfire Touch

Resolve

Shards

357 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 200 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 E, 11 R, 12 E, 13 E, 14 W, 15 W, 16 R, 17 W, 18 W

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

51 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 55% (range 41-68) · 51 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.

Rumble top is the role with the most recorded games on patch 16.14: 699 games, of which 353 were wins. That is the role this page leads with. Measured on this page: top, mid and jungle, the roles that cleared 30 recorded games.

wins 0.1 per 100 vs the top average · 699 games on 16.14 · first-time players measured (641 games) · no clear move vs 16.13 (-5.0, range -19.1 to +9.1)

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 Rumble 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.

Rumble, top

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

Zhonya's Hourglass, Liandry's Torment and Bloodletter's Curse
54.9% 51 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.

Rumble, mid

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

No mid build recorded yet

We have 83 games of Rumble mid on patch 16.14, but no item set repeated often enough to record as a build.

Read the rune pages instead

Rumble, jungle

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

No jungle build recorded yet

We have 37 games of Rumble jungle on patch 16.14, but no item set repeated often enough to record as a build.

Read the rune pages instead

What runes do most Rumble 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.

Rumble, top

10 complete pages were recorded in this role and 3 of them cleared 30 games.

Deathfire Touch Sorcery with Resolve

357 games

Sorcery

Deathfire Touch

Resolve

Shards

357 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

Deathfire Touch Sorcery with Inspiration

135 games

Arcane Comet Sorcery with Resolve

92 games

Keystones picked, one slot on its own

Deathfire Touch: 52.5% 514 games
Arcane Comet: 48.5% 132 games

7 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.

Rumble, mid

10 complete pages were recorded in this role and 0 of them cleared 30 games.

Deathfire Touch Sorcery with Resolve

not enough games yet. 20 games so far, this needs 30. 20 games

Sorcery

Deathfire Touch

Resolve

Shards

20 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

Dark Harvest Domination with Precision

not enough games yet. 19 games so far, this needs 30. 19 games

Deathfire Touch Sorcery with Domination

not enough games yet. 9 games so far, this needs 30. 9 games

Keystones picked, one slot on its own

Deathfire Touch: 48.6% 35 games

6 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.

Rumble, jungle

9 complete pages were recorded in this role and 0 of them cleared 30 games.

Dark Harvest Domination with Sorcery

not enough games yet. 8 games so far, this needs 30. 8 games

Domination

Dark Harvest

Sorcery

Shards

8 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

First Strike Inspiration with Domination

not enough games yet. 6 games so far, this needs 30. 6 games

Conqueror Precision with Inspiration

not enough games yet. 5 games so far, this needs 30. 5 games

Keystones picked, one slot on its own

Dark Harvest: not enough games yet. 14 games so far, this needs 30. 14 games
First Strike: not enough games yet. 7 games so far, this needs 30. 7 games
Conqueror: not enough games yet. 7 games so far, this needs 30. 7 games

4 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 Q, then E, then W

200 recorded games on patch 16.14. Win comparisons only count games that ran long enough to finish the max: 100% of 199 Q-first games and 100% of 1 E-first games. Short games are missing from both sides, and short games are usually decided ones.

not enough games

Rumble top skill path Each filled cell is the skill most players took at that level, from 200 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 E, 11 R, 12 E, 13 E, 14 W, 15 W, 16 R, 17 W, 18 W

Almost nobody maxed E first, so there is nothing to compare. +1 wins per 100 games (range -11 to 12) 200 games Measured 199 games maxed Q first, 1 game maxed E first.

Rumble, mid

Not enough mid level-ups yet

not enough games yet. 25 games so far, this needs 30.

Rumble mid has 83 games recorded, but only 25 games of them carry level-ups. A levelling order off that many games is luck.

How we count level-ups

Rumble, jungle

Not enough jungle level-ups yet

not enough games yet. 15 games so far, this needs 30.

Rumble jungle has 37 games recorded, but only 15 games of them carry level-ups. A levelling order off that many games is luck.

How we count level-ups

When do players buy each piece?

The pieces of the most recorded Rumble top 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

Liandry's Torment: 18:46 906 games
Bloodletter's Curse: 24:40 131 games
Zhonya's Hourglass: 25:56 1,680 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 Rumble?

The opponents Rumble top loses to most and the teammates Rumble 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 Rumble's opponents. Rumbletop has 750 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 Rumble 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 Rumble rank?

Every role we recorded Rumble 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.

Rumble by role

top C +0.1 wins per 100 games (range -1.8 to 2.0) 699 games Pooled: 47 699 games
mid C -0.1 wins per 100 games (range -2.3 to 2.0) 83 games Pooled: 44 83 games
jungle C 0.0 wins per 100 games (range -2.2 to 2.1) 37 games Pooled: 45 37 games
support not enough games yet. 22 games so far, this needs 30.: not scored 22 games
bot not enough games yet. 3 games so far, this needs 30.: not scored 3 games

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

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.