9/6 Jacks or Better: the one game here you can play wrong

What the 9/6 on the label means

Video poker paytables are named by two cells: what a full house pays and what a flush pays, per unit bet. Those two are the knobs manufacturers historically turned to tune a machine’s return, because the hands above them are too rare to move the average much and the hands below too common to touch quietly. A “9/6” machine — full house ×9, flush ×6 — is the classic full-pay Jacks or Better table, the benchmark every cut-down version is measured against.

Multipliers are per unit staked. A pair of tens or worse pays nothing: “Jacks or Better” is a literal description of the last row.

Why the label deserves attention: the industry’s standard analyses put full-pay 9/6 at roughly 99.5% returned under optimal play, while the common “8/5” cut — full house ×8, flush ×5, everything else identical — gives up more than two points. Two cells most players never read are worth more than every lucky streak they will ever have. Checking the full house and flush rows before playing any video poker, anywhere, is the single highest-value habit the game offers.

Where the skill lives

Deal five cards, choose which to hold, draw replacements once, get paid by the table above. The entire game is the middle step. Most hands hold themselves — a made flush, a high pair — but a stubborn minority put two futures in your hand at once, and picking between them is the game:

• Low pair against high cards. A pair of sixes beside a king and a queen. The pair is worthless now but leads to trips and full houses; the honors could pair into the ×1 floor. Optimal play keeps the low pair — the deep hands it reaches outweigh the frequent small save.
• Four to the flush against a made pair of jacks. Guaranteed ×1, or roughly a one-in-five draw at ×6 plus everything a flush hand can become. The draw is right, and taking the sure unit anyway is one of the most common paying mistakes in the game.
• Three to the royal against a made flush. The rare, agonizing one: break a paying ×6 to chase two perfect cards at ×800. Here the made hand usually wins the argument — which is the point: the correct answer switches sides case by case, and feel does not track it.

Optimal Jacks or Better strategy compresses into a ranked list of about thirty hand patterns — hold the highest pattern your five cards contain — that fits on a card and takes an evening to make automatic. We publish no strategy chart in the game because the ranking is standard, printed in every serious treatment of the game for decades. What we publish is the part that varies by house and hides easily: the paytable itself.

The contrast with the rest of the catalog is the reason this article exists. In keno, the risk level changes the payout table and nothing you do changes the odds; dice, limbo and crash resolve themselves the instant you commit. Video poker is the one machine here that leaves a lever in your hands after the bet — and honesty requires saying which direction the lever moves things: skill raises the return toward the published ceiling. It does not lift it over 100%.

The deck is already ordered before you choose

One engine detail is worth understanding because it defines what your decision does and does not touch. When a hand begins, the engine derives a single without-replacement sequence of cards from the committed seed pair — the five you see first, then the replacements, in order. Discard two and you get the sixth and seventh cards of that sequence; discard none and the rest are simply never read.

The consequence cuts both ways and both are worth having. The house cannot react to your holds — the replacement cards were determined, and committed, before you chose, and you can recompute the whole sequence after the round. And you cannot outguess the deck — no timing, no pattern in your discards, changes what sits at position six. The only variable left alive in the round is the quality of your hold. That is what “a game of skill within a fixed price” means, mechanically.

What optimal play buys, honestly priced

The arithmetic of the ceiling: at roughly 99.5% returned, optimal play concedes about half a percent of turnover — the engine’s stated price for the game. Careless play concedes more, invisibly, one mis-held hand at a time; standard analyses of common errors put habitual mistakes at several points of return, which at equal stakes is the difference between our tightest and loosest games.

Two implications, both unglamorous. First, the royal is not the strategy — at ×800 it is the paytable’s crown, but it arrives on the order of once per tens of thousands of optimally played hands, and play that chases it at the expense of the ranking gives back more than the crown is worth. Second, variance still rules the session: half the return of Jacks or Better flows through the ×1 and ×2 rows, which is why the game feels like a slow tide rather than a lottery — and why the sizing logic from risk of ruin applies here with unusually little drama.

Played by the list, this is among the best prices in the building, stated on a table you can read before you sit down. Played by feel, it quietly is not. No other game here puts that difference in your hands, which is exactly why it gets an article.