LIS-Skins vs CSFloat: reading the price difference

Float value and paint seed are the two fields that turn a generic skin into a specific one. Float sets the wear; seed sets the pattern — fade percentage, blue coverage on Case Hardened, the marble arrangement on a Marble Fade. Two identical-looking listings at the same wear can be worth wildly different amounts because of seed alone, and a venue that hides seed is asking you to buy blind on the attribute that moves price most.

A cell that says "not surfaced" means the field is not part of what that venue publishes in the form described here. It does not mean the attribute does not exist on the item.

Two things in that table are genuinely distinctive. First, LIS-Skins exposes both float and paint seed publicly with no key, no account and no cookie. That is rare: most venues gate one of the two behind a login, an API key, or a per-item inspect step you have to run yourself. As a buyer it means you can verify a pattern claim before you register anywhere or deposit anything. The public listing data runs to roughly 2.1 million CS2 listings, which is why it is best treated as something you filter through rather than something you read.

Second, CSFloat is the only side of this comparison that attaches its own predicted price to the item. Whether you agree with that prediction is a separate question, but a number that is computed the same way for every listing and is publicly attached to a permalink is something you can argue with. A price with no permalink behind it is a rumour.

One more asymmetry worth understanding: paint index is what Doppler phase is derived from. LIS-Skins hands you the number and leaves the lookup to you; CSFloat prints the phase. If you know the index-to-phase mapping, LIS-Skins is arguably giving you more — the raw field rather than someone else's interpretation of it.

Fee incidence. CSFloat charges the seller 2%, so an asking price there is close to a net number for the seller and close to the sticker for the buyer. LIS-Skins splits its costs differently, and its current fee structure should be read on its own fee page rather than assumed — a cross-market comparison built on a guessed fee is worse than no comparison at all. Treat any number you have not personally read as unknown, not as zero.

Trade locks. Because LIS-Skins publishes trade-lock state per listing, you can see the discount that locked inventory carries. A locked item is effectively a forward: you pay today for delivery up to a week out and you carry the price risk in between. A meaningful share of any gap you spot between the two venues on the same skin is simply one of them showing you locked stock.

Payment rails and currency. The figure on screen is not necessarily what leaves your bank. Conversion spread, method-specific surcharges, and balance top-up incentives that only exist for people funding from site balance all move the effective price. Price the rail you would actually use, not the headline.

Depth and who sets the floor. Where many copies of a skin are listed concurrently, the visible floor is set by whoever is most impatient to sell. Where inventory is thin, it is set by whoever happens to be listed at all. Same item, two floors, neither of them wrong.

Access and account state. Some payment and withdrawal methods are region-limited. Steam's own rules — mobile authenticator, account age, holding periods, the item's own trade state — apply no matter where you bought. A price you cannot pay at, on an item you cannot receive, is not cheaper.

Staleness. This is the one people miss. LIS-Skins' own documentation says its exports lag the live market by several minutes, so anyone reading that data — you, an aggregator, a price table someone posted on Reddit — is looking at a snapshot rather than the live book. Popular items are therefore seen through a window rather than in full. If something in a lagging snapshot looks dramatically underpriced, the first hypothesis should be that it is already sold, not that you found something. That is not a criticism of the venue; a published export is a generous thing for a marketplace to offer at all. It just means the number in front of you is a recent past price, and you should treat a spectacular one as a question rather than an answer.

A price is a claim about a transaction you can complete. Apply three tests before you believe one. What currency and rail would you be paid in? Site balance is not money until it is withdrawn, and the withdrawal step is often where the real cost hides. Which market can you actually sell into? Your realisable price is set by your exit, not by the venue with the prettiest listing. What account would you need? If the exit requires verification, a region, or a trade state you do not have, the number is decorative.

The practical consequence is that you need one denominator and should stick to it. CSFloat's Final is a defensible yardstick — not because it is authoritative, but because it is computed consistently per item, quoted publicly, and anchored to a permalink someone else can open and check. That is the choice CSWatcher makes: it scores listings on other markets against CSFloat's Final and deep-links you to the listing so you complete the purchase yourself, because on every market other than CSFloat it is alert-only and issues reads only. Both surfaces — the desktop app and the browser app — cover LIS-Skins identically, and each reads all 10 markets. Every market scans and alerts on Spectator, the free plan; a Sniper pass is what lets CSWatcher buy for you where a marketplace supports it.

LIS-Skins genuinely wins in several situations. If you are pattern-hunting, the public float-and-seed data lets you do real filtering before you commit to an account anywhere — that is a concrete advantage over venues that make you register to see the seed. If you are buying a considered item rather than racing for one, the lag costs you nothing. Breadth matters too: with listing counts in the millions, long-tail combinations — odd wears, StatTrak and souvenir variants, unloved skins with good seeds — are simply more likely to exist there than on a thinner book. And if you are comfortable holding through a trade lock, the discount on locked inventory is real money.

It is the weaker choice when timing matters. On a hot item, several minutes of lag means the cheap listings you see are a window into a book that has already moved. It is also weaker when you need the purchase to stay auditable: a stable per-listing URL is what lets you or anyone else re-check a claim a week later, and snapshot data does not give you that. If you are selling rather than buying, the calculation flips again — your exit venue sets your number, and CSFloat's 2% seller fee is one of the very few figures in this space you can compute with directly rather than estimate.

The honest summary: use LIS-Skins' published data to find candidates and check patterns, use a consistent yardstick to judge whether a candidate is actually cheap, and price the fees and rails you personally face before calling anything a bargain.

If you want that comparison run continuously rather than by hand, /marketplaces/lisskins covers how LIS-Skins is watched, what the export lag means in practice, and exactly where the coverage stops. CSWatcher is not affiliated with or endorsed by LIS-Skins.