TFT Star Levels Explained: Is a 4-Star Unit Actually Possible?

TL;DR: Combining copies stops at 3-star: three copies make a 2-star and nine make a 3-star. Set 18 has 4-star units anyway. Riot’s patch 18.4 notes (October 6, 2026) give 4-star values for Camille and Xayah, and in Riot’s public match data from 110,328 ranked games played August 25 to October 4, 2026, 0.46% of final boards had at least one 4-star unit. No unit in that data was above 4-star. This guide covers the copy math, the Set 18 numbers, pool sizes and when 3-starring is worth it.

A common autocomplete question is “TFT is it possible to 4-star”. In Set 18 the answer is yes, just not by buying more copies. Here is how star levels work, and what the match data says about 4-stars.

The direct answer

In Teamfight Tactics, the combine system caps at 3-star: nine copies of a champion. A tenth copy does not make a 4-star, because combining has no fourth tier.

Set 18, Enchanted Wilds, has 4-star units anyway. Riot’s patch 18.4 notes, published October 6, 2026, list Camille’s damage as 150/225/375/640 AD before the patch and 150/225/400/665 AD after it, and Riot says the change buffs her three- and four-star forms. The same notes list Xayah’s ability damage with four values (68/102/155/165 to 73/108/165/175) and fix a bug where her 4-star ability damage was wrong.

Riot’s notes do not list every way a unit can reach 4-star, and the match data shows that 4-stars appear, not how each one was made. So this page gives the numbers and leaves the mechanics to the in-game tooltips.

How the combine math works

Star-ups happen automatically. The instant you own the third copy of a champion at a given tier, the game merges them for you.

You collect	Result	Total 1-star copies used
3 x 1-star	1 x 2-star	3
3 x 2-star	1 x 3-star	9

So a 3-star is nine base copies of the same champion. Each star level is a large stat increase, and 3-star low-cost units can become the single strongest carries in the game, which is exactly why reroll comps exist.

How often 4-star units showed up in Set 18

We counted star levels in Riot’s public TFT match data: 110,328 ranked games on the North America, Europe West, Europe Nordic and East, and Korea servers, played between August 25 and October 4, 2026 (UTC), reached through players ranked Gold, Platinum, Diamond, Master, Grandmaster and Challenger. Each game has eight final boards, the units each player had when their game ended, so that is 882,624 boards. All of these games are from before patch 18.4 went live on October 7, so they do not include its Camille and Xayah buffs.

Measure	Count
Final boards	882,624
Boards with at least one 4-star unit	4,018 (0.46%, about 1 in 220)
Boards with exactly one 4-star unit	3,951
Boards with two or more	67
4-star units in all	4,195
Units above 4-star	0
Most common 4-star unit	Veigar, 535 of the 4,195
Boards with a 4-star that ended with Solar active	441 of 4,018 (11%)

Across every unit on those boards, this is how the star levels split:

Star level	Units	Share
1-star	2,094,306	28.7%
2-star	4,318,265	59.1%
3-star	891,177	12.2%
4-star	4,195	0.06%

Two things to read carefully:

• Solar was active on only 441 of the 4,018 boards with a 4-star. The other 3,577 boards with a 4-star did not have Solar active at the end of the game. Riot’s 18.4 notes say the Solar tooltip now highlights correctly at 7 three-stars, but the data does not show which effect made each 4-star.
• Boards with a 4-star placed well: 2.57 on average, 88.7% in the top four and 32.0% in first. That is a correlation, not proof that the 4-star won those games: a player who lasts longer also has more rounds to build one.

Why combining stops at 3-star

The combine system only knows three tiers, so buying more copies of a champion you already have at 3-star does not make it a 4-star. Riot’s notes do not describe a rule that combines 3-stars.

Even reaching 3-star is limited by the shared pool, so it is worth understanding.

Pool sizes: why some units are easy to star and others are not

Every champion cost has a fixed number of copies that exist in the entire lobby’s shared pool, and the whole lobby draws from it. Cheaper units have far more copies than expensive ones. Exact numbers vary by set, but the shape is always the same:

Cost	Pool size (copies in the game)	3-star difficulty
1-cost	Largest pool	Achievable in a reroll comp
2-cost	Large pool	Achievable in a reroll comp
3-cost	Medium pool	Hard, needs commitment and low contest
4-cost	Small pool	Very rare, usually not the plan
5-cost	Smallest pool	Almost never, and not worth forcing

This is why you 3-star cheap units and not expensive ones. There are simply enough 1-cost copies in the pool for one player to gather nine; there are not enough 5-cost copies for that to be realistic, and other players buying the same unit shrink your odds further.

When 3-starring is worth it (and when it is a trap)

3-starring a unit is a massive power spike, but it costs a lot of gold and locks you into a plan. The decision comes down to which unit and which comp.

Worth it: low-cost reroll carries. Comps built around a 1 or 2 cost carry are designed to 3-star that carry. You stall at a lower level (5, 6, or 7), slow roll your gold down to your interest breakpoint each round, and collect copies from the large pool. The 3-star carry is the entire payoff of the comp. We walk through this fully in the TFT reroll guide.

A trap: chasing 3-star 4 and 5 cost units. Their pools are tiny, so you can spend enormous gold and simply never see the copies, because they may not exist in the pool or are being bought by others. Standard 4-cost comps aim for 2-star 4-costs, not 3-star ones. Going for a 3-star 5-cost is a high-roll dream, not a plan you can build around.

Situation	Verdict
3-star a 1 or 2 cost carry in a reroll comp	Correct, it is the win condition
3-star a 3-cost carry when uncontested	Situationally strong, commit early
3-star a 4-cost	Rare high-roll, do not force it
3-star a 5-cost	Not a realistic plan

Whether to reroll for a 3-star or push levels for stronger units is really an econ and leveling question. Our leveling guide covers which levels to roll at for each cost, and the economy guide explains why slow rolling works.

Quick reference

• 2-star: 3 copies.
• 3-star: 9 copies, the cap for combining.
• 4-star: not from combining, but it exists in Set 18. Riot’s 18.4 notes give 4-star values, and 0.46% of final boards in 110,328 ranked games (August 25 to October 4, 2026) had one.
• 5-star: none in that Set 18 data.
• 3-star cheap carries (1 to 2 cost) in reroll comps; do not force 3-star on 4 and 5 costs.

For the Set 18 roster and traits, see Set 18 champions and traits, and the comps page lists the current Set 18 reroll lines with their items.

Where to go next

• Learn to slow roll correctly in the TFT reroll guide.
• Nail your roll timings in the leveling guide.
• Understand the econ behind rerolling in the economy guide.

Star levels are simple once the copy math clicks: three to 2-star, nine to 3-star, and anything above that comes from the set, not from buying more copies.

Sources

• Riot Games, Teamfight Tactics patch 18.4, published 2026-10-06, read 2026-10-10: Camille’s and Xayah’s values up to 4-star, the Xayah 4-star bug fix and the Solar tooltip fix.
• Riot Games public TFT match data (tft-match-v1): 110,328 ranked games on the NA, EUW, EUNE and KR servers, reached through players on the Gold, Platinum, Diamond, Master, Grandmaster and Challenger ladders, played 2026-08-25 22:47 UTC to 2026-10-04 14:20 UTC, all before patch 18.4. The same games the how long is a TFT game page measures. The counts are ours.

Update log

• 2026-10-10: Rewritten for Set 18. The July version said 4-star units do not exist in normal play; Riot’s 18.4 notes give 4-star values for Camille and Xayah, and 0.46% of final boards in 110,328 ranked games had a 4-star unit. Added the counts, the star-level split and the Solar figure, and removed a line about overlay features.
• 2026-07-12: First version.

Champion, trait, and cosmetic art (c) Riot Games.