An average is not a session promise
RTP means return to player. It describes a theoretical return under a game's defined rules across extensive play. It does not say how many of your next ten attempts will win or what percentage of one deposit you will recover.
The Gambling Commission distinguishes the long-run figure from a single session. That is the key reading habit: keep a mathematical average separate from a personal forecast. A game can produce very uneven short sequences without those sequences explaining its theoretical model.
Use the same basis for house edge
For a fixed game model measured on the same stake basis, a 97% theoretical return corresponds to a 3% expected house share: 100 minus 97. This describes an expectation, not a charge taken visibly from every round.
As an illustrative calculation, 10,000 credits of total wagering at that model implies an expected net loss of 300 credits before considering separate promotions or other conditions. Your realised result can be different. The calculation neither recommends that volume of play nor makes an individual loss predictable.
A deposit and turnover are different
Turnover is the sum of the stakes you place. Reusing a returned amount creates another stake and adds turnover without adding a new deposit. Counting only money first paid into an account misses that repeated exposure.
Keep three figures separate when reading a session: money deposited, total stakes and final money returned or remaining. A screenshot of a balance alone cannot reconstruct all three. Similarly, a gross round return includes the original stake where the game's settlement convention says it does; it is not automatically net profit.
Compare identified games and settings
Use RTP as one technical comparison, alongside the exact title, version and available rules. Do not replace a published range with its highest number or borrow a sequel's figure for the original.
RTP alone does not describe volatility, the chance of a particular multiplier or how long a balance might last. A useful comparison states what is known and leaves unpublished values blank. It should help you understand the product without turning a percentage into a reason to spend more.
Sources
Product information is attributed to the linked source. An external demo link is not an interactive test.