Video Poker

The model used here

The player receives five cards from a standard 52-card deck with no wild cards. They choose any subset to hold, then replace the discarded cards from the remaining 47 cards. All five initially dealt cards stay out of that draw, including the discarded cards. Replacement cards are drawn without replacement.

This is a specified fair-deal model, not a certification of every electronic game. Read the actual product’s deck, draw mechanism, wild-card, bonus and multi-hand rules before applying a probability or strategy.

Follow a hand

1. Read the game and paytable. The name alone does not supply all the payouts or card rules.
2. Check the full stake. Multiply denomination by credits per hand and by the number of paid hands, then include any extra feature cost.
3. Deal and inspect all five cards. Suit, rank and cards that would be discarded can all matter to a hold decision.
4. Confirm the holds before drawing. Under this model, the selected cards stay and the others are replaced once from the remaining deck.
5. Read the final category and total return. A payout equal to the stake returns the spend; it does not create net profit.

Variants need separate calculations

No one return percentage is assigned to each of these broad names. A multi-hand feature may reuse initial cards, so the final hands should not automatically be treated as unrelated independent outcomes.

The complete 9/6 schedule in this example

The shorthand 9/6 describes only the full-house and flush payouts. The cited return also requires every other entry below, the specified deck/draw model and optimal holds for that schedule.

An ace can complete A–2–3–4–5 or 10–J–Q–K–A; the sequence does not wrap through Q–K–A–2–3. The final hand receives its applicable highest category under this paytable rather than adding every overlapping category.

Four fully specified schedule examples

In each row below, royal/straight flush/quads/straight/trips/two pair/high pair/nonpaying payouts are 800 / 50 / 25 / 4 / 3 / 2 / 1 / 0. Only the stated full-house and flush entries differ. Returns use optimal play for the respective complete schedule.

Source: published Jacks or Better payout and outcome-count tables. The figures are recalculated from those published weighted counts and rounded for display. This is not a new exhaustive optimization of every starting hand, a ranking of every available machine, or a promise that an abbreviated strategy reaches the optimum.

The source also lists other schedules with 9/6 full-house/flush entries but different royal, straight-flush or other payouts. Two numbers therefore cannot establish a game’s complete return. Nor does a familiar term such as full-pay mean the best possible offer across every product.

House edge and total action

Theoretical return is the expected total amount paid divided by stake under the specified model. House edge is 100% − theoretical return. On the 9/6 example, a fixed $100 of total action has about $0.46 expected net loss under optimal play. The stated 6/5 schedule instead has about $5.00 per $100.

Total action includes money wagered again after a payout. These expectations are not the amount deducted from every hand, the chance of any prize or a promise of what happens to a particular starting balance.

How a hold is compared

Five cards have 32 possible hold subsets, including keeping none or all five. For each hold, list its legal draws, classify each final hand and average the paytable returns. The largest average identifies the greatest expected return under that specific hand, paytable and draw model; ties between holds are possible.

The published analysis methodology explains how complete hand and hold calculations support a game return. The optimal 9/6 strategy reference includes detailed exceptions: even a card you plan to discard can change which future draws remain. A short list of familiar-looking hands is not a complete optimal strategy for every variant.

A checked one-card example

Suppose the initial hand is K♠ Q♠ J♠ 10♠ 2♠, using the complete 9/6 schedule above.

• Keeping all five cards locks in a flush returning 6 units per unit initially staked.
• Holding K♠ Q♠ J♠ 10♠ and discarding 2♠ leaves 47 equally likely replacement cards. The discarded 2♠ cannot be drawn again in this model.

These counts total 47. The expected total return for that hold is (800 + 50 + 6×6 + 6×4 + 9) / 47 = 919/47, or about 19.55 units. It exceeds the made flush’s 6 in this particular comparison. A nonpaying draw remains possible.

Now change only the royal award to a hypothetical 100-for-1, retaining all other payouts. The same draw counts give 219/47, or about 4.66 units, below the held flush’s 6. Among these two choices, the preferred hold reverses. This counterexample is not a claim about a particular machine’s offer.

These are conditional values after being dealt a strong hand, not the whole game’s theoretical return. The examples compare two specified holds; they are not a complete strategy chart or an excuse to apply the same decision to a different five-card hand.

Rare awards and session variation

The cited optimal 9/6 outcome counts put the royal frequency at about one in 40,391 hands under that model. Its reciprocal is an average frequency, not a schedule or a guarantee that playing that many hands produces a royal.

In a fresh independent-hand model with unchanged rules and strategy, previous missed royals do not make the next hand due. Within one hand, by contrast, knowing the dealt and held cards changes the remaining possibilities. A one-in-47 royal draw from the specific hand above is not the same probability as a royal across all initial hands.

An average return near 100% can coexist with many nonpaying hands and widely varying session results. There is no number of hands after which a person’s experience must equal the theoretical percentage.

Compare definitions in the game comparison.

• Jacks or Better paytables and weighted outcome counts: published optimal-play analyses for the complete schedules used here.
• Optimal 9/6 strategy and exceptions: detailed strategy scope, including the effect of discarded cards. This guide’s two-hold example is not a substitute for the full reference.
• Video poker analysis methodology: how starting hands, legal hold/draw choices and their values relate to a full-game analysis.

Checked September 5, 2026. The full-game returns use the cited counts; the 47-card draw and spending exercises are independently calculated. No universal study-time benefit, non-optimal-play loss range, strategy-aid permission, machine availability or session pace is asserted.