Start with the limits
Video poker is a game of chance with a built-in house edge. No strategy, system or timing makes the odds favor the player over time, and most players lose money over time. Never play with money you cannot afford to lose, and never try to recover losses by playing more. If gambling is a struggle, a local support service can help.
What “optimal” means
In video poker, optimal play means choosing, for every possible starting hand, the set of cards to hold that has the highest expected value under a specific pay table. Expected value is the average result per hand if the same situation were repeated a very large number of times. Computers can calculate this exactly because the game is small: there are only a limited number of five-card starting hands, and for each one there are 32 ways to choose which cards to hold.
For each hold choice, the calculation considers every possible set of replacement cards from the 47 unseen cards, looks up what the pay table pays, and averages. The hold with the highest average is the one a strategy recommends. The result is a strategy chart or a ranked list of holds.
A simple example
Suppose you are dealt four cards of one suit plus an unrelated card, and you hold the four. Nine cards of that suit remain among the 47 unseen cards (13 in a suit, minus the 4 you hold). Your chance of completing the flush on the single draw is 9 out of 47, about 19 percent.
That means roughly 81 percent of the time you will not complete the flush with that draw. Whether holding four to a flush is better than keeping a small made hand depends on the pay table: how much a flush pays against what the alternative pays. This is why the answer to “what should I hold?” is not universal. A change to one payout can change the best hold.
What correct play does
- It lowers the average cost. Holding the wrong cards gives away value. Mistakes cost you on average in the long run, so correct play minimizes the gap between the outcome and the machine’s best possible return for that pay table.
- It is consistent. Following one tested chart avoids hold decisions based on hunches or feelings about the machine.
- It depends on the game. Each pay table and variant, especially those with wild cards, has its own strategy. A chart for one version can lead to costly errors in another.
What correct play does not do
- It does not remove the house edge. Even with perfect holds, a machine with a return below 100 percent keeps an average advantage. Perfect play gets you to the machine’s best long-run return, not above it. This relationship is described in house edge versus RTP.
- It does not control the cards. The random number generator decides what you are dealt and what replaces your discards. The hold decision only changes how you respond.
- It does not guarantee a session result. Variance in video poker is large. Even perfect players can have long losing stretches, because much of the return comes from rare hands. See the article on the royal flush and variance for why.
- It does not make the next hand different. Hands are independent events. A machine does not owe you a good hand after poor ones.
Average results versus a single session
The long-run figure only appears after a very large number of hands, far more than a typical session. The law of large numbers describes how averages settle, but small samples can be dominated by luck. With perfect strategy, a player can still finish a session well down, and some players will finish up. Neither result proves anything about skill.
The gap between average and actual outcomes is part of why standard deviation matters. A machine with a lower edge and large swings can cost you a lot in a short time, while the formula for the average hides that. Thinking about risk of ruin alongside the percentage is more informative than the percentage alone.
The cost of learning and of fatigue
Strategy charts must be memorized or consulted, and errors tend to rise with fast play and long sessions. Even small, frequent mistakes add up, though the amounts depend on the game. Playing quickly also increases the number of hands per hour, which raises expected loss in proportion to speed of play.
Misleading claims
Be cautious about any claim that a strategy, a pattern in the cards or a “cycle” makes a machine beatable. Some printed pay tables can have a long-run return close to or above 100 percent in theory when played perfectly, but such claims often ignore the variance, the large bankroll needed, the possibility that conditions have changed, and the cost of mistakes. This article does not recommend seeking them out, and they should never be presented as a dependable way to make money.
A reasonable takeaway
Optimal play is best understood as damage control. It makes a game with a built-in cost less costly. It is a worthwhile idea for anyone who chooses to play for entertainment, but it is not a plan to win, and the decision to play should be based on whether the entertainment is worth the expected cost.