Skip to content
CSGO Top

Guides

CS2 Trade-Up Math, Explained Without the Hype

Most CS2 trade-up contracts are negative-EV. Here's the math, the pool sizing, and the exit-rate questions to ask before pulling the trigger.

By Marek H. · Senior Skins Editor
Published April 12, 2026 Updated May 21, 2026

Most players who run trade-ups in 2026 lose money. That’s not a controversial statement — it’s just what the math says when you compute the expected value of a typical contract against current Steam Market prices. This guide walks through the math, the float mechanics, and the small set of contracts that are still genuinely profitable.

The 10-input rule

A CS2 trade-up contract takes ten inputs of the same rarity and outputs one item one tier higher. The output is randomly drawn from the collections represented by the inputs, weighted by how many inputs come from each collection. That weighting is the first lever.

If 9 of your 10 inputs come from a single collection, you have a 90% chance of pulling from that collection’s higher-tier pool. The remaining 10% goes to the 10th input’s collection. Pure single-collection contracts (10/10) are simpler to analyze but rarer in practice because most collections don’t have ten viable inputs at the right price.

Float averaging

The output float is the average of all ten input floats, normalized to the destination skin’s float range:

output_float = (sum_of_input_floats / 10) × (max_float - min_float) + min_float

This is why picking inputs by float matters as much as picking them by price. A contract built from 0.0–0.07 inputs targets the Factory-New side of the destination skin’s range. A contract built from 0.15–0.38 inputs targets Field-Tested or Well-Worn output, which usually drops the destination value substantially.

Why most contracts are −EV

For a single-collection contract, expected value is:

EV = (1/N) × Σ (price_of_output_i) − total_cost_of_inputs − Steam fees

Where N is the number of higher-tier outputs in the collection. Most collections have one or two high-value outputs and three to five low-value outputs. The unweighted average across the pool is usually below the cost of ten priced inputs once you factor in Steam Market fees (~13% effective). That’s the structural reason trade-ups skew −EV.

When trade-ups are still +EV

A handful of niche contracts are still profitable in 2026:

  • Single-collection contracts where the worst output still exceeds the input cost. These are rare but they exist in older collections where input prices have decayed faster than output prices.
  • Float-targeted contracts on high-pattern skins where Factory-New outputs trade at a steep premium.
  • Sticker-led contracts where the output’s sticker value floor changes the math (this only works if your inputs already carry useful stickers).

What to ask before running a contract

  1. Is the destination collection priced where I think it is, today, on Steam? (Use Steam directly — third-party feeds lag.)
  2. Are all 10 inputs reachable at the float range I need, without paying premium pattern markups?
  3. After Steam’s 13% fee on the output sale, does the math still work?
  4. Do I have a realistic exit (buyer pool) for the worst-case output?

If you can’t answer all four, the contract is probably negative-expected-value and you’re paying for entertainment. That can be a perfectly fine reason to run it — just make sure you know that’s what’s happening.

A note on third-party trade-up sites

Most third-party trade-up sites operate at significantly worse house edges than Steam itself, often 20–40% versus Steam’s ~13% fee. The math we walked through here assumes Steam-native contracts. If you’re running trade-ups on a third-party platform, the EV is almost always materially worse, regardless of how the site presents its odds.

Frequently asked questions

Are CS2 trade-ups profitable in 2026?
Most aren't. The 10-input contract is mathematically negative-expected-value once you adjust for Steam Market fees and the 50%+ float weighting toward worse outcomes. There are still profitable trade-ups inside narrow collections, but they require careful float sourcing.
What's the difference between a "covert" and a "rare" trade-up?
Covert trade-ups consume 10 classified-tier inputs and output a single covert from one of the 10 inputs' collections. Rare trade-ups (mil-spec → restricted, etc.) follow the same pattern at lower tiers. The math is the same; only the prices change.
Why is float so important in trade-ups?
The output float is a weighted average of the 10 input floats, mapped onto the destination skin's float range. This means a contract designed around a specific output float (e.g. 0.07 for Factory New ceilings) requires deliberately picking input floats — not "any 10 inputs you have lying around".