A Three or more shelves agree
Three or more marketplaces quote the item above the cap and their median sits within 25% of the cheapest. 197 categories, 18,218 copies, $76,585,587.
Absent from Steam: 100.0%.
Exhibit IV-B / the price ladder / 2026-09-23
Valve’s market stops
where the money starts.
Every CS2 item with a live price on five marketplaces was placed in a price band by its cheapest ask, and each band was measured for one thing: what share of it Valve’s own Community Market still carries. Below Valve’s listing cap the answer is 95.3%. Above it, the answer is 1.9%. The drop is not gradual and it does not fall where demand falls — it falls exactly at the published rule. Of the 248 item categories that two or more marketplaces independently price above the cap, 100.0% have no Steam listing at all.
30,266item categories measured
95.3%of sub-cap items carried by Steam
1.9%of above-cap items carried by Steam
$78.6Mfloor value of the corroborated set
01 / the measurement
The cliff lands on the rule, not on demand
Each of the 30,266 item categories below is placed in a band by its cheapest live ask across BUFF163, CSFloat, Skinport, YouPin 898 and C5Game, and only items quoted on two or more of them are counted, so no single optimistic listing can invent a band member. The upper panel is the share of each band that Valve’s own market still carries.
Share of each price band carried by the Steam Community Market
upper: steam coverage · lower: items per band, log scale
at or below Valve’s listing capabove the cap — Valve’s rule forbids the listingitem categories in the band
Coverage means a live Steam ask existed for that category at the moment of capture. The lower panel is logarithmic because the catalogue is overwhelmingly cheap: reading it linearly would flatten every band above $300 into the baseline.
What the shape rules out
If expensive items were simply unpopular, coverage would sag gradually as price rose. It does not. It holds at 76.8% through the $300–$1k band, then collapses to 14.5% in the band immediately beneath the cap and 2.9% in the band immediately above it. A demand effect does not know where Valve set its ceiling. A rule does.
| Price band | Item categories | With a live Steam ask | Steam coverage | Copies listed off Steam |
|---|---|---|---|---|
| 1¢–10¢ | 4,632 | 4,632 | 100% | 8,223,227 |
| 10¢–$1 | 7,152 | 6,931 | 96.9% | 5,488,315 |
| $1–10 | 8,568 | 8,213 | 95.9% | 3,939,538 |
| $10–100 | 5,941 | 5,780 | 97.3% | 1,328,255 |
| $100–300 | 2,329 | 2,121 | 91.1% | 364,466 |
| $300–1k | 1,292 | 992 | 76.8% | 187,504 |
| $1k–1.8k | 193 | 28 | 14.5% | 40,281 |
| $1.8k–3k | 70 | 2 | 2.9% | 9,592 |
| $3k–10k | 67 | 1 | 1.5% | 6,706 |
| $10k–30k | 18 | 0 | 0% | 1,366 |
| $30k–100k | 3 | 0 | 0% | 21 |
| $100k+ | 1 | 0 | 0% | 2 |
Pattern and phase variants are excluded from this ladder: Steam does not list a blue-gem tier or a Ruby phase as its own market entry, so counting them would understate Steam’s coverage. Measured without the two-marketplace requirement the same ladder reads 79.1% below the cap and 6.5% above it across 37,204 items — a starker gap, built on weaker evidence, which is why it is not the figure quoted.
02 / evidential grading
An asking price is not a price
Anyone can list a P2000 for a round ten million yuan. Before a figure on this page is used for anything, it is graded by how many marketplaces independently quote the item above the cap and how tightly the middle of those quotes clusters. Only grades A and B are quoted — here or anywhere else in this dossier.
B Two shelves agree
Two marketplaces agree above the cap, or three agree less tightly. 51 categories, 176 copies, $2,038,814.
Absent from Steam: 100.0%.
C One shelf — not used
A single marketplace carries the only above-cap ask. 171 categories, 247 copies. Shown in the gallery for completeness and excluded from every total on this page.
These are requests for a price, not evidence of one.
What this correction cost, and what it bought
Dropping grade C removes 171 categories and $3,825,125 from the headline. What is left is smaller and much harder to argue with: 248 categories, 18,394 listed copies, $78,624,401, and 100.0% of them — every single one — absent from Valve’s market. The grade-C rows are the reason the figure moved; they are listed, not hidden.
Listings set aside as artefacts
Any quote more than three times the cheapest ask on the same item is set aside. Converted at the rate the data itself implies, most of them are round CNY figures — the signature of a placeholder or a platform ceiling rather than an offer. Because every total on this page uses the cheapest ask, none of these figures could ever have inflated a number; they are excluded so that none of them can be quoted as a headline either.
| Item | Cheapest live ask | Set aside |
|---|---|---|
| Souvenir P2000 | Wicked Sick (Minimal Wear) | $148,954 | $1,489,543 |
| ★ Specialist Gloves | Crimson Kimono (Minimal Wear) | Phase 1 | $37,239 | $119,163 |
| Souvenir USP-S | Printstream (Factory New) | $2,000 | $59,582 |
| ★ Bowie Knife | Crimson Web (Factory New) | $2,957 | $43,030 |
| ★ Karambit | Marble Fade (Factory New) | 1st | $11,916 | $38,874 |
| Sticker | Ninjas in Pyjamas (Holo) | Katowice 2014 | $5,000 | $31,748 |
| Souvenir AWP | Dragon Lore (Battle-Scarred) | $4,225 | $29,157 |
| ★ Karambit | Doppler (Minimal Wear) | Ruby | $7,122 | $27,273 |
03 / the atlas
All 419 categories, with the object each price refers to
Every category whose cheapest ask anywhere clears the cap, with its catalogue render, its evidential grade, the length of the bar showing its price on a logarithmic scale against the rest of the set, and every marketplace quote behind it. Struck-through quotes are the artefacts set aside in section 02.
Renders are the catalogue images for the same items, downscaled and stored in this folder so the page works offline and on the IPFS mirror. Variants of one weapon share the parent render because the pattern tier is not a separate object. The bar under each price is logarithmic and comparative only — it encodes rank within this set, not value.
04 / where it went instead
The shelves that carry what Steam will not
The same listings, grouped by the marketplace holding them. This is a count of where copies are listed, which is not the same as where they are owned and is not read here as such.
Floor value of above-cap listings, by marketplace
lowest ask x copies listed on that shelf
Jurisdictions, in the order above: YouPin 898 — China · BUFF163 — China · C5Game — China · CSFloat — United States · Skinport — Germany / Malta. The same physical item can be listed on more than one of these shelves, so these figures must not be summed with any other market total.
Floor value by price tier
how the money is distributed above the cap
05 / scarcity you can count
How thin the supply actually is
Each CS2 item carries a wear value, and the marketplaces publish how many copies sit in each band of it. That turns “rare” from an adjective into a number — and it is the number that explains why a single compromised account is worth the effort of a bespoke phishing kit.
AWP | Dragon Lore
150 listed · from $11,350
AWP | Dragon Lore (Factory New)
M4A4 | Howl
105 listed · from $6,628
M4A4 | Howl (Factory New)
AK-47 | Wild Lotus
113 listed · from $14,373
AK-47 | Wild Lotus (Factory New)
AWP | Gungnir
173 listed · from $11,618
AWP | Gungnir (Factory New)
AK-47 | Fire Serpent
278 listed · from $663
AK-47 | Fire Serpent (Field-Tested)
Case Hardened (all tiers / blue gem)
76 listed · from $1,638
AK-47 | Case Hardened (Minimal Wear) | Tier 1
Bands are the marketplaces’ own float buckets; only bands with at least one live listing are drawn. The price beside each bar is the cheapest live ask in that band, so the premium attached to a low float is directly readable rather than asserted.
06 / movement, measured off Steam
What the third-party shelves actually realised
Valve publishes no usable history at this tier, and this dossier claims none. Skinport does publish its own completed sales over the trailing 24 hours, 7, 30 and 90 days, and that is the whole of the time dimension available here: a median realised price per trailing window, indexed so the 90-day median is 100. No Steam data is involved in either chart.
Median realised price, trailing window
indexed to the 90-day median · 6,758 items with 20+ sales
Completed sales per day
same items, same windows
These are two charts rather than one with two vertical scales: an index and a daily count share no unit, and drawing them on a common axis would invite a comparison that is not valid. The windows are nested and trailing — the 90-day figure contains the 24-hour figure — so this is not a daily series and must not be read as one. What it supports is narrow and stated plainly: over the trailing windows Skinport published, the median realised price sat below its own 90-day median and the daily sale count fell with it.
Skinport realised sales. The windows are nested and trailing, so each point is the median realised price over the trailing window, indexed to the 90-day median. It is not a daily series and no Steam data is involved.
07 / what it is made of
Sixteen-year-old stickers and a handful of gloves
The set above the cap is not a broad slice of the game. It concentrates in a few rarities and a few collections, several of which can no longer be obtained at all — which is exactly why the supply cannot respond to the price.
Categories above the cap, by rarity
count of item categories
Categories above the cap, by collection
top ten collections
By category: skin 366, sticker 48, container 3, sticker_slab 2.
08 / the most anyone actually paid
Highs backed by a transaction, not by a listing
Every other price on this page is an ask — what a seller wants. These are sales: the highest completed transaction Skinport recorded for each item, over the trailing 90 days. It is the only figure in this dossier that somebody actually paid, and each row links to the marketplace so the number can be checked against its source rather than against us.
21,216items with a completed sale in the window
106sold at or above Valve’s listing cap
$14,192highest single completed sale
67/419above-cap categories with any recorded sale
The thirty highest completed sales in the window
| Item | Highest sale | Median sale | Sales | Cheapest ask now | Steam | Check it |
|---|---|---|---|---|---|---|
| AWP | Gungnir (Factory New) | $14,192 | $11,343 | 5 | $11,490 | absent | Skinport CSFloat Steam |
| StatTrak™ M4A4 | Howl (Factory New) | $12,387 | $12,387 | 1 | $13,836 | absent | Skinport CSFloat Steam |
| AWP | Dragon Lore (Factory New) | $11,779 | $11,063 | 16 | $10,941 | absent | Skinport CSFloat Steam |
| AK-47 | Wild Lotus (Minimal Wear) | $11,693 | $11,693 | 1 | $9,946 | absent | Skinport CSFloat Steam |
| AWP | Gungnir (Minimal Wear) | $10,437 | $10,437 | 1 | $9,610 | absent | Skinport CSFloat Steam |
| StatTrak™ AK-47 | Blue Laminate (Minimal Wear) | $9,772 | $37 | 76 | $30 | $40 | Skinport CSFloat Steam |
| StatTrak™ M4A4 | Howl (Minimal Wear) | $9,463 | $9,463 | 1 | $9,500 | absent | Skinport CSFloat Steam |
| AWP | Dragon Lore (Minimal Wear) | $9,104 | $9,026 | 6 | $8,524 | absent | Skinport CSFloat Steam |
| ★ Sport Gloves | Hedge Maze (Minimal Wear) | $8,139 | $8,139 | 1 | $6,819 | absent | Skinport CSFloat Steam |
| AK-47 | Wild Lotus (Field-Tested) | $7,617 | $7,390 | 4 | $7,194 | absent | Skinport CSFloat Steam |
| StatTrak™ M4A4 | Howl (Field-Tested) | $7,615 | $7,615 | 1 | $7,100 | absent | Skinport CSFloat Steam |
| AWP | Gungnir (Field-Tested) | $7,149 | $7,006 | 2 | $7,165 | absent | Skinport CSFloat Steam |
| AWP | Dragon Lore (Field-Tested) | $6,711 | $6,194 | 15 | $5,988 | absent | Skinport CSFloat Steam |
| M4A4 | Howl (Factory New) | $6,679 | $6,613 | 3 | $6,628 | absent | Skinport CSFloat Steam |
| ★ Sport Gloves | Superconductor (Factory New) | $6,084 | $6,084 | 1 | $7,969 | absent | Skinport CSFloat Steam |
| AWP | Gungnir (Battle-Scarred) | $5,750 | $5,413 | 3 | $5,899 | absent | Skinport CSFloat Steam |
| ★ Sport Gloves | Pandora's Box (Minimal Wear) | $5,520 | $5,508 | 3 | $4,998 | absent | Skinport CSFloat Steam |
| ★ Sport Gloves | Vice (Factory New) | $5,385 | $5,158 | 2 | $5,700 | absent | Skinport CSFloat Steam |
| M4A4 | Howl (Minimal Wear) | $5,342 | $4,989 | 19 | $4,997 | absent | Skinport CSFloat Steam |
| ★ Sport Gloves | Hedge Maze (Field-Tested) | $5,334 | $4,862 | 5 | $3,724 | absent | Skinport CSFloat Steam |
| StatTrak™ AK-47 | Vulcan (Factory New) | $5,275 | $1,418 | 12 | $1,280 | absent | Skinport CSFloat Steam |
| AK-47 | Wild Lotus (Well-Worn) | $4,978 | $4,978 | 1 | $4,942 | absent | Skinport CSFloat Steam |
| AWP | Dragon Lore (Well-Worn) | $4,931 | $4,931 | 1 | $5,445 | absent | Skinport CSFloat Steam |
| AK-47 | Hydroponic (Factory New) | $4,802 | $4,351 | 2 | $3,648 | absent | Skinport CSFloat Steam |
| ★ Talon Knife | Case Hardened (Minimal Wear) | $4,588 | $466 | 16 | $442 | $630 | Skinport CSFloat Steam |
| AWP | Medusa (Factory New) | $4,484 | $4,484 | 1 | $4,171 | absent | Skinport CSFloat Steam |
| ★ Specialist Gloves | Pillow Punchers (Factory New) | $4,359 | $4,359 | 1 | $4,302 | absent | Skinport CSFloat Steam |
| ★ Sport Gloves | Pandora's Box (Field-Tested) | $4,359 | $3,666 | 11 | $3,245 | absent | Skinport CSFloat Steam |
| M4A4 | Howl (Field-Tested) | $4,289 | $4,225 | 8 | $4,022 | absent | Skinport CSFloat Steam |
| AWP | Dragon Lore (Battle-Scarred) | $4,124 | $4,124 | 1 | $4,279 | absent | Skinport CSFloat Steam |
Same item, same 90 days, wildly different price
Restricted to items with a $100 median and at least five completed sales, because on a cheap skin one freak sale swamps the ratio. What is left is dominated by Case Hardened finishes, where the pattern decides the price and the marketplace files every pattern under one name — the same mechanism measured by float in section 05.
| Item | High vs median | Highest sale | Median sale | Sales | Check it |
|---|---|---|---|---|---|
| ★ Talon Knife | Case Hardened (Minimal Wear) | +884% | $4,588 | $466 | 16 | Skinport CSFloat Steam |
| Desert Eagle | Cobalt Disruption (Factory New) | +859% | $1,210 | $126 | 115 | Skinport CSFloat Steam |
| AWP | BOOM (Minimal Wear) | +814% | $1,232 | $135 | 89 | Skinport CSFloat Steam |
| AK-47 | Case Hardened (Well-Worn) | +769% | $1,720 | $198 | 70 | Skinport CSFloat Steam |
| AWP | Graphite (Factory New) | +761% | $1,469 | $171 | 124 | Skinport CSFloat Steam |
| ★ Talon Knife | Case Hardened (Field-Tested) | +509% | $2,437 | $400 | 48 | Skinport CSFloat Steam |
| ★ Kukri Knife | Case Hardened (Minimal Wear) | +448% | $723 | $132 | 69 | Skinport CSFloat Steam |
| ★ Kukri Knife | Case Hardened (Field-Tested) | +415% | $549 | $106 | 128 | Skinport CSFloat Steam |
| ★ Ursus Knife | Case Hardened (Minimal Wear) | +371% | $597 | $127 | 12 | Skinport CSFloat Steam |
| AK-47 | Case Hardened (Field-Tested) | +369% | $1,037 | $221 | 125 | Skinport CSFloat Steam |
| AK-47 | Red Laminate (Minimal Wear) | +360% | $552 | $120 | 18 | Skinport CSFloat Steam |
| AWP | Asiimov (Field-Tested) | +351% | $528 | $117 | 439 | Skinport CSFloat Steam |
| AK-47 | Vulcan (Field-Tested) | +337% | $865 | $198 | 241 | Skinport CSFloat Steam |
| StatTrak™ AK-47 | Nightwish (Factory New) | +322% | $625 | $148 | 31 | Skinport CSFloat Steam |
| AWP | Lightning Strike (Factory New) | +319% | $2,065 | $493 | 198 | Skinport CSFloat Steam |
| StatTrak™ AWP | Redline (Minimal Wear) | +312% | $574 | $139 | 24 | Skinport CSFloat Steam |
| StatTrak™ AWP | Asiimov (Battle-Scarred) | +287% | $524 | $135 | 37 | Skinport CSFloat Steam |
| AK-47 | Redline (Minimal Wear) | +277% | $562 | $149 | 112 | Skinport CSFloat Steam |
| AWP | Crakow! (Factory New) | +275% | $379 | $101 | 72 | Skinport CSFloat Steam |
| StatTrak™ AK-47 | Vulcan (Factory New) | +272% | $5,275 | $1,418 | 12 | Skinport CSFloat Steam |
This is a 90-day high, NOT an all-time high, and it covers ONE marketplace. No all-time figure is claimed anywhere in this dossier. An item with no Skinport sale in the window is absent from this exhibit entirely, which is why the set is smaller than the census.
Verification links are built from the catalogue's own platform identifiers. They point at each marketplace's page for the item so a reader can check the figure against the source rather than against us. Outbound links carry rel="nofollow": they are citations, not endorsements, and several of the destinations are marketplaces this investigation is critical of.
09 / provenance and limits
Where this came from and what it cannot show
Every figure on this page is recomputed from two snapshot files whose hashes are published alongside them. Nothing here is transcribed from another site’s article.
Source and access
cs2.sh API, GET /v1/prices/latest and GET /v1/schema, single snapshot, captured 2026-09-23T00:44:27.871388581Z, priced in USD.
The data was obtained under a paid commercial subscription to the provider, using a licensed key, under the provider’s own terms. No marketplace was scraped and no credential of any marketplace was used. The key itself is held outside this repository and appears in no file here.
How an item qualifies
Every base item with at least one live ask on BUFF, CSFloat, Skinport, YouPin or C5Game is placed in a price band by its CHEAPEST such ask, so the band never depends on Steam. "Steam coverage" is the share of the items in that band carrying any live Steam ask at the moment of capture. Pattern and phase variants are excluded from the ladder because Steam does not list them separately; including them would understate Steam. Renders are the catalogue images for the same items, stored locally.
What we could not obtain
The subscription tier used serves current snapshots only. Historical and archive endpoints were not available, and no historical series is claimed anywhere on this page beyond the trailing-window sales Skinport publishes itself in section 06.
One snapshot, not a time series. Absence of a Steam listing at the moment of capture is not proof an item may never appear there; it is the measured state of the shelf at that moment, and for prices above the cap it is also what Valve's published rule requires: Valve's Community Market explicitly caps listings at $1,800 per item (with a maximum Steam Wallet balance of $2,000; see Steam Community Market FAQ archived 2026-09-21 ↗ and Steam Subscriber Agreement § 3.C archived 2026-09-24 ↗).
Grading
An item's price is always its cheapest live ask, so a high outlier can never inflate a total. Grading governs which rows may be quoted. Any quote more than three times the cheapest ask on the same item is set aside as not a live price. Grade A: three or more marketplaces still quote above the cap and their median sits within 25 per cent of the cheapest. Grade B: two marketplaces agree, or three agree less tightly. Grade C: a single shelf - an asking price, not a market price. Only grades A and B are quoted anywhere in this dossier.
Outlier rule: quote > 3.0x the cheapest ask on the same item.
Why the totals are a floor
For each marketplace the item’s lowest ask is multiplied by the number of copies that shelf lists. Every other copy there is priced at or above that lowest ask, so the true posted value is higher by construction, not by estimate. It is not a sale, not a cash flow, and not a valuation of the CS2 economy.
Check it yourself
The derived datasets behind every chart on this page are published in full: the atlas dataset, the grading and artefact screen, the underlying census and the manifest with the SHA-256 of each raw snapshot.
A correction recorded in the open
An earlier draft of the companion register led with a Souvenir AK-47 | Point Disarray asked at $148,954 on one Chinese marketplace while Steam priced it at $119. That row is grade C: a single shelf, a single copy, and no second marketplace agreeing. It is an asking price, not a price, and it has been demoted here and on that page. The corroborated leader is Sticker | Titan (Holo) | Katowice 2014 at $66,285, quoted by 4 marketplaces within 5.2% of each other — and absent from Steam entirely.