ShareA 99% Return to Player in video poker does not mean that 99% of hands pay. The percentage describes the average amount returned relative to the amount staked, under the website’s stated conditions. How often a completed hand receives any award is a separate calculation. A game can have a return close to 100% while more than half its hands produce no award.
Return to player, usually shortened to RTP, is an average proportion of total stakes returned over time. Each possible result contributes according to both its probability and its payout. Paying-hand frequency is a different measurement: every hand that wins anything (even a small amount) counts toward this frequency.Total stakes means the amount wagered across the hands being measured, including credits wagered again after being returned. It is not simply the amount deposited before a session. A percentage of turnover and a percentage of an opening balance describe different things, even when both are displayed as percentages.
Start with the complete paytable
In video poker, the final five cards are assessed against a paytable, the list of awards for qualifying hands. Hand names alone cannot establish an RTP for Cafe Casino video poker games. The award attached to each qualifying hand and the choices made before the draw both enter the calculation. A stated return therefore needs a named variant, its complete paytable, and the strategy assumed.
Even a familiar name, such as Jacks or Better, is insufficient to identify an RTP. Different paytables can attach different awards to the same named hands. Those changes can affect which cards the calculation assumes you retain. A strategy chosen for one paytable is not automatically appropriate for another.
Two percentages from one game
Let’s take the following numbers for standard 9/6 Jacks or Better: a full house pays nine units per unit staked, and a flush pays six. The calculation assumes the maximum-credit paytable, including a royal flush paying 800 units per unit staked, and optimal play.
The remaining awards per unit staked are 50 for a straight flush, 25 for four of a kind, four for a straight, three for three of a kind, two for two pair, and one for a qualifying high pair.
Under those conditions, theoretical RTP is approximately 99.54%. The probability of a nonpaying final hand is approximately 54.54%, leaving a paying-hand frequency of about 45.46%. Both figures describe the same game model. Their difference comes from what each calculation measures.
These probabilities refer to the completed hand after the draw. They cannot be replaced with a chart showing the frequency of poker hands on the initial five-card deal. Holding and replacing cards changes that distribution, so mixing the two sets of figures would produce a calculation for the wrong stage of the game.
To figure out paying-hand frequency, add the probabilities of every category receiving an award. A high pair and a royal flush each qualify as one paying result when they occur.
Calculating RTP adds another step: multiply each category's probability by its award, then add those contributions. A rare result with a large award can contribute appreciably to the average return, despite appearing infrequently. Counting paying hands alone removes that information.
In this reference table, a pair of jacks or higher returns one unit for each unit staked. That hand counts toward the 45.46% paying-hand frequency, even though it only returns the stake. An award and a net increase in credits are therefore different events.
What optimal play means
This video explains how players can approach video poker, and that different strategies have different RTPs, with the best approaches resulting in an RTP that is around 99%, depending on the version. The qualifications belong with the number: “certain variants” limits which games the statement covers, while “optimal play” specifies the decisions assumed.
Here, optimal play means choosing holds that maximize the average return for the applicable rules and paytable. It does not mean producing a paying result on every hand.
A completed result alone cannot reveal whether the preceding hold matched that strategy. The calculation evaluates the possible replacements before the draw; the eventual hand is only one of those possibilities. Understanding the strategy assumption helps explain why the published return is conditional.
A session measures something different
Repeated samples of the same process can produce different averages. Quanta's explanation of how averages themselves vary shows that individual results and averages across repeated samples follow certain patterns. Still, a theoretical average cannot specify the result of one video poker session.
A session's observed return can be calculated by dividing its total awards by its total stakes. That describes what happened during those hands. The theoretical RTP instead comes from the game's possible outcomes, and the strategy assumptions attached to them.
Likewise, a session's share of paying hands may not be 45.46%. Neither a short sequence of awards nor a run of nonpaying hands changes what the two published measures mean. To interpret a percentage, identify whether it counts returned credits or completed hands, then keep the specified paytable and strategy attached to it.
