How to Calculate Trade-Up EV in CS2

What is expected value in a trade-up?

Expected value, or EV, is what a trade-up is worth on average. Take every skin the contract could give you, multiply each one's value by the chance of getting it, and add them up. That total is the EV of the output. Subtract what you paid for the inputs and you have the EV of the contract.

If the number is positive, the contract makes money on average across many repeats. If it is negative, it loses on average. That is all EV says. It does not say you will win this time, and for most trade-ups you will not, because the average is held up by a rare good outcome.

This guide walks through the calculation by hand so you understand what the trade-up calculator is doing when it gives you a number. It assumes you know the contract rules, which are in how trade-up contracts work.

How do you calculate trade-up EV step by step?

Six steps. The order matters because the later ones change the earlier ones.

1. List the outcomes. For each collection among your inputs, list every skin of the output rarity. Those are your possible outputs.
2. Work out each outcome's probability. Each input has a 1-in-10 chance of being the chosen input (1-in-5 for a Covert contract). The output comes from the chosen input's collection, evenly across that collection's outputs. So an outcome's chance is (inputs from its collection ÷ total inputs) × (1 ÷ number of outputs in that collection).
3. Work out each outcome's float and wear. The output float comes from the average of the inputs mapped onto the output skin's range. That decides whether you get a Factory New or a Field-Tested, which can be a very different price. See float capping explained.
4. Price each outcome after fees. Use what you would actually receive for that skin at that wear, on the venue you will sell on, after its fee.
5. Multiply and add. Each outcome's after-fee price times its probability, summed. That is the expected output value.
6. Subtract the input cost. What you paid for all inputs, including any fees on buying them. The result is the contract's EV.

Then, before you act, add a seventh step: look at the chance of profit, not just the EV. A contract can be positive EV and lose 80% of the time. Whether you can stomach that is a separate question from whether the maths works.

Which prices should you use for inputs and outcomes?

The most common EV mistake is mixing price types. Use the price that reflects what actually happens to your money.

Two refinements. First, use recent sale prices, not listing prices. A skin listed at $100 that last sold at $85 is worth $85 for EV purposes. Second, use the price for the wear your output will actually be. If your float calculation says the output lands in Minimal Wear, price the Minimal Wear version, not the Factory New one you were hoping for.

Prices also move. A contract worth doing on Monday can be worth skipping by Friday. Recalculate before you buy the last input, not after.

How do fees change the EV?

Fees hit the output, and they hit it harder than people expect.

On the Steam Community Market, Valve's Community Market FAQ lists a Steam Transaction Fee of 5% (minimum $0.01) and a Counter-Strike 2 game fee of 10%. Both are added to the buyer's price on top of what the seller receives. So if the buyer pays $115 for your output, you receive about $100. And that $100 is Steam Wallet credit that cannot be withdrawn, which is fine if you are buying more inputs there and not fine if you want cash. The Steam fee calculator converts between buyer price and seller receipt.

On third-party marketplaces the seller fee varies by site and sometimes by item. Use the actual figure for the site you will use. The marketplace fee calculator applies whatever rate you enter, and our fee comparison page lists verified rates.

The effect on EV is straightforward. If every outcome loses 13% to fees, the expected output value drops 13%. A contract that looked like a 10% edge before fees is a 4% loss after them. Always calculate with fees in. A calculator that shows EV before fees is showing you a number you will never receive.

How does float change which outcomes count?

Float does not change which skin you get. It changes which version of that skin you get, and that changes the price.

Every outcome skin has its own float range. Your output float is the average of your inputs' normalised floats, mapped onto that range. So the same set of inputs can give a Factory New on one outcome skin and a Field-Tested on another, depending on their ranges.

For EV, that means you price each outcome at the wear your float calculation predicts for that specific skin. A common pattern:

• Outcome A has a range of 0.00 to 0.50. Your inputs average out to 0.05. Factory New. Price the FN version.
• Outcome B has a range of 0.10 to 0.70. Same inputs give about 0.16. Field-Tested. Price the FT version, which may be a fraction of the FN price.

The trade-up calculator does this per outcome. If you are calculating by hand, use the float cap calculator for each outcome's range. The full method is in float capping explained.

Buying lower-float inputs costs more. That cost has to be weighed against the higher output prices it unlocks, which is just another EV calculation layered on top.

What does a worked example look like?

All numbers below are made up to show the method. They are not real prices for any real skin.

Setup. You use 10 Restricted inputs. Seven come from Collection A, which has two Classified outcomes. Three come from Collection B, which has three Classified outcomes. Each input costs $4 to buy, so inputs total $40. You will sell the output on the Steam Market, where you receive roughly 87% of the buyer price.

Outcomes and probabilities.

Expected output value: $18.27 + $7.61 + $13.05 + $2.61 + $1.74 = $43.28.

EV of the contract: $43.28 minus $40 of inputs = +$3.28, about 8% on the input cost.

Chance of profit: You profit only when the output is worth more than $40 after fees. That is A1 (35%) and B1 (10%), so 45% of the time. The other 55% of the time you lose between $14 and $23.

That is what a positive EV trade-up usually looks like. It makes money on average, it loses more often than it wins, and one outcome does most of the work. If B1's price dropped from $150 to $110 tomorrow, the EV would fall to about minus $0.20 and the whole contract would flip negative.

What is variance and why does positive EV still lose?

Variance is how far single results spread around the average. Trade-ups have high variance because the outcomes are lopsided: many small losses, a few big wins.

In the example above, over 100 contracts you would expect roughly 45 wins and 55 losses, netting about $328 in total. But in any run of 10 contracts you could easily see 8 losses. That would cost you around $150 before the maths catches up. If you cannot afford that stretch, positive EV does not help you, because you will be out of money before the average arrives.

Three things follow:

• Bankroll matters more than EV. Only trade up with money you can leave in the game through a bad streak.
• Small edges are fragile. An 8% edge disappears with a modest price move or a fee you forgot. Bigger edges survive mistakes.
• One contract proves nothing. Winning one does not mean you found a good contract. Losing one does not mean it was bad. The odds show up over dozens.

Knife trade-ups since the 22 October 2025 update push this to the extreme. Five Coverts into a case's knife pool can have a dozen outcomes with prices spread across a wide range, and a random phase or pattern on top. The EV can be calculated, but the variance is large, and the pattern lottery cannot be priced at all.

How do you find trade-ups worth doing?

There is no list of guaranteed contracts, because any published good contract gets its inputs bought up until it is no longer good. But there is a method.

1. Start from an outcome you understand. A skin whose price you can verify, whose float range you know, and whose collection has few other outputs.
2. Look for collections with one expensive output and few cheap ones. Fewer outputs means each one is more likely, and one expensive one carries the EV.
3. Look for cheap inputs with the right floats. The input side is where the edge usually comes from. Bulk lots of low-float Restricted skins from old collections are the classic source.
4. Run it in the calculator with fees on. The trade-up calculator shows outcomes, probabilities, float-adjusted prices and EV in one screen.
5. Check the chance of profit and the worst case. Decide if you can live with the losing runs.
6. Recalculate right before you commit. Prices move. So do rules; Valve can change the contract at any time, as it did in October 2025 and May 2026.
7. Track your results. Keep a log of inputs, cost, outcome and sale. Over time, the log tells you whether your process works better than your memory does.

EV is the tool that turns trade-ups from gambling into a decision. It does not remove the risk. It just tells you which risks are priced in your favour, and only if you calculate it honestly, with real prices and real fees.