Full-pay machines and why they became hard to find

A full-pay machine is not generous, and on most games it is not even a winner. The phrase means something narrower than it sounds: for each family of paytable, somebody wrote down the best version that casinos actually put on a floor, and every other version is measured against it. For Jacks or Better the reference is the 9/6 machine — nine coins per coin staked for a full house, six for a flush, 99.54% back with correct play.

Those two rows do all the work, because the rest of the table does not move. Royal flush 800 per coin at the maximum bet, straight flush 50, four of a kind 25, straight 4, three of a kind 3, two pair 2, a pair of jacks or better 1 — identical on every Jacks or Better machine you will sit at. The names are built from the only two numbers that vary, which is why 9/6 and 8/5 are machine names rather than scores out of ten.

What a trimmed row costs

You can price the cuts with arithmetic rather than a simulator. Played correctly on 9/6, a full house lands on 1.1512% of deals and a flush on 1.1015%. Removing one coin per coin staked from a row removes exactly that fraction of a coin per hand, so the cost of a cut is simply the frequency of the hand being cut.

Every step down the ladder is 1.15 for a full-house coin and 1.10 for a flush coin, which is the two frequencies above rounded to two places. Subtract them from 99.54 in any order and you land on the published returns to within a hundredth of a point; the hundredth is the strategy being re-optimised for the cheaper table. The ordering of 9/5 and 8/6 falls out of the same arithmetic — the full house is the slightly commoner hand, so the machine that keeps it is the better one, by six hundredths of a point.

Why the cut is invisible from the stool

Nothing about a trimmed machine feels different. The deck is a deck, the hand frequencies are unchanged, and the hold list barely moves: both cuts come out of hands you have already made rather than out of the draws you choose between. A cheaper flush does devalue a four-card flush draw slightly, so a handful of marginal comparisons flip, but the 9/6 priority list plays an 8/5 machine at a cost measured in hundredths. Two numbers on the glass are the entire difference, and 2.24 points of return go with them.

That invisibility is the commercial point. A trimmed paytable needs no new cabinet, no new maths and no explanation to the player, and once paytables became downloadable it did not even need a new glass.

Why they thinned out

Full pay made sense when video poker was being established. A machine whose arithmetic was published and nearly break-even drew players who would never sit down at a blackjack table, and the house still kept 0.46% of everything staked — thin in theory, and considerably fatter in practice, because almost nobody played the chart correctly.

Two things closed that gap. Correct strategy became free: first books and printed cards, then trainers, then a phone in every pocket, so the money actually held drifted down towards the theoretical 0.46%. And the floor around the machines changed, as multi-line slots arrived keeping several times as much per square foot of the same carpet. A game that returns 99.54% to a competent player has to be justified against that comparison every year.

The marketing moved to the top of the table instead. Bonus quad rows, progressive meters and multi-hand all advertise a jackpot rather than a return, and all of them are easier to put on a sign than a full house paying nine. Full pay survived longest at low denominations and on floors with regulars who compare paytables and will walk; it went first where nobody was going to leave over one row.

Reading a floor

Three habits follow. Check the two rows against the ladder rather than the claim on the sign — an advertised payback percentage is usually a floor-wide average across every game in the building, and a machine's hold percentage is a historical number about other people's play rather than a property of the paytable. Check the denomination, because the same game one step up is often the better table. And check which family you are in first, since a full-pay version of a worse game is still a worse game.

The underlying habit is to treat the paytable as the machine. Return, variance, correct strategy and how long a stack lasts are all downstream of those printed rows, and the rows are the only part of a machine you can read before you put money in. A 6/5 machine and a 9/6 machine are the same game in every respect but one, and that one is worth four and a half points.