The 99.54 percent printed on a 9 and 6 Jacks or Better machine is a promise about a player who makes the correct discard on every single hand. You are not that player yet, and the difference between you and that player is the real price of the game.
That is what separates video poker from a slot. On a slot the number is fixed the moment you choose the game. Here the number is a ceiling you have to play your way up to, hand by hand.
The good news is that the correct play is derived rather than felt, so it can be written down and followed. This page is the writing down.
Every figure below is either taken from the video poker guide on this site or it is arithmetic from those figures, and the table says which.
Where Bad Play Overtakes a Bad Pay Table
Everybody knows the pay table matters. Far fewer people know where their own errors sit next to it, so here is the line, drawn with three returns this site already publishes.
Jacks or Better at 9 and 6 returns 99.54 percent with perfect play. At 8 and 5, which is the version you will actually meet most often, it returns 97.30. At 6 and 5, the worst of the six pay tables on the guide, it returns 95.00.
| Pay table | Return with perfect play | How big a strategy leak puts you below it |
|---|---|---|
| 9 and 6 Jacks or Better | 99.54% | The machine you want to be sitting at |
| 8 and 5 Jacks or Better | 97.30% | 2.24 points of error on a 9 and 6 machine |
| 6 and 5 Jacks or Better | 95.00% | 4.54 points of error on a 9 and 6 machine |
All three returns are published on the video poker guide here, which carries the full six row ladder from 9 and 6 down to 6 and 5. The right column is the subtraction that follows from them.
Read the right column as a pair of thresholds. Leak 2.24 points and you have thrown away the whole advantage of finding the best machine in the building over the ordinary one. Leak 4.54 and you are playing the best machine at the price of the worst.
That is a narrower margin than it sounds. The pay table is one decision made once, before the first hand. The strategy is a decision made again every twenty seconds for as long as you sit there.
I am not going to tell you how many points the average player leaks, because I have no source I would stand behind for that number. What I can do is show you where leaks of that size come from, and all three of them are avoidable in an afternoon.
The Three Leaks, Priced
Two of these have published costs. The third does not, but its size can be bounded, which is almost as useful.
| The leak | What it costs | How long it takes to close |
|---|---|---|
| Playing the wrong pay table | 2.24 points down to 8 and 5, 4.54 down to 6 and 5 | Five seconds of reading before you sit down |
| Betting fewer than five coins | About 1.4 points, because the royal drops from 800 per coin to 250 | One press of the bet button |
| Misplaying hands that contain a royal draw | Attacks the two points the royal contributes | Learning eight ranked rules |
The pay table and five coin figures are published on this site. The third is bounded rather than measured: the royal supplies about two points of the published return, so errors on those hands cannot cost more than that, and they can cost most of it.
Add the first two together and you are already at 5.94 points, which is more than the whole gap between the best and worst pay table. Two decisions, neither of which involves any skill at cards.
One clarification on the five coin rule, because it is the one people get backwards. If five coins is more money than you want to risk, lower the coin value instead. Five coins at the smallest denomination always beats four coins at a larger one.
Six Mistakes the Priority Chart Already Rules Out
The video poker guide publishes the hold priority chart in full, eight ranked rules from a pat royal down to two unsuited high cards. What it does not do is name the errors those rules exist to prevent, so that is this section.
Every row below is the chart read backwards. If the chart puts A above B, then keeping B when you hold both is a mistake, and these are the six that come up most.
| What players do | What the chart says | Why the instinct is wrong |
|---|---|---|
| Break two pair to keep the higher pair | Two pair ranks above a high pair | Two pair is already a paying hand with two ways to improve to a full house |
| Throw away a paying high pair to draw three to a royal | A high pair ranks above three to a royal | You are discarding a certain payout to chase a draw that misses almost every time |
| Keep a low pair over four cards to a flush | Four to a flush ranks above a low pair | A low pair pays nothing by itself, and nine cards complete the flush |
| Break a low pair to chase an inside straight | A low pair ranks above four to an outside straight | An inside straight is weaker still, with four outs instead of eight |
| Hold two unsuited high cards over two suited ones | Two suited high cards rank above two unsuited | Same straight and pair chances, plus every flush and royal the suit adds |
| Add a high kicker to a pair | No rule on the chart holds a pair plus a fifth card | The kicker removes one of the three cards that could make your trips or quads |
Each row is derived directly from the eight rule chart published on the video poker guide here. If a hand is not covered by one of those eight rules, the chart is not the place to guess from.
Notice what the six have in common. Five of them break a hand that already has value in order to chase one that might, and the sixth keeps a card that does nothing but reduce the draw.
The Correct Play That Feels Worst
There is one line on the chart almost nobody follows the first time they see it. Four cards to a royal ranks above a made flush, a made straight and a made full house.
So when you are dealt a flush that happens to contain four royal cards, the chart says break the flush. You throw away a paying hand and draw one card.
The reason is the royal itself. It pays 800 per coin at five coins, and the guide puts its contribution at roughly two of the 99.54 points, earned at about one royal every 40,000 hands. A hand that is one card away from it is rare enough that the chart is willing to spend a made flush on it.
This is also the clearest demonstration of why charts beat instinct. No feeling at the screen produces that play, and no amount of experience at the table produces it either. It comes out of arithmetic, which is exactly why somebody did the arithmetic once so that nobody has to feel their way to it.
One boundary on it. The chart ranks four to a royal above those made hands, not above a pat royal, a pat straight flush or four of a kind, which sit at the top of the list and get kept as dealt.
What an Hour Costs on Each Machine
Percentages are easy to shrug at. Hours are not, and video poker is fast, so the two numbers meet quickly.
Take an unhurried five hundred hands an hour at five coins of 0.25, which is 1.25 a hand and 625 euros of turnover in that hour. The expected cost of the hour is that turnover times the house edge.
| How you are playing | Effective return | Expected cost of one hour |
|---|---|---|
| Perfect play, 9 and 6 machine | 99.54% | 2.88 euros |
| Perfect play, 8 and 5 machine | 97.30% | 16.88 euros |
| Perfect play, 6 and 5 machine | 95.00% | 31.25 euros |
| Four coins instead of five, 9 and 6 machine | About 98.14% | About 11.6 euros |
| A 9 and 6 machine played with a 4.54 point leak | 95.00% | 31.25 euros |
The return figures are published on this site. The euro column applies the same 625 euros of hourly turnover to each return, so the rows compare like with like. A four coin player at the same coin value also stakes less per hour, which lowers that row again without making the machine any better.
The four coin row is the cheapest fix in gambling. Pressing the bet button once a hand moves the hour from about 11.6 euros to 2.88, and it requires no knowledge of cards at all.
The last row is the argument of this page in euros. A player who found the 99.54 percent machine and plays it with a 4.54 point leak pays exactly what the 6 and 5 machine charges a flawless player, which is close to eleven times the hourly cost of doing both things right.
Speed matters on top of all of it. A thousand hands an hour at the same stake doubles every euro figure in that table without changing a single percentage.
One Chart Does Not Cover Every Machine
Everything above is Jacks or Better. That is the game whose pay tables and hold priorities are published on the video poker guide here, and it is where the arithmetic on this page comes from.
Other machines on the same screen are not the same game. When a variant turns cards wild, pays a premium for particular four of a kinds, or pays nothing until a higher minimum hand, the value of each draw moves and the correct hold moves with it.
So the chart is not a general video poker chart, and carrying it across to a different pay schedule is its own leak. The same holds for the three returns in the table above, which belong to Jacks or Better and to no other variant.
The practical rule is narrow and it costs nothing. Learn one game properly, check its pay table every time you sit down, and treat an unfamiliar variant as a game you do not yet have a strategy for rather than one your existing strategy nearly fits.
Autohold Is a Convenience, Not a Strategy
Most video poker software offers to preselect your holds. It is a reasonable way to speed up hands you already know, and a poor way to learn the ones you do not.
The problem is that the feature does not tell you which strategy it is applying, and nothing obliges it to apply the optimal one for the pay table in front of you. A hold that is correct on one pay schedule is not automatically correct on another.
So the useful way to use it is in reverse. Decide your own hold first, then look at what the software selected, and when the two disagree, go back to the chart rather than to the button. The disagreements are where the learning is.
Two other interface habits are worth building. Check the pay table on the glass before every session rather than assuming the machine is the one you played last time, and confirm the coin counter reads five before the first deal.
How I Checked These Numbers
The three returns and the eight rule hold chart come from the video poker guide on this site, which publishes the full pay table ladder and the chart in full. The slot comparison comes from the all casino games page here. Both carry the date they were last checked.
The error table is that chart read in reverse. I did not add a single new ranking to it, because the moment a tips page starts inventing hold rules it stops being checkable.
The euro figures are mine and all of them are one multiplication. A return of 99.54 percent leaves 0.46 percent of turnover with the house, so 625 euros of turnover costs 2.88. You can repeat every number in the tables on a phone.
One figure is deliberately missing. I have not told you what the average player’s strategy costs in points, because I have no source for it I would stand behind, and a page about precision is a poor place to start guessing.
Where to Go Next
The decisions are above. These pages carry the pay tables and the numbers behind them.
Where to Play Video Poker for Real Money
A pay table is chosen by the operator and not by you, so where you play decides which of those three returns you are offered. Each name here has a full review behind it, with all four marks written out.
FAQ About Video Poker Tips and Strategy
Does strategy really change what video poker returns?
Yes, and more than the pay table does. The 99.54 percent printed on a 9 and 6 machine is the figure for a player who makes the correct discard every hand. A 9 and 6 machine is worth only 2.24 points more than the common 8 and 5 version, so a strategy leak larger than that has already cost you more than finding the better machine gained.
Which pay table should I be looking for?
On Jacks or Better, the one paying 9 per coin for a full house and 6 for a flush. That is the 99.54 percent build. The same game pays 97.30 percent at 8 and 5 and 95.00 percent at 6 and 5, and all three look identical until you read those two lines on the glass.
Why does everyone insist on betting five coins?
Because the royal flush pays 800 per coin only at five coins and drops to 250 per coin below that, which strips about 1.4 points off the return. If five coins is more than you want to risk, lower the coin value instead of the coin count.
Is it ever correct to break a made flush?
Yes, when the flush contains four cards to a royal. The priority chart ranks four to a royal above a made flush, a made straight and a made full house, because the royal supplies about two of the 99.54 points on its own.
Should I keep a high card alongside a pair?
No. No rule on the chart holds a pair plus a fifth card, and the kicker removes one of the three cards that could turn the pair into trips or quads. Hold the two cards of the pair and draw three.
Can I just use the autohold button?
Use it to check yourself rather than to decide. The software does not tell you which strategy it applies, and a hold that is correct on one pay schedule is not automatically correct on another, so when your choice and its choice differ, go back to the chart.
Can perfect play make video poker profitable?
No. The best pay table discussed here returns 99.54 percent with flawless play, which is still below 100, so the expectation stays negative. Perfect play makes the game cheap to play for a long time, and that is the entire claim.
Why does my session never feel like 99.54 percent?
Because about two of those points arrive in the royal flush, which lands roughly once every 40,000 hands. Strip the royal out and the game returns about 97.6 percent, and that is the number almost every real session actually runs at.
Does playing faster change what I lose?
It changes the amount, not the percentage. Five hundred hands an hour at 1.25 a hand is 625 euros of turnover, costing 2.88 euros at 99.54 percent. Double the speed and the hour costs twice as much, with the return exactly where it was.
Is video poker a better bet than slots?
On the numbers, a 9 and 6 machine played correctly returns 99.54 percent, against 96.80 percent for the highest verified return among the slots reviewed here. The condition is the one that matters: the slot figure needs no skill at all, and the video poker figure needs every hand played right.
Who Wrote This
I came to this from finance, where nobody asks whether an instrument feels lucky. They ask what it costs to hold and for how long. Video poker is the one casino game that answers both questions precisely, provided the player holds up their end of the calculation.
I am responsible for everything published here. The returns and the hold priorities come from the video poker guide on this site, which carries the date each was last checked. The euro figures are my arithmetic and every one of them is a single multiplication you can repeat. If one stops holding, write to me and I will correct it and date the change.
18+. Gambling can be addictive. Nothing on this page makes video poker profitable, because 99.54 percent is still less than everything you put through the machine. What perfect play does is make the game slow to take your money, and that is the only honest promise anybody can make about strategy. Decide what a session costs before you sit down, bet five coins at a denomination you can afford to, and treat a withdrawal as the end of the session rather than an interruption of it. If gambling has stopped being entertainment, free and confidential help is available in Slovenia through NIJZ.
