Video Poker Strategy Guide: How to Make the Best Decisions Every Hand

Why video poker even has a "strategy"

Slot machines give you nothing to decide. You press spin, the reels land, and the outcome is already determined the instant you commit your bet — there's no button that changes your odds. Video poker is built differently. After the machine deals your five cards, you get exactly one real decision per hand: which of those five cards to hold before the draw, and which to discard and replace. That single decision — repeated, hand after hand — is the entire skill of the game, and unlike a hunch or a "feel," it has a mathematically correct answer for every possible five-card hand you can be dealt.

That's a strong claim, so it's worth unpacking. A standard video poker machine deals from a 52-card deck (Joker Poker variants use a 53-card deck with one wild joker, but the core mechanic is identical). For any five cards you're dealt, there are exactly 32 possible ways to hold them — hold all five, hold four and draw one, hold none and draw five, and every combination in between (that's 2⁵, the number of ways to include or exclude each of the five cards from your hold). Each of those 32 hold decisions leads to a different, calculable set of possible outcomes once the discarded cards are replaced. Strategy is simply the process of identifying which of those 32 holds produces the best average result — and doing that consistently, hand after hand, is what separates a strategy player from someone pulling a lever and hoping.

This guide is the strategy companion to the full beginner's guide to how the game works, which covers the mechanics of dealing and drawing in more depth. If you haven't played a hand yet, start there; if you already understand the basic step-by-step of dealing and drawing and want to know how to actually make good decisions once the cards are in front of you, you're in the right place.

Expected value: the concept that makes strategy possible

Every hold decision in video poker can be evaluated using a concept called expected value, or EV. In plain terms, EV is the average payout you'd get from a specific hold if you could somehow repeat that exact same five-card starting hand thousands of times and let the draw play out every possible way it could go. You can't actually do that at the machine — you only see one draw per hand, and that draw is genuinely random — but the math behind it is fixed and knowable, because the deck is finite and every remaining card has a known, calculable chance of appearing.

Here's a simplified way to see it. Say you're deciding between two holds on the same hand: Hold A discards three cards, and there are a fixed number of ways those three replacement cards could land, each with its own payout (some improve your hand, most don't, a few make it worse in the sense of not improving it at all). Multiply each possible outcome's payout by its probability, add them all up, and you get Hold A's expected value. Do the same for Hold B, Hold C, and every other viable hold on that hand, and whichever number comes out highest is the mathematically correct play — not because it guarantees a win, but because across a large enough sample of hands with that same starting position, it pays back more on average than any other option.

This is the crucial distinction to internalize: EV tells you nothing about what happens on any single hand. You can make the highest-EV decision and still draw nothing, and you can make a poor decision and get lucky. What EV guarantees is your long-run outcome — over thousands of hands, the highest-EV choice on each one compounds into meaningfully more money kept than any other approach, including pure guessing. This is exactly the same logic that underpins blackjack basic strategy: a fixed, correct decision for a given situation that pays off in the aggregate rather than on any individual hand. The difference is that blackjack strategy responds to what the dealer is showing, while video poker strategy responds entirely to your own five cards and the specific pay table you're playing.

Strategy is pay-table-specific — there is no single "video poker strategy"

This is the single most important thing to understand before you memorize anything: there is no one universal strategy chart that works for every video poker machine. The correct hold for a given hand depends on the exact payouts attached to every hand type on that specific machine, because EV is calculated using those payouts. Change the payout on a full house from 9-for-1 to 6-for-1, for example, and you've changed the relative value of holding toward a full house versus some other hand — which can flip which hold is correct in a borderline situation.

This matters at two levels. First, between game types: Jacks or Better, Deuces Wild, Joker Poker, Bonus Poker, and other variants each reward different hand types differently — Deuces Wild pays nothing for a pair that isn't at least three-of-a-kind-equivalent in value but pays enormous multipliers for four deuces, which reshapes the entire hold hierarchy compared to a pair-based game like Jacks or Better. A hold that's obviously correct in Jacks or Better can be a clear mistake in Deuces Wild, and vice versa. For a full breakdown of how strategy differs between game types, see our dedicated variants guide before assuming any chart you've seen applies to the machine in front of you.

Second, and less obvious to new players, strategy differs between two machines running the same game with different pay tables. Two cabinets can both say "Jacks or Better" and still pay meaningfully different amounts for a full house and a flush — commonly written as "9/6" (9-for-1 on a full house, 6-for-1 on a flush) versus "8/5" or the much worse "6/5." These aren't cosmetic differences. They change the correct strategy in specific borderline hands, and more importantly they change the entire game's long-run payback percentage. We cover exactly how pay tables set your long-run odds in detail — the short version is that a full-pay 9/6 Jacks or Better machine returns approximately 99.54% with mathematically perfect play, while an 8/5 table drops to roughly 97.3%, and a 6/5 table — visually almost identical on the cabinet — falls to around 95%. Always confirm the posted pay table on the paytable screen before assuming a strategy chart you've learned applies. It's the first habit worth building, before any hold decision.

The hold hierarchy: how to rank what you're dealt

While the exact numbers shift by pay table and game variant, a general hierarchy of hold priorities applies broadly across Jacks-or-Better-style games, and understanding the reasoning behind the order matters more than memorizing it, because the reasoning is what lets you adapt when the pay table changes. Here's that hierarchy, from strongest to weakest, with the logic behind each tier. For the underlying hand values referenced here, see the full hand ranking order.

A dealt Royal Flush. If the machine deals you a Royal Flush outright, hold all five cards. This isn't really a decision — it's the single highest-paying hand in the game, typically rewarded at a bonus multiplier far beyond a normal payout curve, and there is no possible draw that improves on already having it. Every other hand type on this list is here because it isn't yet a Royal Flush.

Four cards to a Royal Flush. This is where strategy starts to feel counter-intuitive to new players. If you hold four of the five cards needed for a Royal Flush — say 10-J-Q-K of the same suit, with a fifth unrelated card — this draw is usually correct to hold even over a made straight, made flush, or made full house that you might already have in hand. The reasoning is entirely about the size of the prize relative to its odds. A Royal Flush pays so disproportionately more than every other hand that even though you'll only complete it a small fraction of the time from a four-card draw, the payout when you do is large enough to outweigh the near-certain smaller payout from keeping a made hand. This is a pure expected-value calculation: multiply a modest probability by an enormous payout, and it can exceed the value of a near-certain small payout. It's the clearest example in the whole game of why "the play that feels safest" and "the play with the highest EV" aren't always the same thing.

A made Straight Flush, Four of a Kind, Full House, Flush, Straight, or Three of a Kind. Once you have one of these made hands and you're not sitting on a four-card Royal draw instead, hold every card contributing to it and don't break it apart chasing something bigger. These hands already pay well, and breaking any of them to draw for an improvement almost always destroys more expected value than it creates, because the odds of a meaningfully better outcome from a broken draw rarely justify walking away from a guaranteed payout you already hold in your hand.

Two pair. In Jacks or Better and similar pair-based games, hold both pairs and draw one card, hoping to improve to a full house. Do not break one of the pairs to keep a single high card as a "kicker" hoping for a different pair or trips — a made two pair already guarantees a payout, and the only way to improve it is drawing the exact card that completes a full house from your one remaining draw slot, which is a low-probability, high-value improvement you want to give yourself the best possible shot at, not dilute with an unrelated hold.

A high pair (Jacks or better). This is where Jacks or Better gets its name: only a pair of Jacks, Queens, Kings, or Aces qualifies for a payout at all — pairs of 2s through 10s pay nothing on their own in standard Jacks or Better. When you're dealt a high pair, the correct play in the overwhelming majority of cases is to hold just the pair and draw three fresh cards, discarding everything else, even a card that looks tempting to keep. We'll walk through exactly why in the worked example below, because this is one of the most common and costly mistakes new players make.

Four cards to a flush. Four cards of the same suit, with no pair and no straight already in hand, is usually a stronger hold than a low pair or a straight draw, because a flush pays well and you have nine remaining cards of your suit left in the deck (13 per suit, minus the four you're holding) that complete it on the draw.

A low pair (2s through 10s) versus three cards to a straight or a flush. This is one of the genuinely close calls in the hierarchy, and the correct answer depends on the specific cards and pay table. A low pair guarantees nothing on its own — remember, it doesn't pay in Jacks or Better — but it carries strong improvement potential: it can become three of a kind, a full house, or four of a kind, and it also keeps you "live" for a decent chance at ending the hand with at least your original bet back if you improve to trips or better. A three-card straight or flush draw, by contrast, needs two specific cards to land in the right two draw slots, which is a much steeper ask than a pair needing one more matching card. In most standard pay tables, a low pair beats a three-card straight or flush draw for this reason — but an open-ended four-card straight or a stronger draw with extra outs (like three cards to a straight flush) can flip that comparison. This is exactly the kind of close, pay-table-sensitive decision a proper strategy chart resolves precisely, rather than something to eyeball by feel.

Nothing worth holding. Sometimes you're dealt a genuine garbage hand — no pair, no three-or-more cards toward a made flush or straight, nothing but scattered, unconnected cards. In that situation, the correct play is to discard all five and draw a completely fresh hand. New players sometimes instinctively hold a lone high card (an Ace or King) out of habit, but if it isn't part of a pair, a flush draw, or a straight draw, it carries no real improvement value and holding it only reduces the number of fresh cards you get to see. Five fresh cards give you a better shot at a strong new hand than four fresh cards plus one card that was never going to do anything for you.

Worked example: why a "kicker" next to a high pair is a mistake

Say you're dealt Jack-Jack-King-4-9, no suits matching, no straight possibility. You already have a paying hand — a pair of Jacks. The tempting instinct for a new player is to hold the King alongside the pair, reasoning that if another King lands you'll have two pair, which pays more than a single pair.

That instinct is measurably wrong, and it's worth understanding exactly why. Holding just the two Jacks and drawing three fresh cards gives you three chances to land another Jack (for four of a kind), or to pair up any of the three new cards you draw alongside your existing Jacks-only hold... but more importantly, it maximizes the number of ways your hand can improve at all. Holding Jack-Jack-King and drawing only two cards, by comparison, narrows your outs significantly: now you need one of those two new cards to specifically pair the King (and there are only three Kings left in the deck to do it with), while also losing a meaningful share of your chances to draw into three-of-a-kind or four-of-a-kind Jacks, since you're only seeing two new cards instead of three.

Run the actual comparison and holding just the pair comes out ahead in expected value on essentially every standard Jacks or Better pay table. The King "kicker" hold does occasionally produce two pair, but not often enough, and the payout for two pair isn't large enough relative to a pair, to offset the much larger loss in overall improvement chances from seeing two fresh cards instead of three. The general rule this example teaches: once you have a made paying hand, adding an unrelated "just in case" card to the hold almost always costs you more in reduced draw potential than it could possibly gain.

Worked example: low pair versus a four-card draw

Now consider a genuinely close decision. Say you're dealt 8-8-9-10-Jack, with the 9, 10, and Jack all the same suit — spades, for example — and no suit match on either 8. You have two live options: hold the pair of 8s and draw three cards toward trips or better, or hold the three-card straight-and-partial-flush (9-10-Jack of spades) and draw two cards.

Here the comparison depends closely on exactly what you're holding and the pay table in play. A pair of 8s guarantees no payout on its own but has strong improvement odds toward three of a kind, a full house, or quads across three fresh draw cards. The three suited connectors, on the other hand, are only a three-card draw toward either a straight or a flush — needing two specific cards to land correctly out of two draw slots — which is a materially harder ask than improving a pair, even though a completed straight or flush would pay considerably more than three of a kind. Now change one detail: if instead of three scattered high cards you were dealt four cards to a flush already (say 8-8 plus two spades instead of three), that four-card flush draw — needing only one more spade out of nine remaining in the deck — would usually beat the low pair outright, because a four-card draw with nine outs is a much stronger, higher-probability position than a three-card draw needing an inside sequence.

This is exactly the kind of decision where "it depends" is the honest, correct answer, and where a real strategy chart for your specific pay table earns its keep — the difference between a three-card draw and a four-card draw, or between an inside straight draw and an open-ended one, is often the entire swing between which side of the comparison wins.

Reading a strategy chart, rather than memorizing every number

Serious video poker players don't carry the full EV calculation for every one of the roughly 2.6 million possible five-card hands in their head — nobody does. What they use instead is a strategy chart: a printed or on-screen reference that lists every meaningful hand category in ranked order, from strongest hold down to weakest, specific to the exact pay table they're playing. These charts exist for every common variant and pay table combination, and reputable ones are built directly from the underlying EV math, not guesswork.

The practical skill isn't memorizing the chart cell by cell — it's learning to recognize which category your dealt hand falls into (a made flush? a four-card royal draw? a low pair with a weak kicker?) and then finding that category's position in the ranked list quickly enough to use it at normal playing speed. Most charts are organized top-to-bottom in EV order specifically so you can scan down the list and stop at the first category that matches your hand — that's your correct play. With repetition, recognizing hand categories becomes close to instant, which is really what "knowing strategy" means in practice: pattern recognition backed by math you don't have to re-derive every hand.

Common strategy mistakes worth naming

A handful of errors account for most of the expected value lost by new video poker players, and all of them are avoidable once you know to watch for them:

• Breaking a made hand to chase a Royal Flush without an actual four-card Royal draw. A four-card Royal draw specifically means four of the five cards needed, of the correct suit, already in hand. Holding three Royal-quality cards and hoping is a much weaker position than the math supports breaking a made flush or straight for — that's a mistake, not the aggressive-but-correct four-card Royal hold described earlier.
• Holding an unnecessary kicker next to a pair. Covered in detail above — it feels intuitive and it's measurably wrong in the large majority of cases.
• Discarding a made low pair to chase a gutshot (inside) straight draw. An inside straight draw needs one specific rank to complete, out of four available cards of that rank in the deck, a much narrower path than most players intuitively credit it with, and it's usually a weaker hold than a pair that already has real improvement potential of its own.
• Not recognizing a truly worthless hand. Holding a single unconnected high card out of habit, rather than drawing a fresh five, understates how much a clean slate is worth when nothing in your current hand is contributing to a plausible improvement.

Max-coin play: not just a betting-size choice, a strategy input

It's tempting to think of bet size as a separate decision from strategy — how much you wager, versus which cards you hold. In video poker they're linked in a way that doesn't come up in most other casino games. The Royal Flush payout on the overwhelming majority of video poker machines is structured to reward exactly five coins wagered per hand: typically around 800-for-1 at max coin (five coins bet), compared to roughly 250-for-1 per coin at anything below max coin. That's not a small rounding difference — it's a deliberate jump built into the pay table specifically to reward max-coin players, and it exists because the Royal Flush is rare enough that operators can afford a disproportionate bonus at the top bet level without materially changing the machine's overall return.

Because that bonus only exists at max coin, every standard strategy chart is calculated assuming you're betting max coin. Betting below max coin doesn't just reduce your win size proportionally — it can change which hold is actually correct in Royal-draw situations, because the huge EV boost that makes a four-card Royal draw worth holding over a made flush or straight comes specifically from that oversized max-coin jackpot. Strip that bonus away by betting one to four coins, and some of those aggressive Royal-chasing holds are no longer justified by the math. In practical terms, this means max-coin play isn't a separate "how much to bet" decision layered on top of strategy — it's a precondition for the strategy chart you're using to actually be correct. If your bankroll can't comfortably support max-coin bets at a given denomination, the better fix is to drop down to a lower coin denomination and still bet max coin there, rather than bet fewer coins at a higher denomination.

Bankroll and variance: strategy sets your expectation, not your outcome

Correct strategy on a known pay table gives you a fixed, calculable long-run return-to-player percentage — that's the whole point of the math behind it. But "long run" is doing real work in that sentence. Any individual session, even one played with flawless strategy, is still subject to genuine variance, and video poker's variance is concentrated heavily around its rarest, highest-paying hands. A Royal Flush might show up once every 40,000-plus hands even under correct play, which means it's entirely normal to play for a long time without one and see your session results run below the machine's theoretical return, just as it's possible to hit one early and run well above it. Strategy doesn't smooth that out on any single sitting — it only guarantees that, averaged across a very large number of hands, your results converge toward the pay table's built-in return.

That variance is exactly why bankroll sizing matters as much as hold decisions. A useful starting principle is to size your total session bankroll as a meaningful multiple of your per-hand max-coin bet — enough hands' worth of buffer to ride out an ordinary cold streak without being forced to stop before variance has had a fair chance to even out. Betting a stake that eats a large share of your bankroll per hand leaves very little room for a normal run of unimproved hands, regardless of how correctly you're holding.

Strategy and pay table selection work together, not separately

It's worth being direct about something that surprises new players: perfect strategy applied to a bad pay table will still underperform mediocre strategy applied to a good one. A player making textbook-correct decisions on a 6/5 Jacks or Better machine — one paying only 6-for-1 on a full house and 5-for-1 on a flush — is playing against a built-in return around 95%, no matter how well they hold. A player making decent, mostly-correct decisions on a full-pay 9/6 table is working against a return closer to 99.54%, a gap of several percentage points that dwarfs what strategy refinement alone can claw back. Strategy is what you control once you've sat down; pay table selection is what determines the ceiling on how much that strategy can actually be worth. Before you play a session seriously, it's worth checking how pay tables set your long-run odds so you know what return you're playing toward in the first place — the two skills, reading a pay table and reading a strategy chart, are meant to be used together, not as substitutes for each other.

It's also worth remembering that this pay-table sensitivity isn't unique to Jacks or Better. Full Pay Deuces Wild, played with correct strategy, can actually exceed 100% return (theoretical figures around 100.76% have been published for the full-pay version) — though the overwhelming majority of Deuces Wild machines on casino floors are not full pay, and a short-paying version of the same game can sit well below 100%. The lesson is the same across every variant: the game's name on the cabinet tells you the rules, but the specific pay table tells you the number that actually matters.

Frequently asked questions

Do I need to memorize a complete strategy chart before I play for real money? Not to start. Understanding the general hold hierarchy covered above — made hands beat draws, high pairs beat unnecessary kickers, four-card draws beat three-card draws — gets you most of the way to correct play. A full chart specific to your pay table sharpens the close calls, and many players keep one open on a second screen or reference an app while playing rather than memorizing every cell.

Does strategy change between video poker variants? Yes, substantially. Deuces Wild, Joker Poker, Bonus Poker, and other variants pay different hand types differently, which reshapes the entire hold hierarchy compared to standard Jacks or Better. See how strategy differs between game types before applying a Jacks or Better chart to a different game.

Is video poker strategy the same thing as card counting in blackjack? No, and the distinction matters. Card counting tracks the composition of a shrinking physical deck across many consecutive hands to gain a shifting edge over time. Video poker strategy is a per-hand decision made from a freshly shuffled virtual deck (or a physical deck reset every hand on most machines), evaluated purely against fixed, known payout odds. There's nothing to track between hands — every hand starts from the same known probabilities, which is what makes a single, unchanging chart useful at all.

Can I use a strategy app or on-screen chart while playing online? Generally yes, since you're playing on your own device at your own pace with no other players or dealer waiting on you. Many online video poker games even include a built-in "strategy hint" or auto-hold feature. Using an external reference is a normal, accepted part of learning the game properly rather than something to be self-conscious about.

Does using correct strategy guarantee I'll win? No. Strategy sets your long-run expected return on a given pay table — it doesn't change the outcome of any single hand, which remains genuinely random. Two players making identical, correct decisions on the same pay table will still see very different results over a short session; the value of strategy shows up over a large number of hands, not a single sitting.

What's the single biggest mistake new players make? Holding an unnecessary kicker next to a made pair, as shown in the worked example above. It feels intuitive — "more cards toward two pair can't hurt" — but it measurably reduces expected value on the large majority of hands where it comes up, because it shrinks the number of fresh cards you see without a large enough payoff to justify the trade.

Does bet size change which hold is correct? It can, specifically around Royal Flush draws. Standard strategy charts are built assuming max-coin play, because the Royal Flush's outsized bonus payout only applies at five coins wagered. Betting below max coin can make some aggressive Royal-chasing holds no longer correct, since the reward that justified them isn't present at a lower bet level.

Why do some strategy charts disagree with each other? Almost always because they're built for different pay tables or different game variants, not because one is wrong. A chart built for 9/6 Jacks or Better won't exactly match one built for 8/5 Jacks or Better or for Deuces Wild — always confirm a chart matches your specific machine's posted pay table before trusting it.

Is it ever correct to hold fewer than five cards on a dealt Royal Flush? No. A dealt Royal Flush is already the best possible outcome in the game; there's no five-card draw that could improve on it, so holding all five and drawing zero new cards is always correct.

How much does correct strategy actually improve my results compared to guessing? Substantially. The gap between optimal strategy and casual, instinct-based play is commonly cited in the several-percentage-point range on standard Jacks or Better pay tables — a meaningful shift in return that compounds significantly across a long session, similar in scale to the difference basic strategy makes in blackjack.

Should beginners start with Jacks or Better before trying other variants? It's a reasonable starting point because its hold hierarchy — pairs, flushes, straights, and Royal draws — maps closely to standard poker hand values most players already recognize, making the logic behind each hold decision more intuitive while you're still building the habit of thinking in terms of expected value.

Quick glossary reference

A handful of terms come up constantly once you start applying strategy: hold (a card kept before the draw), draw (the replacement of discarded cards), expected value / EV (the average payout of a decision across all possible outcomes), kicker (an unrelated card held alongside a pair, usually a mistake), four-to-a-flush (four cards of one suit needing a fifth to complete), and pay table (the specific payout schedule for a given machine that determines correct strategy). For the full glossary, see our complete terms guide, and for the underlying rules referenced throughout this guide, see the full rulebook.

Where to go next

Strategy is the part of video poker you control, but it only reaches its full value when paired with a pay table worth playing on in the first place — the two decisions work together, not separately. From here, the highest-value next steps are confirming you understand the full hand ranking order so you can recognize hand categories instantly, checking how pay tables set your long-run odds before choosing a machine, and reading through how strategy differs between game types if you're planning to branch out beyond Jacks or Better. Put together, those three pieces — hand recognition, pay table awareness, and variant-specific strategy — are what turn video poker from a game you're guessing at into one you're playing with a real, calculable edge over guessing.