There’s a gambling mechanic built directly into CS2 itself: the Trade-Up Contract. Ten skins of one rarity go in, one skin of the next rarity comes out — no casino, no license, no provably fair tab, just Valve’s code and your inventory. It’s slower and nerdier than case sites, but played with a spreadsheet instead of hope, it’s one of the few formats where the player can genuinely control the math.
The mechanics
Feed ten skins of the same rarity tier (e.g., ten Mil-Spec blues) into a contract and you receive one random skin of the next tier up (a Restricted purple) from the collections your inputs came from. The outcome pool is determined by the input mix: if all ten inputs are from one collection, the output is from that collection; mixing collections spreads the odds proportionally. Crucially, the output’s float (wear) is the average of your input floats — and this is where the game is actually played.
Float engineering: the whole skill
A Factory New output (float under 0.07) can be worth 5–20x its Battle-Scared sibling. Since output float mirrors input average, buying ten cheap low-float inputs lets you “engineer” a Factory New result. The community standard: keep your input average below (target float cap ÷ scaling) — in practice, average float under 0.07 for FN outcomes, with each collection having its own float ranges to account for. This is arithmetic, not luck, and it’s the only corner of skin gambling where diligence directly manufactures value.
The expected value reality
Every trade-up has a calculable EV: sum of (probability of each possible output × its market price) minus cost of ten inputs. Three honest observations:
- Most trade-ups are negative EV — if a profitable contract were obvious, bots would arbitrage it within hours (they do; margins compress fast).
- Profitable windows appear — after case drops, collection updates and market panics, input prices lag output prices for days. The edge goes to people with spreadsheets and price alerts, not vibes.
- Variance is brutal at the top — the profitable trade-ups usually hinge on hitting one of several possible outputs. A 30% chance at the money skin means seven-in-ten contracts lose, at volume, on purpose.
Trade-ups vs case sites
Compared to case opening (30–60% expected loss on official cases, 5–20% on third-party sites), a well-chosen trade-up is dramatically better math — sometimes actually positive. The trade-offs: no instant gratification (sourcing ten good-float inputs takes time), no provably fair verification (Valve’s roll is a black box you simply trust), and marketplace price risk between buying inputs and selling the output. It’s a grinder’s format.
The sane workflow
- Pick a target output with real liquidity (check daily sales volume, not just price).
- Calculate required average input float for the desired wear.
- Source inputs patiently — buy orders below market, never panic buys.
- Compute EV before every contract; skip anything below ~90% return.
- Do it at volume or treat it as entertainment — single contracts are lottery tickets with extra steps.
Trade-up contracts are the skin economy’s quiet corner: no flash, no rake, no house — just arithmetic and patience. If the casino side of this hobby ever feels rigged, this is the format that answers to a spreadsheet instead of a server seed. And when you do hit something worth selling, our cash-out guide covers turning it into crypto at the lowest cost.