PRACTICAL GUIDES
CS2 cases, trading or marketplaces: how platform types differ
Compare mechanics, pricing, fees and withdrawal paths without invented ratings.
1. Identify each business model
Case opening sells a random outcome; a trading service quotes one item against another; a marketplace connects a seller and buyer at a price. One site may combine models, so compare the specific transaction rather than the platform label alone.
- Random outcomes are clearly identified
- The party setting the price is known
- Item ownership before settlement is understood
2. Calculate total cost, not the headline
For cases, review opening price and published odds; for trading, calculate the spread between incoming and outgoing valuations; for a marketplace, include sale, deposit and withdrawal fees. Compare the final money or skins after every deduction.
- Every stage fee is included
- The effective trading spread is calculated
- Case odds and possible items are reviewed
3. Verify liquidity and withdrawal path
A valuable item shown in a catalogue may not be available to withdraw now. Check stock, Steam trade lock, minimums, substitutions, cash redemption and KYC. On P2P markets, also account for sale time and cancellation risk.
- The item is available, not merely displayed
- Withdrawal method and timing are known
- KYC, limits and substitution rules are checked
4. Match the model to the task
A transparent trade quote may suit a predictable item swap; marketplace depth matters for a target purchase price; case opening remains random entertainment. Calling one type “best” without a user goal and verified facts is not useful.
- The transaction goal is defined first
- Randomness is not presented as a purchase
- The decision uses price, rules and availability
FAQ
Is a trading site always cheaper than a marketplace?+
No. Compare the current item price, spread and all fees.
Does case price equal expected value?+
No. Expected value requires probabilities for every outcome.
Can all sites be compared with one rating?+
A single score hides different models; task-specific verified facts are more useful.