Breaking a made hand in video poker means discarding one or more cards from a combination that already pays because another hold has higher expected value. It is sometimes correct, but far less often than players who chase every attractive draw assume.
The deciding question is not “Do I already have a winner?” or “How large is the possible jackpot?” It is:
Which legal hold returns the most credits on average across every possible draw from this exact five-card deal?
That answer depends on the game, paytable, number of coins wagered, and cards already visible.
A made hand is only one candidate hold
A made hand is a final combination that would receive an award if the player pressed Draw without discarding it. Examples include:
- a high pair in Jacks or Better;
- two pair;
- three of a kind;
- a straight;
- a flush;
- a full house.
The hand ranking tells you its current award. It does not prove that holding all five cards is the best decision.
Every hold has an expected value:
[ EV(H)=\frac{\sum_{j=1}^{N_H}A_j}{N_H} ]
where:
- (H) is the selected group of held cards;
- (N_H) is the number of possible draws after that hold;
- (A_j) is the paytable award from draw (j).
With five cards dealt from a 52-card deck, 47 unseen cards remain. If four cards are held, there are 47 possible one-card draws. If two cards are held, the number of three-card draws is:
[ \binom{47}{3}=16{,}215 ]
A correct analyzer prices all of them, including misses, ordinary pairs, straights, flushes, and premium hands.
Exact example: break a paying pair for four to a royal
Assume full-coin 9/6 Jacks or Better, with the usual per-coin awards:
- royal flush: 800;
- straight flush: 50;
- four of a kind: 25;
- full house: 9;
- flush: 6;
- straight: 4;
- three of a kind: 3;
- two pair: 2;
- jacks or better: 1.
The deal is:
10♠ J♠ Q♠ K♠ K♦
The player has a paying pair of kings. Two natural candidates are:
- hold both kings;
- discard K♦ and hold four to the royal flush.
Holding the four suited royal cards produces 47 possible draws. The largest branches include:
- A♠: royal flush, 800 credits;
- 9♠: straight flush, 50 credits;
- seven other spades: flush, 6 credits;
- three non-spade aces and three non-spade nines: straight, 4 credits;
- two remaining kings: paying pair, 1 credit;
- all other draws: no award or another lower result according to the cards.
Summing every branch gives:
[ EV(10\spadesuit J\spadesuit Q\spadesuit K\spadesuit)=\frac{924}{47}\approx19.6596 ]
credits per coin wagered.
Holding the pair of kings requires enumerating 16,215 three-card draws. Under the same paytable, its exact expected value is approximately:
[ EV(KK)\approx1.5365 ]
The four-card royal draw is not merely a little better. It is worth more than twelve times as much on average in this specific deal. Keeping the pair because it is “guaranteed” sacrifices substantial expected return.
The royal still misses most of the time. Correct expected-value play is not a promise that the draw will win.
Exact counterexample: keep the straight
Now consider:
5♠ 6♠ 7♠ 8♠ 9♦
The hand is already a straight paying 4 credits per coin in 9/6 Jacks or Better. The first four cards are also an open-ended four-card straight-flush draw.
That draw looks powerful because 4♠ or 9♠ completes a straight flush. But across the 47 possible one-card draws, holding 5♠-6♠-7♠-8♠ produces:
- 2 straight flushes;
- 7 ordinary flushes;
- 5 ordinary straights;
- many pairs or misses.
Its exact expected value is:
[ EV(5\spadesuit6\spadesuit7\spadesuit8\spadesuit)\approx3.4468 ]
Keeping the made straight returns exactly 4 credits. Therefore:
[ 4.0000>3.4468 ]
Breaking this straight is a mistake under the stated paytable, even though the draw can make a much larger hand.
These two examples show why slogans fail. “Never break a winner” loses value in the first hand. “Always chase a straight flush” loses value in the second.
The paytable can reverse a close decision
A made-hand decision cannot be separated from the paytable.
Changing any of these rows can alter the ranking:
- royal flush;
- straight flush;
- four of a kind;
- full house;
- flush;
- straight;
- two pair;
- high pair.
Bonus variants place more return in quads. Deuces Wild changes the meaning of wild cards and removes the ordinary high-pair floor. Some games pay only one unit for two pair; others pay two. A royal may pay 250 per coin at one through four coins but 800 per coin at five coins.
That is why a strategy learned for Jacks or Better cannot be transferred unchanged to Bonus Poker, Double Bonus, Double Double Bonus, or Deuces Wild. The page Expected Value of a Hold explains the general calculation; this page focuses on the emotional conflict created when the lower-EV choice is already paying.
Four to a royal is the major exception, not a universal excuse
Four to a royal is powerful because one remaining card can produce the game’s largest standard award, while several other draws may still make straights, flushes, or high pairs.
But four suited cards are not automatically four to a royal. Compare:
- 10♠-J♠-Q♠-K♠: one card from a royal, with straight and flush backup;
- 2♠-5♠-8♠-J♠: an ordinary four-card flush draw;
- 5♠-6♠-7♠-8♠: an open-ended straight-flush draw.
Each has a different set of completing cards and a different award distribution.
For detailed royal conflicts, continue to Four to a Royal. Do not generalize that page into breaking made hands for weak suited fragments.
Penalty cards can change the answer
A penalty card is a discarded card that removes a useful completion from the unseen deck. It can slightly lower the EV of a hold even though it is not retained.
Suppose two candidate holds are extremely close. A discarded card may:
- remove one flush completion;
- block a straight possibility;
- reduce the number of high-pair draws;
- occupy a rank needed for a full house or four of a kind.
The visible fifth card in a four-card draw is therefore not irrelevant. Basic strategy charts often group hands that are identical in most deals, but advanced exceptions arise because exact suits and ranks change the draw inventory.
This is why penalty cards matter most near strategy boundaries, not in obvious gaps such as 19.6596 versus 1.5365.
Common made-hand conflicts
High pair versus premium draw
A high pair usually has solid value because it starts with a guaranteed one-credit award and can improve to trips, two pair, a full house, or quads. Breaking it requires a strong alternative, commonly four to a royal.
An ordinary three-card straight or four-card inside straight is not enough merely because it looks exciting.
Two pair versus one pair
In 9/6 Jacks or Better, two pair pays 2 and normally should be held. In some bonus variants, two pair pays only 1 while premium quads pay much more. Even then, the correct choice is paytable-specific and can be close. Use the strategy chart for the exact game rather than treating “bonus” as permission to discard a pair.
Straight or flush versus four to a royal
Four to a royal commonly outranks a made straight or flush because the royal award is so large. This is one of the few situations where discarding a respectable completed hand is routine under the proper chart.
An ordinary four-card flush or straight-flush draw may not have enough value to justify the break, as the 5♠-6♠-7♠-8♠-9♦ example shows.
Full house or four of a kind
These are almost always retained in ordinary single-hand games. Claims that a player should break a full house for a speculative royal draw usually come from misreading the cards, using an unusual promotion, or confusing a different variant’s rules.
Expected value and certainty answer different questions
Suppose the made hand pays (M) credits and the alternative hold has expected value (D):
[ \Delta EV=D-M ]
- If (\Delta EV>0), breaking the hand has higher long-run value.
- If (\Delta EV<0), keeping the made hand has higher long-run value.
- If the difference is tiny, paytable accuracy and penalty cards become especially important.
The made hand may have lower variance because it locks in an award. The draw may have higher EV and higher variance. Strategy charts rank expected value, not emotional comfort or session preservation.
A player may deliberately choose the lower-variance option for entertainment or bankroll reasons, but it should be described honestly as a deviation from maximum-EV strategy.
What the machine must do—and what it does not decide
A regulated video poker device must correctly display the game rules, accept the selected holds, draw from the approved virtual deck process, evaluate the final hand, and award the configured paytable. It does not advise the player which cards should be held.
Gaming Laboratories International’s GLI-11 gaming-device standard includes requirements for game rules, paytable information, random selection, and reporting used in certification work. Nevada’s published technical standards for gaming devices likewise address draw-poker device behavior and permitted configuration controls. Jurisdictional rules govern the machine; player strategy remains a mathematical choice.
If a player discards a paying combination and the replacement cards miss, the machine has not “taken” a completed award improperly. The draw instruction replaced the original hand. A real dispute concerns whether the selected holds, draw, final evaluation, or posted payout were recorded correctly.
A disciplined decision process
Before breaking a made hand:
- Identify the exact game and paytable.
- Confirm the active coin level, especially the royal award.
- Name the made hand and its current award.
- Name the alternative hold precisely.
- Check the correct strategy chart or analyzer.
- Look for suit or rank penalty cards if the decision is close.
- Accept that the higher-EV draw can lose immediately.
The phrase “made hand” describes the current screen, not the optimal action. The right hold is the one with the highest expected return under the actual rules.
For related decisions, read Hold or Draw Decisions, Common Video Poker Strategy Mistakes, and Video Poker Variance. A strategy chart is most useful when it resolves exactly this conflict: a visible small award versus a less certain but more valuable draw.