One honest caveat before anything else: this analysis covers the trade-up contracts our own calculator currently flags as profitable, roughly 7,800 of them. It is not a random sample of every possible trade-up, profitable or not, so we can't use it to answer "what share of all trade-ups lose money." What it can answer, cleanly, is a different and arguably more useful question: among the trade-ups that do work on paper, what do you actually get for your money?
The pattern
We split the dataset by input cost and looked at two things: the percentage return, and the actual dollar profit. Cheap contracts, the kind built from Consumer Grade junk worth a few cents each, have absurd-looking returns. The median Consumer Grade trade-up returns 122.8% on your money. That sounds like the headline. It isn't the real story. The median cost of a Consumer Grade contract is $0.06. A 122% return on six cents is nine cents of profit. Across every contract under $1 of input cost, the median dollar profit is $0.10. You could run a hundred of these and still not clear a coffee. Move up to contracts that actually put real money on the table, input cost above $20, and the percentage return collapses to a median of 14.6%. The correlation between how much you put in and what percentage you get back is -0.38: the more you spend, the worse your edge gets, consistently.
Where the real dollars are, the odds get worse too
Restricted-grade contracts, the tier that can output knives and gloves, are the only rarity where median profit in dollar terms is meaningful ($6.73 vs under $1 for every cheaper tier). They're also the tier with the lowest median success rate of any category, 55%, barely better than a coin flip. Look specifically at the highest-cost contracts, input over $20, and it gets sharper: 61% of them have a success rate at or below 55%. The median success rate for this group is 53.3%, against 65% for the dataset as a whole. The single best dollar-profit contracts in the whole dataset, a Galil AR | Metallic Squeezer trade-up projecting $58 of profit on a $52 stake, and a near-identical Nova | Yorkshire play, both sit at exactly 50% success rate. A literal coin flip, for the two best-paying trade-ups our calculator currently finds. StatTrak contracts tell a smaller version of the same story: median return of 15.0%, tightly clustered (very little variance compared to non-StatTrak), which is really just another way of saying the market has already priced most of the edge out of them.
What this actually means
There isn't a version of trade-up contracts where you get both a strong percentage edge and low variance and real dollar amounts, at least not right now, not in what's currently profitable. You get to pick two, usually one:
Meaningful edge, real money, and a coin flip. High success-rate variance, real capital at risk, real profit if it lands.
Meaningful edge, tiny stakes, and near-certain outcome. Consumer Grade contracts return over 100% with success rates around 79%, but the dollar amount is so small it's not really an investment, more a rounding exercise.
What's actually missing from the profitable set is the safe, well-paying middle. It isn't in this data because, structurally, it doesn't really exist: markets tend to price out free lunches, and trade-up contracts round-trip through Steam and third-party market fees that eat further into whatever edge remains once you account for actually cashing out.
If someone tells you a trade-up contract is "basically free money," ask which bucket it's in. If it's real money, check the success rate before anything else, because based on what's currently profitable, there's a good chance it's close to 50%.
