Video poker math becomes much easier when you separate four questions: what the paytable pays, how often outcomes occur, which hold has the best expected value, and how much total action you create. Most confusion comes from mixing those questions together.
Begin with one hand, not with RTP
RTP is a property of the whole game under stated assumptions. A player decision happens one hand at a time. The bridge between the two is expected value.
Suppose a deal offers several legal holds. For each hold, the remaining deck creates a set of possible draws. Those draws lead to final hands with known payouts. The average value of those possible outcomes is the expected value of the hold.
Expected Value of a Hold =
Sum of (probability of each resulting hand × payout of that hand)The highest-EV hold is the mathematically preferred play. It can still lose on the next draw. Expected value ranks decisions; it does not promise outcomes.
That distinction is the foundation of hold or draw decisions and the video poker analyzer.
Build the whole game from weighted payouts
Once a strategy defines how every possible deal is played, we can estimate how often each final hand occurs. Multiply each final-hand probability by its payout and add the contributions.
RTP = Sum of (final-hand probability × final-hand payout)
House Edge = 1 - RTPThe Wizard of Odds Jacks or Better tables show these return contributions by hand. This is more informative than seeing one headline percentage because it reveals where the return comes from.
A rare royal flush can contribute meaningful return even though most sessions never see one. Frequent hands such as pairs, two pair and trips contribute in smaller pieces but occur much more often. The final percentage is the sum of the entire distribution.
Change the paytable and the weighted sum changes immediately. Cut a flush payout, and flushes contribute less. Raise a quad category, and that category contributes more. Strategy can also shift because the relative value of candidate holds changes with those payouts.
Work a dollar example from bet to expected cost
RTP percentages feel abstract until you connect them to coin-in.
Assume a player bets $1.25 per hand and plays 500 hands.
Coin-In = $1.25 × 500 = $625If the theoretical RTP under the chosen strategy is 99.54%, then:
House Edge = 1 - 0.9954 = 0.0046, or 0.46%
Expected Loss = $625 × 0.0046 ≈ $2.88The $2.88 is a long-run average cost of that amount of action under those assumptions. It is not a forecast. The actual session might win hundreds, lose hundreds, or finish close to even.
That gap between expectation and result is where variance enters.
Variance explains why a good percentage can feel terrible
A game can have a high RTP and still produce severe short-term swings. Video poker returns are not paid in smooth fractions of a percent. They arrive through discrete hands, some frequent and small, some rare and large.
If a meaningful share of return depends on royals or premium quads, many ordinary sessions can run below the long-term average simply because those rare events did not occur. A royal-heavy game can therefore feel expensive for a long time and then produce a huge jump in one hand.
Variance measures the spread of possible results around the average. It is not the same as house edge.
- House edge describes the long-run theoretical price.
- Variance describes how widely results can swing around that expectation.
- Bankroll risk asks whether your money can survive those swings at the chosen denomination and pace.
See video poker RTP vs variance and video poker bankroll risk when those distinctions become more important than the headline return.
Strategy error is a mathematical change, not just a skill issue
Published return figures normally assume a defined strategy, often optimal or near-optimal play. If a player repeatedly chooses lower-EV holds, the actual average return falls.
Imagine two holds:
Hold A EV = 1.48 credits
Hold B EV = 1.41 creditsChoosing Hold B once costs 0.07 credits in expectation. Nothing forces that hand to lose. But repeated inferior choices create an accumulated strategy penalty.
This is why “I played a 99% machine” can be misleading. The machine may have a 99% theoretical schedule under correct play, while the patron’s own decisions produce a lower personal return.
The same applies when a strategy chart is copied from the wrong variant. Jacks or Better, Deuces Wild and Double Double Bonus can present similar-looking hands but assign different values to the holds.
Paytable math in a real hand
Consider A♠ K♠ Q♠ 8♥ 3♦ in Jacks or Better.
Three suited high cards create a royal-flush route. The mathematical question is not whether a royal would be exciting; it is how the average return of holding A♠ K♠ Q♠ compares with all competing holds.
Now move the same visual pattern into a bonus or wild-card game. Payout rows and hand possibilities change, so the hierarchy can move. This is why strategy is inseparable from paytable math.
The Wizard of Odds optimal 9/6 Jacks or Better strategy demonstrates the connection between a specific paytable and its strategy. The video poker summary tables show how broad game families contain multiple return schedules.
Five numbers that answer different questions
| Number | What it tells you | What it does not tell you |
|---|---|---|
| Payout | Reward for one final hand | How often that hand occurs |
| EV of a hold | Average value of one decision | Whether the next draw wins |
| RTP | Long-run return under stated assumptions | Your short-session result |
| House edge | Long-run casino advantage | Size of short-term swings |
| Coin-in | Total action | How much cash you initially inserted |
Keeping those definitions separate prevents most beginner math errors.
A sixth number, hands per hour, matters because it converts theoretical cost into a pace:
Average Theoretical Loss Per Hour =
Hands Per Hour × Bet Per Hand × House EdgeAgain, that is an average price of action, not a prediction of the next hour.
Why the casino cares about coin-in more than your starting bill
From the casino side, video poker is measured through meters, coin-in, paytable configuration, theoretical hold and actual results. A player can insert $100 and wager far more than $100 over a session by recycling wins. The economic exposure is tied to repeated wagers.
Marketing can use coin-in and theoretical value for player offers. Slot management can compare banks and denominations. Accounting reconciles game meters and payouts. Surveillance and technicians deal with jackpots, disputes, machine events and integrity issues.
This operating view explains why comps should not be treated as free money. If a player creates $2,000 of extra coin-in to earn a small benefit, the extra theoretical cost can exceed the benefit depending on the game and offer.
Common math claims that should trigger caution
“The machine owes me because I am below RTP.” RTP is not a running debt to the player. Past results do not force near-term correction.
“A 99% game means I lose 1% of my bankroll.” The percentage applies to total action, not simply your opening bankroll.
“A correct hold should win more often.” A correct hold maximizes average return. It may prefer a lower-frequency but higher-value outcome over a more frequent small result.
“A bigger royal means a better game.” Only the full weighted paytable answers that.
“Bad strategy only affects a few hands.” Even small EV leaks accumulate when repeated at high speed.
Use the right tool for the right math question
Use the video poker analyzer to compare holds. Use the house edge calculator to convert return into edge. Use the expected loss calculator to connect edge with coin-in. Use the variance simulator to understand possible swing ranges rather than to predict a result.
Then continue to video poker expected value, expected value of a hold, and video poker expected loss per hour. Those pages take the four core ideas here—weighted outcomes, strategy, total action and variance—and examine them separately.