CSBoard vs CSFloat

A CS2 skin is not one product. Two Karambit Fades at the same wear can differ in real value by a large multiple depending on where the fade sits, and two Case Hardened AK-47s at identical float can be worth wildly different amounts depending on how much blue lands on the playside. What a marketplace publishes per listing determines whether you can find that difference yourself or whether you have to paste every inspect link into a third-party checker.

Read that table as two different jobs. CSFloat answers "is this priced correctly?" — it hands you a number and a link you can send to someone else. CSBoard answers "is this the right copy?" — it hands you the two cosmetic percentages that decide whether a pattern-sensitive skin is a trophy or a filler.

The sticker gap runs the other way and matters more than people expect. On sticker-heavy AK-47s and AWPs, applied craft can be a large share of the asking price, and if a venue does not surface stickers as a listing field you are inferring value from the image. For anything where the craft is the point, CSFloat's listing data is the more complete one.

One structural difference worth knowing before you try to compare prices at all: CSFloat gives every listing a real permalink, and CSBoard does not in the same way. That sounds like a UI detail. It is not. A permalink is what lets you save a listing, share it for a second opinion, come back to it after a bank transfer clears, or prove to yourself what the price actually was. Without one, a CSBoard listing you liked is something you have to find again.

Identical items sit at different prices on different venues for reasons that have nothing to do with one market being "better".

• Seller fees. Fees are the largest single wedge between two quotes for the same skin. CSFloat charges sellers 2%. For comparison, BUFF163 is 2.5%, Skinport is 12%, and the Steam Community Market takes 15% — which is why a Steam price and a CSFloat price for the same item are not describing the same thing. CSBoard's fee structure is not something to take on hearsay: check it on CSBoard's own fee page before you model a resale, because a fee assumption you got from a comparison article is the fastest way to misprice an item.
• Liquidity and depth. A venue with fewer simultaneous listings of a given skin has a wider spread. Ask prices sit higher because a seller who is the only one holding that item can wait; but when a seller there does want out, the discount can be steeper than on a deep book, because there is no queue of comparable listings to anchor against. Thinner markets are more volatile in both directions, which is exactly why they are worth watching.
• Payment rails. Card, crypto and balance-based settlement each carry different processing costs, different chargeback exposure and different withdrawal friction, and every one of those gets priced into the ask. Two venues quoting the same nominal number are not offering the same deal if one settles instantly to a bank and the other settles to site credit.
• Trade locks. A tradable item and a Steam-trade-locked item are different assets with different prices, and lock status is a large part of why a listing looks cheap. Always establish the lock state before treating a discount as real — a week of holding risk is a real cost, and the market prices it.
• Regional access and currency. Availability, KYC requirements and the display currency all shift effective prices. A price shown in your local currency after conversion is not the price you will be paid in if you later sell into a market that settles in something else.
• Cosmetic percentages are priced in unevenly. This is the CSBoard-specific one. Because CSBoard publishes fade % and blue % openly, buyers there can see exactly which copies are premium — which tends to push the good percentages up and, more usefully, means the ordinary percentages are not accidentally overpriced. On venues where those numbers are hidden, listing prices cluster around a generic average, so exceptional copies are sometimes underpriced and mediocre ones overpriced. Transparency does not make a market cheap. It makes it less random.

A price you cannot realise is not a price. That is the whole answer, and most price comparisons ignore it.

Before you treat any listed number as the value of a skin, ask three questions. What currency would you actually be paid in? If the answer is site balance rather than money in a bank account, the number is a store credit, not a valuation. Which market can you actually sell into? If your realistic exit is a 12%-fee or 15%-fee venue, the sale price you should be modelling against is the one net of that fee, not the headline. What account would you need? Verification tiers, regional restrictions, trade-hold status and Steam account age all determine whether a given venue is genuinely available to you as a buyer or a seller. A quote on a market you cannot transact on is trivia.

Given that, the sensible discipline is to pick one liquid, permalinked, low-fee venue as your yardstick and measure everything against it — not because it is always the cheapest, but because it is the one number you can consistently check and act on. In practice that is CSFloat's Final estimate: it is published per listing, it is attached to something you can link to, and the fee to sell there is a known 2%. This is exactly how CSWatcher approaches it — it scores listings from other markets against CSFloat's Final rather than against each market's own internal notion of "value", because a discount is only meaningful relative to a price you could genuinely realise.

The failure mode to avoid: comparing a CSBoard ask to a Steam Community Market ask and concluding one is a bargain. Those two numbers describe different assets settling in different money at different fees. Compare like to like or do not compare.

CSBoard genuinely wins when the pattern is the product. If you are hunting a Karambit Fade above a specific fade percentage, or a Case Hardened where the blue coverage on the playside is the entire reason to buy, CSBoard is the one venue in this comparison where that criterion is a published number on the listing instead of something you derive from an inspect link. That collapses the search from "open fifty inspect links in a pattern checker" to "read the listing". For fade and blue hunting specifically, no amount of predicted-price sophistication elsewhere makes up for the marketplace simply telling you the number.

It also wins on friction. Browsing CSBoard requires no key and no login — you can look before you commit anything. That is a real advantage when you are price-checking rather than buying.

CSFloat wins in most other cases. If you want a predicted fair price rather than your own judgement, if the item's value is in its stickers, if you want a permalink you can save or share, if you want a known 2% seller fee when you eventually exit, or if you want the deepest book for mainstream skins — CSFloat is the stronger venue. It is also the one whose price other people will recognise, which matters when you are negotiating.

Two honest caveats about relying on CSBoard data. First, browsing without a key means listings carry no timestamp — you can tell that a listing is one you have not seen before, but not when it appeared, so "newest first" is not something you can establish from the data alone. Adding a free read-only API key fixes this: the official feed timestamps listings. Second, if you are automating any of this, CSBoard is reachable from desktop software only.

The practical answer for most buyers is not "pick one". It is: use CSBoard to identify the right copy of a pattern-sensitive skin, use CSFloat's Final as the yardstick for whether the asking price is sane, and settle wherever the fees, currency and trade-lock status leave you better off after costs.

If you want the mechanics of how CSWatcher watches CSBoard's published fade % and blue % and alerts on them, see /marketplaces/csboard. CSWatcher is not affiliated with or endorsed by CSBoard, and is not affiliated with or endorsed by CSFloat.