There’s a moment that catches most new players off guard. You sit down at a video poker machine expecting something like slots — press a button, watch the outcome, repeat — and then you realize the game is asking you to make a decision. Which cards do you keep? Which do you throw away? That decision changes your expected outcome. Not by much on any single hand, but over hundreds of hands it compounds into the difference between a game that returns 99.5 cents per dollar and one that returns 94 cents.
That gap is the whole story of video poker. The screen looks like a slot machine. The math doesn’t.
This guide covers how the game actually works, what the numbers on the pay table mean in real money terms, which variant to start with, how to make correct hold decisions, and why sessions go wrong even when you play perfectly. No invented statistics, no fluff — just what you need to play with your eyes open.
How Video Poker Works: The Mechanics
Video poker is based on five-card draw poker. You’re dealt five cards from a virtual 52-card deck (53 if the game includes a Joker). You choose which cards to keep — called holding — and discard the rest. The machine draws new cards to replace your discards, and your final five-card hand determines whether you win and how much.
That’s the entire structure. Every variant builds on it.
The deal is controlled by a random number generator, certified by independent testing laboratories. The RNG doesn’t remember your previous hands. It doesn’t know you’ve been playing for two hours. Each deal is a fresh draw from a fresh virtual deck. This matters because it means the probability of every possible outcome is fixed and calculable — and that’s precisely why strategy exists. Unlike slots, where the outcome is entirely predetermined by the RNG before you even press spin, video poker lets your hold decision change which outcomes are possible on the draw.
Here’s a concrete example. You’re dealt: J♥ J♦ 7♣ 9♠ 2♥. You have a pair of Jacks — the minimum qualifying hand in Jacks or Better. You hold both Jacks and discard the other three cards. The draw gives you: J♥ J♦ J♣ K♠ 4♦. Three of a Kind. You win.
Had you discarded one of the Jacks trying to chase something else, your probability of winning that hand would have dropped substantially. The decision mattered. It always does.
At a 25-cent denomination machine, a standard bet is five coins — $1.25 per hand. Players typically get through 400–600 hands per hour. That’s $500–$750 in coin-in per hour at that denomination. The percentage difference between a correct hold and an incorrect one, multiplied across that volume, becomes real money fast.
Hand Rankings and the Pay Table
Every video poker machine displays a pay table — the grid showing what each winning hand pays per coin wagered. Before you play a single hand, you should read this grid. It tells you the actual value of the game you’re about to play, and two machines with identical names can have completely different values.
The winning hands, from highest to lowest, in standard Jacks or Better:
Royal Flush — A♠ K♠ Q♠ J♠ 10♠. Five highest cards of the same suit. The top payout, typically 800-for-1 at maximum coins (more on this shortly).
Straight Flush — 7♦ 8♦ 9♦ 10♦ J♦. Five consecutive cards of the same suit. Usually pays 50-for-1.
Four of a Kind — K♠ K♥ K♦ K♣ 3♠. Four cards of matching rank. Pays 25-for-1 in standard Jacks or Better.
Full House — Q♠ Q♦ Q♣ 9♥ 9♠. Three of a kind plus a pair. This is one of the two numbers you use to classify a machine — more on that in a moment.
Flush — A♥ 9♥ 6♥ 4♥ 2♥. Five cards of the same suit, not in sequence. The other number that classifies the machine.
Straight — 5♠ 6♦ 7♣ 8♥ 9♠. Five consecutive cards, mixed suits. Usually pays 4-for-1.
Three of a Kind — 8♠ 8♦ 8♣ J♥ 2♦. Three matching cards. Pays 3-for-1.
Two Pair — A♠ A♦ 6♣ 6♥ Q♠. Two different pairs. Pays 2-for-1.
Jacks or Better (pair) — J♠ J♦ 4♣ 9♥ 2♠. A pair of Jacks, Queens, Kings, or Aces. Pays 1-for-1, meaning you get your bet back.
Now, about those two critical numbers. The pay table column for five coins is what you look at. Find the Full House row and the Flush row, then divide each number by five to get the per-coin payout. A machine that shows 45 for Full House and 30 for Flush at five coins is a 9/6 machine — 45÷5 = 9, 30÷5 = 6. This is the full-pay Jacks or Better standard, with an RTP of 99.54% under optimal play.
A machine showing 40/25 at five coins is an 8/5 machine. That sounds close. It’s not. The RTP drops to 97.3% — a difference of 2.24 percentage points. At $1.25 per hand and 500 hands per hour, that’s a swing from roughly $2.88 in expected loss per hour to roughly $14.38. Same game name, very different financial proposition.
Always check these two numbers before sitting down.
What RTP Actually Means in Practice
Casinos express the advantage on every game as a percentage of each bet they statistically keep over time. The flip side is the percentage returned to players — the RTP figure. For 9/6 Jacks or Better with optimal play, that’s 99.54%, meaning the house keeps 0.46 cents of every dollar wagered on average.
To put that in context: European roulette, often cited as a player-friendly table game, carries a 2.7% house edge (97.3% RTP). Most online slots run between 94% and 96% RTP. A 9/6 Jacks or Better machine with a player using correct strategy sits well above both.
Here’s the practical version of that comparison. Say you wager $100 total — across 80 hands at $1.25 each. On 9/6 JoB with optimal play, your theoretical expected loss is $0.46. On a 96% RTP slot, that same $100 in coin-in produces an expected loss of $4.00. Same money, same time. One costs you nearly nine times more on average.
The caveat — and it’s a real one — is that RTP is a long-run average, not a per-session guarantee. This is where variance enters the picture, and it’s the thing most guides skip over.
Video poker has concentrated payouts. The Royal Flush in Jacks or Better accounts for roughly 2% of the game’s total expected return. That single hand, occurring on average once every 40,000 hands — about 67 hours of play at 600 hands per hour — carries an outsized weight in the math. In any session where you don’t hit a Royal, you’re working with an effective RTP closer to 97.5% from all the other hands combined. That’s why sessions end negative even with perfect decisions. The math is working correctly; you just haven’t collected the Royal’s contribution yet.
This isn’t a flaw. It’s how the game is structured. Understanding it prevents you from concluding that strategy doesn’t work after a bad session.
The Five-Coin Rule and Why It’s Non-Negotiable
The single most important mechanical rule in video poker is also the most misunderstood. Maximum coins — five coins — must always be bet. Not because it changes the probability of your hands. It doesn’t. The cards fall the same way regardless of how many coins you put in. The reason is the Royal Flush payout structure.
On most machines, betting one through four coins pays 250-for-1 on a Royal Flush. Bet five coins, and that jumps to 800-for-1. Every other hand pays proportionally the same per coin across all bet levels. The Royal’s payout is the only one that doesn’t scale linearly — and it scales dramatically upward.
Work through the numbers: at one coin, a Royal on a 25-cent machine pays $62.50 (250 × $0.25). At five coins, that same hand pays $1,000 (800 × $0.25 × 5). You wagered five times as much, but you received 16 times the one-coin payout. The Royal Flush is mathematically subsidizing the entire five-coin bet relative to four-coin play.
Playing four coins instead of five effectively drops your RTP by over 1.5 percentage points on a 9/6 machine — erasing most of the advantage that made video poker worth playing in the first place.
If the denomination feels too high to bet five coins comfortably, the solution is simple: move to a lower denomination and bet five coins there. A 10-cent machine at five coins is $0.50 per hand. A 25-cent machine at five coins is $1.25. Play at the level where max coins fits your session budget, never at a level where you’re shaving coins to make it fit.
How to Make the Right Hold Decision
Strategy in video poker is about expected value — which hold decision returns the most money averaged across every possible draw outcome from that position. There’s a mathematically correct answer for every hand you can be dealt. You can look it up on a strategy chart while playing online, and you should. Using a strategy card isn’t cheating; it’s the same as a blackjack player using basic strategy — entirely permitted and entirely rational.
For Jacks or Better, the hold priority works from strongest made hands down to drawing hands. The logic: a hand that’s already a winner stays a winner on draw (though it can improve), while drawing hands are probabilistic bets that your incomplete hand completes.
A few decisions that cost players money most often:
Low pair versus four to an open-ended straight. Say you hold 8♦ 8♣ 7♥ 9♠ 5♦. You have a pair of Eights (a made hand, not a qualifying hand — below Jacks or Better’s minimum) and four cards to an open-ended straight needing a 6 or a 10 to complete. The straight draw looks appealing. Hold the pair. A low pair improves to a qualifying hand or better on the draw significantly more often than a four-card straight draw completes. The pair gives you trips, two pair, a full house, or quads as possible draw outcomes — the straight gives you a straight or nothing. The pair wins this comparison.
Low pair versus four to a flush. Same logic. A four-card flush draw completes roughly one in five times. A low pair has higher expected value because the number of ways it can improve to a paying hand — including quads — outweighs the flush’s single path to a win.
Three to a Royal Flush versus a low pair. This one surprises people. If you have three cards to a Royal (e.g., J♠ Q♠ K♠ plus two unrelated cards), and one of those unrelated cards pairs one of your royals (say you also have a J♦), the three-to-a-Royal hold is still ranked above the low pair in Jacks or Better strategy. The Royal draw’s potential payout is large enough, and the partial hands that come from a three-to-Royal draw are varied enough, that it edges out the low pair in expected value.
When to hold nothing. If your five dealt cards contain no pair, no flush draw of three or more suited cards, no connected straight draw of three or more, and no face cards — discard everything and draw five new cards. This happens. It’s the correct play.
The priority list for Jacks or Better has roughly 30 tiers from Royal Flush at the top to “discard all” at the bottom. You don’t need to memorize all of them to play well. The decisions that most players get wrong are the contested ones — pair vs draw scenarios — and those are exactly the ones worth drilling with a free training app before committing real money.
Choosing Your Starting Variant
If you’re new to video poker, play Jacks or Better first. Not as a rule of thumb, but as a mathematical recommendation. Here’s why it’s the right starting point.
Jacks or Better has the most thoroughly documented strategy of any variant. The pay table is easy to classify — those two numbers, Full House and Flush, tell you almost everything you need to know. The variance is the lowest of all common variants; the standard deviation sits around 4.42, meaning swings between big pays are narrower than in most alternatives. And the minimum qualifying hand — a pair of Jacks — is straightforward. You always know whether your hand pays.
Once Jacks or Better strategy becomes instinctive, you can move to other variants. But “move to other variants” means learning new strategy, not just applying JoB decisions to a different game. Each variant shifts the hold priorities in specific ways.
Deuces Wild replaces the four 2s in the deck with wild cards — cards that substitute for any other card. This sounds like a pure advantage, and in terms of qualifying hand frequency, it is. But the pay table adjusts accordingly: you need at minimum three of a kind to win anything. A pair of Aces doesn’t pay. The strategic implications are significant. You always hold any deuces regardless of the other four cards. Natural Royal Flush (no wild cards used) is the premium hand. Playing JoB strategy on Deuces Wild will lose money faster than random play in some spots — the games are that different.
Bonus Poker is built on the JoB framework but adds premium payouts for certain four-of-a-kind hands. Strategy adjusts to be slightly more aggressive about chasing quads in specific situations. Variance is moderately higher than JoB — standard deviation around 5.08 on Deuces Wild (as a reference point for what higher SD means in practice) — meaning longer cold stretches between big wins.
Double Double Bonus Poker takes the quad premiums further and adds a kicker mechanic: four Aces with a 2, 3, or 4 as the fifth card pays at a dramatically elevated rate compared to four Aces without a qualifying kicker. The variance on DDB is among the highest of common variants, with a standard deviation around 6.48. That’s a meaningful difference from JoB’s 4.42 — in practical terms, your session swings are much larger, and you need more bankroll to survive equivalent playing time at the same stake.
For a first-time player: Jacks or Better, 9/6 pay table, five coins. Learn the hands, learn a few of the key hold decisions, and get comfortable with how the game moves before adding strategic complexity.
Bankroll Planning That Makes Sense
Generic advice says to bring 200–300 betting units. That’s not wrong, but it doesn’t tell you much without connecting it to time and variant. Here’s a more practical framing.
At a 25-cent denomination, five coins per hand costs $1.25. Playing 600 hands per hour, your coin-in is $750 per hour. For a four-hour session, that’s $3,000 total wagered. On 9/6 JoB, your expected loss on $3,000 in coin-in is $13.80 (at 0.46% house edge). That sounds fine — but variance means the actual session result spreads around that expected value.
The standard deviation on a hand of JoB is 4.42 betting units. Across 2,400 hands at $1.25 per hand, the formula for estimating the 95th percentile swing is: 2 × 4.42 × √2,400 × $1.25. That works out to roughly $540. So 95% of four-hour sessions on 25-cent 9/6 JoB should fall within $540 of the expected result. A session that ends $300 down isn’t evidence of bad play; it’s normal variance.
On DDB with SD 6.48, the same calculation gives you approximately $790 as the 95th percentile swing. Same stake, same time, same denomination — nearly 50% more bankroll exposure due to the variant’s structure. This is the practical cost of chasing bigger quad payouts.
A session stop-loss at 50–60% of your planned bankroll is a reasonable discipline. If you’ve burned through more than half the session budget, continuing isn’t more likely to recover losses than quitting — the math doesn’t change — but variance won’t reverse a bad session just because you keep playing. Walking away preserves bankroll for a future session where variance runs the other way.
On the winning side: if you’re up 30–50% of your session bankroll, locking in some profit by reducing session length is a valid choice. The game’s expected return doesn’t change based on whether you’re ahead or behind. Setting a win goal doesn’t improve your mathematical outcome; it prevents variance from giving back a good session result.
Multi-Hand Video Poker: Lower Variance, Same EV
Many online casinos offer 3-play, 5-play, and 10-play video poker. The mechanic: you make one hold decision, and the machine deals an independent draw for each hand simultaneously. The expected value per hand is identical to single-play — the underlying math doesn’t change. What changes is variance.
Playing five hands at $0.25 per hand (total bet $1.25, same as a single-hand 25-cent five-coin bet) produces a smoother result curve than single-hand play at the same total stake. Five independent draws average out the outcomes, reducing the swing around your expected return. You won’t have the single-hand experience of hitting one Royal that defines your session; but you also won’t have the single-hand experience of getting nothing for three hours straight.
For players whose primary concern is session duration and variance management, multi-hand play at a lower denomination-per-hand is a reasonable approach. The tradeoff is that you can’t replicate the same total coin-in on multi-hand at a lower denomination without significantly more hands per decision — the math of the deal stays constant.
One practical note: on multi-hand video poker, the pay table per hand should be the same as the single-hand version of the same game. Verify this before playing — some platforms offer multi-hand versions with different (usually worse) pay tables than their single-hand counterparts.
Reading Pay Tables Quickly
You don’t need more than 30 seconds to evaluate a video poker machine’s value. Here’s the process.
Look at the pay table displayed at the top of the screen. Find the five-coin column — the rightmost one in most layouts. Locate the Full House row and the Flush row. Divide each number by five.
If the Full House row shows 45 and the Flush row shows 30: 45÷5 = 9, 30÷5 = 6. This is a 9/6 machine — the full-pay Jacks or Better standard. Play it.
If the Full House row shows 40 and the Flush row shows 25: 8/5. Short-pay. RTP drops to 97.3%. Factor that into your session expectations, or look for a better machine.
If the Full House row shows 35 and the Flush row shows 25: 7/5. Avoid. The house edge here makes video poker’s advantage over slots nearly disappear.
Reference benchmarks worth memorizing:
- 9/6: Full-pay Jacks or Better — 99.54% RTP with optimal play
- 8/5: Short-pay Jacks or Better — 97.3% RTP
- 10/7: Full-pay Double Bonus — approximately 100.17% RTP with optimal play
- 15/9: Full-pay Bonus Poker — approximately 99.17% RTP
Also check the Royal Flush payout at five coins. It should show 4,000 (which is 800-for-1 × 5 coins). If the five-coin Royal pays the same rate as the one-coin Royal scaled up — that is, 1,250 instead of 4,000 — the max-coin bonus is absent, and you’re playing a materially worse game regardless of the rest of the pay table.
On progressive jackpot machines, the Royal’s standard payout is replaced by an accumulating jackpot. The question isn’t whether the progressive is large — it’s whether the jackpot amount brings the effective contribution of the Royal Flush back up to its fair-value level relative to the 40,000-hand frequency. For standard Jacks or Better, the Royal contributes approximately 2% to total RTP. If the pay table is otherwise identical to 9/6 JoB but the base Royal pays less, the progressive needs to compensate for the shortfall before the game returns to full-pay value.
Four Mistakes That Cost Money
Playing fewer than five coins. This one’s covered above, but it deserves emphasis because it’s the most common and most costly error. Never bet four coins on a machine configured for five.
Sitting down without checking the pay table. Every major online casino and mobile platform runs multiple video poker variants, often at the same denomination. A 9/6 Jacks or Better and an 8/5 Jacks or Better can sit side by side. Players who don’t check end up paying a 2%+ premium for no reason.
Breaking up a made straight or flush to chase a Royal Flush. If you have a straight — say 8♠ 9♦ 10♣ J♥ Q♠ — the hand is already worth 4-for-1. Breaking it to chase a Royal Flush (if some of those cards are suited and consecutive toward a Royal) is only correct in very specific situations, and it’s almost never correct when you hold a complete straight or flush. The made hand’s EV beats the draw hand’s EV in most cases. Check a strategy chart before making this discard.
Concluding that strategy is wrong because a session went badly. Video poker’s variance means that correct decisions regularly produce losing sessions in the short run. A player using perfect JoB strategy can run $400 negative in a four-hour session and have done nothing wrong. The strategy improves expected return over time. It doesn’t guarantee any particular session. If you walked away from a session down, the question to ask isn’t “did the strategy fail?” — it’s “did I make the correct hold decisions?” Those are different questions with different answers.
The Role of Loyalty Programs
At online casinos, cashback programs and reload bonuses interact with video poker in a way they don’t with most other games. Because video poker’s house edge is so small on full-pay machines, even a modest cashback rate changes the effective economics noticeably.
A game with 98.5% RTP — a slightly short-pay Jacks or Better, for instance — combined with a 0.5% cashback program produces an effective combined return of approximately 99%. The game that looked mildly unfavorable becomes close to neutral. At 1% cashback, it crosses into neutral or slightly positive territory depending on the specific pay table.
The practical calculation: estimate your total coin-in for a session (hands per hour × bet per hand × hours), multiply by the cashback percentage, and that’s the dollar value of the program. At 500 hands per hour × $1.25 per hand × 3 hours = $1,875 in coin-in. A 0.5% cashback rate returns $9.38. On a game where your expected loss is $26 at 1.4% house edge, the program reduces net expected loss to around $17.
This doesn’t make a bad pay table good. A 6/5 Jacks or Better machine has such a poor return — around 95% RTP — that no reasonable loyalty program compensates. But for near-full-pay games, programs push the economics further in the player’s direction.
One thing to watch: most casino bonuses restrict video poker’s contribution to wagering requirements, often to 10–20% of slots’ contribution rate. Before using a deposit match bonus for video poker, check the terms. A bonus requiring 40x wagering may effectively require 200x or 400x wagering if video poker counts at 10% or 5%.
Final Thoughts
Video poker rewards a specific kind of attention. You don’t need to memorize 30 tiers of hold priority before your first session. But you do need to check the pay table, always bet five coins, and learn a handful of the contested hold decisions that separate a 97% effective RTP player from a 99.5% effective RTP player.
The game hasn’t changed in its fundamentals since it was introduced in the 1970s and refined through the 1980s and 90s. The math is documented, the strategy is calculable, and the correct decisions are known. What changes is how many players bother to learn them.
Start with 9/6 Jacks or Better. Use a strategy chart — there’s no penalty for having it open on your phone while you play online. Bet five coins. Set a session budget and a stop-loss, and treat a losing session as the statistical event it is rather than evidence of anything going wrong.
The difference between video poker and slots isn’t just the lower house edge. It’s that you’re a participant in the outcome, not just a spectator. That participation has value — but only if you use it correctly.
Play within your means, take regular breaks, and treat any session loss as the cost of entertainment. If gambling stops being enjoyable, that’s the only signal that matters.
