Video poker is the closest thing to a solved game in the online casino catalogue. The deck composition is fixed, the draw is a single decision node, and optimal play is computable exactly by enumerating all 32 hold combinations against the remaining deck. For US players in the Realtime Gaming ecosystem, the difference between a well-chosen paytable played correctly and a poor paytable played by habit exceeds five percentage points of return, which is larger than the entire house edge of most slot titles. This page covers paytable identification, exact return figures, variance profiles, and the limits of the auto-hold button.
Why the Paytable Is the Whole Game
Every video poker variant shares an underlying deck and hand ranking. What differs between machines with identical names is the payout schedule, and the schedule determines the return. The industry shorthand names a game by the per-coin payouts for its two most sensitive rungs. In Jacks or Better, those rungs are the full house and the flush, so “9/6” means nine coins for a full house and six for a flush on a one-coin bet.
| Variant shorthand | Full house (per coin) | Flush (per coin) | Return with optimal play | House edge |
|---|---|---|---|---|
| 9/6 Full Pay | 9 | 6 | 99.54% | 0.46% |
| 9/5 | 9 | 5 | 98.45% | 1.55% |
| 8/6 | 8 | 6 | 98.39% | 1.61% |
| 8/5 | 8 | 5 | 97.30% | 2.70% |
| 7/5 | 7 | 5 | 96.15% | 3.85% |
| 6/5 | 6 | 5 | 95.00% | 5.00% |
Each single-coin reduction in the full house rung costs about 1.1 percentage points; each reduction in the flush rung costs about 1.1 points as well. The game looks identical on screen, uses the same graphics, the same hand rankings, and the same optimal strategy with only marginal changes. The only visible difference is two numbers in the paytable grid. US-facing RTG lobbies carry a mix of schedules across their Jacks or Better, Bonus Poker, Double Bonus, and Joker Poker offerings, so the paytable must be verified inside the game rather than assumed from the title.
Two machines labelled Jacks or Better can return 99.54 percent and 95.00 percent respectively. The difference is three characters in the paytable grid and is worth more than four and a half percentage points, which no amount of skilled play can recover.
The 9/6 Jacks or Better Schedule
| Hand | 1 coin | 2 coins | 3 coins | 4 coins | 5 coins | Probability (optimal play) |
|---|---|---|---|---|---|---|
| Royal flush | 250 | 500 | 750 | 1,000 | 4,000 | 0.00248% |
| Straight flush | 50 | 100 | 150 | 200 | 250 | 0.0109% |
| Four of a kind | 25 | 50 | 75 | 100 | 125 | 0.2363% |
| Full house | 9 | 18 | 27 | 36 | 45 | 1.1512% |
| Flush | 6 | 12 | 18 | 24 | 30 | 1.1015% |
| Straight | 4 | 8 | 12 | 16 | 20 | 1.1229% |
| Three of a kind | 3 | 6 | 9 | 12 | 15 | 7.4449% |
| Two pair | 2 | 4 | 6 | 8 | 10 | 12.9279% |
| Jacks or better | 1 | 2 | 3 | 4 | 5 | 21.4585% |
| Nothing | 0 | 0 | 0 | 0 | 0 | 54.5430% |
The five-coin column is not linear. At one through four coins the royal flush pays 250 per coin; at five coins it pays 800 per coin. That single non-linearity is worth roughly 1.4 percentage points of return. Playing 9/6 Jacks or Better at less than max coins reduces the return from 99.54 percent to about 98.15 percent, which converts a near-break-even game into an ordinary one. The correct adjustment for a smaller bankroll is to lower the denomination and keep five coins, never to reduce the coin count at a higher denomination.
Deuces Wild and the Variance Question
Deuces Wild replaces the low-pair minimum with four wild cards, shifting the entire distribution. All four deuces are wild, the minimum paying hand becomes three of a kind, and the paytable adds rungs for five of a kind, wild royal flush, and four deuces. Full Pay Deuces Wild, historically the 25/15/9/5/3/2/2/1 schedule, returns 100.76 percent with perfect play, which is why it is essentially extinct online. The schedules actually deployed return meaningfully less.
| Schedule family | Five of a kind | Straight flush | Four of a kind | Optimal return |
|---|---|---|---|---|
| Full Pay (25/15/9/5/3/2/2/1) | 15 | 9 | 5 | 100.76% |
| Not So Ugly Ducks | 16 | 10 | 4 | 99.73% |
| Illinois Deuces | 12 | 9 | 4 | 98.91% |
| Colorado Deuces | 16 | 13 | 4 | 98.77% |
| Airport Deuces | 12 | 9 | 4 | 97.06% |
Variance is where Deuces Wild differs most from Jacks or Better. Expressed as the variance of the per-hand return in units of the total bet, 9/6 Jacks or Better sits near 19.5, while typical Deuces Wild schedules sit near 25 to 26, and Double Double Bonus Poker exceeds 40. The practical consequence is bankroll requirement. A player targeting a 95 percent probability of surviving 10,000 hands without ruin needs roughly two to three times the bankroll on Double Double Bonus that the same player needs on Jacks or Better, despite the two games having comparable headline returns.
| Variant | Optimal return | Approx. variance | Top hand cycle (hands) | Bankroll character |
|---|---|---|---|---|
| 9/6 Jacks or Better | 99.54% | 19.5 | 40,391 (royal) | Moderate; the reference game |
| 8/5 Bonus Poker | 99.17% | 20.9 | 40,000 approx. (royal) | Similar to Jacks or Better |
| Not So Ugly Ducks Deuces | 99.73% | 25.8 | 45,282 (natural royal) | Higher swing, more frequent quads |
| 9/6 Double Double Bonus | 98.98% | 42.0 | 40,799 (royal) | Severe; kicker hands drive return |
| Joker Poker (Kings or Better) | up to 100.65% full pay | 32 approx. | Varies by schedule | High swing, schedule dependent |
Optimal Strategy and the Cost of Approximation
Optimal video poker strategy is derived by computing the expected value of all 32 possible hold patterns for a dealt hand against the 47-card remaining deck and choosing the maximum. Human players use ranked hold charts that reproduce this result to within a few hundredths of a percentage point. The most common and most expensive strategy errors in Jacks or Better are these:
- Holding a low pair over four cards to a flush. The four-flush is the higher expected value hold; the low pair is not.
- Breaking a made flush or straight to draw at a royal. Correct only when four cards to the royal are present and the made hand is a flush, not a straight flush.
- Holding a kicker with three of a kind or two pair. Never correct in Jacks or Better; it costs expected value on every occurrence.
- Keeping three cards to a straight or flush with no high cards. Almost always inferior to holding a single high card or drawing five fresh cards.
- Ignoring penalty cards. When two hold patterns are close, the specific cards discarded can block an inside straight or a flush draw and flip the decision. Penalty card awareness is worth roughly 0.02 to 0.04 percentage points.
- Discarding a dealt straight flush draw with a low pair present. The four to a straight flush is a strong hold and outranks most low pairs.
Typical intuitive play returns roughly 97 to 98 percent on a 9/6 machine, giving away 1.5 to 2.5 percentage points to strategy error alone. Combined with a poor paytable, casual play on an 8/5 machine can run below 95.5 percent effective return, which is worse than most video slots in the same lobby.
Auto-Hold Logic: Useful but Not Authoritative
Most RTG video poker builds include an auto-hold or expert-hold feature that pre-selects cards after the deal. The implementation is a strategy table embedded in the game client, and it is important to understand what that means:
- Auto-hold is optimized for the specific paytable loaded. If the game engine ships a shared strategy table across multiple paytables, marginal hands can be misplayed on the non-default schedule.
- Auto-hold typically ignores penalty cards. Rank-based hold charts do not evaluate which specific cards are discarded, so genuinely close decisions can go the wrong way.
- Auto-hold does not manage bankroll or coin count. It will happily play optimally at one coin, where the royal flush bonus is not in play and the return is well below the headline figure.
- Auto-hold cannot be assumed to be a house feature that disadvantages the player. There is no mathematical reason for a certified build to embed a deliberately suboptimal chart, and lab math verification would flag a systematic deviation. The realistic risk is approximation, not sabotage.
The reasonable posture is to treat auto-hold as a fast default and override it on the recognized marginal cases: four to a flush versus a low pair, four to an inside straight with three high cards, and any dealt hand containing three or four cards to a royal alongside a made paying hand.
Max coins is not a betting style, it is part of the paytable. The royal flush jumps from 250 to 800 coins per coin wagered only on the fifth coin, and playing short-coin discards roughly 1.4 percentage points of return before a single strategy decision is made.
Session Math for US Players
Video poker’s low edge makes hourly cost calculation straightforward. At 600 hands per hour on a $1.25 total bet (five coins at 25 cents), turnover is $750 per hour. On 9/6 Jacks or Better at 99.54 percent, expected loss is about $3.45 per hour. On an 8/5 schedule at 97.30 percent, the same play costs about $20.25 per hour, nearly six times more for an identical-looking game. Add typical strategy error of two percentage points and the 8/5 figure rises past $35 per hour.
Two structural notes specific to the US offshore environment. First, video poker is frequently weighted at zero to twenty percent toward bonus wagering requirements or excluded entirely, precisely because of its low edge; the excluded-games clause must be checked before a bonus balance is active. Second, royal flush payouts on quarter and dollar denominations can exceed common weekly withdrawal caps, so the practical settlement timeline for a large video poker win may span several payment cycles.
Finally, the low house edge of a full-pay schedule is not an income opportunity. A 0.46 percent edge is still negative expectation, and the return is concentrated in a royal flush that appears roughly once per 40,391 hands. Play with a fixed bankroll, treat the hourly expected cost as the price of the entertainment, and use deposit and session limits. If play stops feeling optional, confidential support is available at all hours through 1-800-GAMBLER.