TF2 trading research · Beginner friendly
Are TF2 Trade-Ups Profitable?
Some TF2 trade-ups can have positive expected value at particular prices. That is a conditional estimate, not a guaranteed profit, and the result changes when the prices or available listings change.
Calculate expected profit before buying
Add up the cost of the ten inputs. For every possible output, multiply its net selling value by its probability. Add those weighted values together to get expected value (EV). Expected profit is EV minus input cost; ROI is expected profit divided by input cost.
Use what you could realistically receive for an output. A seller asking 10 keys does not mean a buyer will pay you 10 keys immediately. On this site, backpack.tf research uses eligible sell listings for input costs and eligible buy orders for output liquidation. Steam mode uses asks for inputs and after-fee bids for outputs.
Worked example: positive ROI with an 80% chance of losing
Illustrative example only: a contract costs 10 ref and has five equally likely outputs. Four would sell for 5 ref each and one for 40 ref. EV is (4 × 5 + 40) ÷ 5 = 12 ref. Expected profit is 2 ref and ROI is 20%.
However, four of the five outcomes lose 5 ref. The chance of profit is only 20%. A 20% ROI is not a 20% chance of profit, and running the contract five times does not guarantee you receive every output once. Simulations demonstrate this variation; they do not predict your next result.
Research with your budget
Find a ten-item contract
Compare quoted costs, expected returns and loss risk. These are modeled recipes, not reserved stock or guaranteed profits.
Free access includes one positive-EV preview when available, plus non-profitable practice routes. Other profitable recipes remain premium. Prices are rechecked in the main calculator.
Why a profitable-looking trade-up can turn unprofitable
A low sell listing may be bought before you reach it. A high buy order may request paint, spells or Strange parts that your output would not have. An order may have enough demand for one item, not a large batch. These differences matter more than a headline percentage.
The board is a research shortlist, not an execution guarantee. A positive margin must survive the actual purchase cost, any selling fees and price changes. Unquoted outcomes mean profitability is unknown, not that those outcomes have zero value.
Does repeating a positive-EV trade-up guarantee profit?
No. Repetition does not remove the risk of loss or make stale quotes valid. Statistical averages are useful only when the probabilities and attainable prices are reasonably accurate. You also need enough capital to survive losing runs.
Record the cost of inputs you actually acquired and the proceeds from outputs you actually sold. Compare realized profit separately from modeled EV. Steam Wallet proceeds and keys/refined are different forms of value, so do not combine them as if both were withdrawable cash.