How pattern tiers work

1. T1Ranks ~1 - 50.5%
2. T2Ranks ~6 - 202%
3. T3Ranks ~21 - 707%
4. T4Ranks ~71 - 15015%
5. -Rank ~151+~85%

Ranks are given for 1001 seeds; a board missing a render uses the same percentages of what was measured. These boundaries are defaults, not quotas: each one can slide up to half its width to land on a real break in the scores, so a tier can hold 3 seeds on one board and 8 on another. The same rule runs on every skin and every board, so a new analysis gets its tiers the moment it is published.

Every finish has 1001 pattern seeds, 0 to 1000. We render each one and measure a score, usually the share of a colour on a zone: the blue on the top of an AK-47 Case Hardened, the fade on a Karambit blade. Each measurement is its own board, and one skin can have several.

1,001

seeds per finish, 0 to 1000

~15%

of top seeds can hold a tier

0

prices, votes or opinions used

• Why is a tier so narrow?

  Tier sizes are not fixed: each boundary slides to a real break in the scores. On one board T1 can hold 3 seeds, on another 7. A narrow tier is the measured shape, not an error.

• Why is a famous seed not ranked first?

  A rank compares measured scores only - for example how much blue sits on the top zone. A seed can be famous for how its pattern looks while another seed covers more of that zone. The board answers one question: which seeds score highest on this measurement.

• Why does one seed get different tiers on different boards?

  Each board is its own measurement. A seed can be T1 on Blue Top and hold no tier on Gold, the same way a tall person is not automatically heavy. A tier only speaks for the measurement that made it.

• Can a seed's tier change over time?

  Yes. Tiers are computed when the board is read, never stored. A re-analysis can move scores, and seeds within 0.6 points of a boundary - the dashed chips - are the first that could flip.

• Why does a board show no tiers at all?

  When the best seeds descend as evenly as the rest, or the whole board is nearly flat, every boundary would be invented. Showing ranks only is the honest answer - the fade boards do exactly this.

• Does a higher tier mean a higher price?

  No. No prices, sales or votes feed the tiers. The market prices looks, demand and rarity; we only rank one measured number on one board.

• Why four tiers and not more?

  Four is enough to name the groups these distributions actually form. The defaults cover about the top 15% of seeds; past that, scores blend into the middle of the board where extra tiers would only relabel the same seeds.

• What does 'seed X not measured' mean?

  We only tier seeds we can render. A missing render is skipped, the percentages apply to the measured count, and the board lists which seeds are absent.

Two seeds with the same score share a rank (60, 60, then 62) and always land in the same tier. If a tier boundary falls between two equal scores, it moves up so both go to the lower tier, unless that would leave the tier empty.

AWP | Fade · Orange · Front

• Too uniformIf the scores at the top and bottom of the tiered range are less than 1.8 points apart, no seed is visibly rarer than another: the board shows ranks only.
• No real eliteIf the best seeds descend no faster than the rest of the board, there is no elite group to name and every boundary would be invented: the board shows ranks only.
• Close to typicalA seed less than 1.2 points above the median seed never gets a tier, even inside the tiered range.
• Too few seedsA board with fewer than 200 measured seeds shows ranks only.
• Reverse boardsA reverse fade ranking is the exact inverse of the normal fade board, so tiers are only given on the normal one.
• Other viewsTiers use the Front view only, the side you see in game. Back, Top and the other views show ranks only.
• What a tier does not sayA tier is not a price and not a community pattern name like Blue Gem. It ranks one measured number against the other seeds of the same board, nothing more.