18+ · INTERNATIONAL EDITION   /   TERMS BEFORE TEMPTATIONGambling should never be a financial plan. Get support ↗
CasinOffersThe offer-terms journal

GAME MATHEMATICS

RTP is not a refund

A long-run percentage cannot tell you what one deposit, one session or one unfinished bonus will return.

By CasinOffersResearch-based · No funded testingChecked 18 September 2026

THE POINT TO TAKE AWAY

RTP refers to long-run wagers under a game model. It is not a promise about a deposit, a session, a withdrawal or recovering previous losses.

The denominator is wagers, not deposits

Return to player is a long-run return measure expressed relative to wagers. Research on RTP communication describes that distinction; the measure is not a promise about an individual session. [1]

For a hypothetical game with a 96% theoretical RTP, the corresponding modelled house edge is 4% of turnover. The figures in this guide are teaching examples, not audited values for a named casino game. They do not mean that depositing 100 units produces a 96-unit withdrawal.

A deposit and total turnover can be very different numbers because a balance can be used for repeated stakes. Confusing those denominators is one reason a plausible-sounding calculation becomes misleading.

What the arithmetic can say

Suppose 1,000 units are wagered under a simplified model with a constant 4% expected house edge. The model’s expected net loss is 40 units. “Expected” describes a mathematical average, not a scheduled deduction or the most likely result for a particular short session.

The calculation does not tell you the distribution of outcomes, how often large prizes occur or whether the balance lasts long enough for the planned turnover. Those are separate questions. Replacing “expected” with “guaranteed” would be wrong even if the multiplication itself were correct.

A reader should not use this example to budget the amount needed to finish a bonus. It cannot identify a safe deposit or assure completion.

Previous losses do not create a personal refund entitlement

An advertised RTP is not an account-level target the game has promised to restore. Being below a theoretical percentage in one session does not entitle that account to compensating wins in a later session.

The fact that a long-run proportion is defined does not say how any particular sequence must end. A model can have a stable expectation while individual sessions vary substantially. Nothing in this explanation predicts a “due” win.

Chasing losses to return to an imagined percentage adds more wagers; it does not turn the percentage into a debt owed to the player. Gambling should not be treated as a way to repair financial losses.

Why a bonus calculator does not use RTP to promise value

Our wagering calculator models a contractual target: base × multiplier ÷ contribution fraction. It does not forecast money returned. Adding an RTP input could make the display look more sophisticated without answering whether the player completes the requirement or whether the terms allow the chosen play.

A single theoretical number cannot account for balance depletion, game-specific restrictions, caps, expiry and account conditions. Presenting a neat “profit” output from those missing pieces would suggest more certainty than the inputs support.

The tool therefore reports turnover and its assumptions. It leaves outcome prediction out rather than making an unsupported claim about likely bonus value.

Read the actual game information, but keep the limit

For a named game, use the information supplied for the actual version being offered, not a percentage copied from a different title or version. This guide has not audited any game’s settings or random outcomes.

Understanding a percentage can help you reject a false guarantee. It does not remove gambling risk or create a winning method. A decision not to play remains valid regardless of how attractive a published number looks.

If an article describes RTP as a guaranteed refund, a personal payout schedule or a way to recover losses, the problem is with the interpretation, not with your failure to play long enough.

Sources & scope

Public-source research checked on 18 September 2026. Operator statements are not hands-on testing. Each source retains its regional and evidence limits.

  1. Weiss-Cohen et al.: return-to-player information research (2025)Research abstract; not an individual game audit · Checked 2026-09-18
    Limited / indexed evidence