How to Choose a Video Poker Bankroll

Choose your video poker bankroll based first on what you can afford to lose—not an amount supposedly needed to win. Then check the full stake, paytable, session length, and acceptable risk of running out of money. If the planned session does not fit your budget, lower the stake, shorten the session, or do not play. You do not need to replenish the bankroll to keep going.

Expected loss is usually much smaller than the amount that could be consumed by an unfavorable run of hands. For example, in the 9/6 Jacks or Better model below, a $1.25 full stake produces an expected loss of $34.50 over 10 hours, yet a $250 bankroll was exhausted before the end in about 46.0% of simulated sessions. The average does not describe an individual session’s risk.

Session, trip, and lifetime bankrolls answer different questions

For a session, assess the risk of exhausting your budget within a fixed number of hands. For a trip, set an overall affordable loss limit. Unlimited play is a separate risk-of-ruin problem.

• Session bankroll: money set aside for one time-limited session.
• Trip budget: the total loss limit across all planned sessions. Dividing it into daily amounts does not make additional losses affordable.
• Unlimited-horizon risk of ruin: the probability of exhausting a bankroll when the stake stays fixed and there is no limit on the number of hands.

In the negative-expected-value model, with no cashback or replenishment, a finite bankroll eventually runs out with probability 1 if fixed-stake play continues indefinitely. That does not mean every short session loses the entire bankroll.

Keep three events separate: ending a session at a loss, exhausting the bankroll during the session, and eventual ruin over an unlimited horizon. Their probabilities differ. For broader budget boundaries, see the bankroll management guide.

How to check the full stake and paytable

Count the cost of the entire hand, not one credit, and check the complete paytable in the game you intend to play.

For a conventional five-credit bet:

The final column is not a prediction of how many hands you will last. Payouts can fund subsequent bets, so the actual number depends on their sequence.

The name 9/6 Jacks or Better identifies the full house and flush payouts, not the entire schedule. The model used here has these gross returns per full stake:

These are gross payouts, not net profits. A hand’s net result is its payout minus the stake. A return of 1 gives the stake back; it does not earn a profit equal to the stake.

Before playing:

1. Open the game’s paytable and check every payout.
2. Confirm the denomination and credits in the full stake.
3. Check what activates any enhanced royal flush payout.
4. Compare the total cost per hand with your budget.

In the conventional schedule, a royal flush pays 250 credits per credit bet at one through four credits, but 4,000 credits total at five credits—800 per credit. However, the max-coin royal bonus depends on the actual paytable. Do not impose a “must bet five coins” rule on a game that already pays a flat 800 times the stake.

If the enhanced payout requires five credits, prefer a lower denomination with the qualifying stake rather than spending beyond your budget. If that is still too expensive, shorten the session or do not play.

Historical Jacks or Better paytable in game rules hosted by Loto-Québec

Loto-Québec implementation example from the English document “EN04-09-2024,” page 6. Source: official PDF.

This is a historical example, not confirmation of a casino’s current paytable or a five-credit activation requirement. The multi-hand game described there is not identical to the conventional single-line model, and its default strategy is not always optimal.

Why equal turnover does not mean equal risk

Expected loss depends on RTP and turnover, while the size of possible swings also depends on the payout distribution.

RTP is the long-run percentage of wagers returned through gross payouts. It is not the probability of winning a hand or finishing a session with a net profit.

An independent calculation from the exact final-hand counts for optimal 9/6 Jacks or Better gives RTP of 99.543904% and standard deviation of 4.417542 full-stake units. The money examples below use rounded values of 99.54% and 4.42. For 8/5 Jacks or Better, RTP is 97.30% with its corresponding optimal strategy. The schedules and underlying counts appear in the Wizard of Odds Jacks or Better calculations, updated August 3, 2026. This is specialist mathematical analysis, not provider certification.

At the same $7,500 turnover:

Expected loss is about 5.87 times higher in the second case. That does not establish a universal percentage increase in bankroll: the ratio of average losses is not the ratio of ruin probabilities. This comparison does not assess the two games’ standard deviations.

Turnover is the sum of all wagers, including recycled payouts. The formula turnover = full stake × number of hands does not determine the starting cash required.

How often does a bankroll run out over 2 or 10 hours?

In the single-line 9/6 Jacks or Better model at $1.25 per hand, bankroll exhaustion becomes much more likely as the session grows longer. The simulations track the first time the bankroll reaches zero, rather than just the ending result.

The assumptions are:

• optimal 9/6 Jacks or Better strategy;
• independent, single-line hands;
• $1.25 full stake;
• 600 hands per hour;
• no cashback, taxes, or replenishment;
• no early stop at a profit target.

Each bankroll-and-duration combination used 50,000 simulated paths, with NumPy 2.3.5 and random seed 20261012. These are computational simulations, not real-money play.

To reproduce the method, take the exact optimal-strategy final-hand probabilities from the linked Wizard of Odds table and match them to payouts 800, 50, 25, 9, 6, 4, 3, 2, 1, 0. Independently sample a payout for each hand, subtract the full stake, and add the gross payout. Record the first exhaustion and stop that path; do not allow play to continue after the bankroll runs out. Here, every starting bankroll and payout is a multiple of the full stake, so inability to fund the next bet coincides with reaching zero.

The 95% Wilson intervals for the exhaustion probabilities, expressed as percentages, are:

No observed exhaustion among the 50,000 two-hour paths starting at $500 does not mean zero risk. These intervals describe sampling uncertainty only. They do not cover strategy mistakes, a different paytable, or mismatch between the model and an actual game.

At $500, about 94.5% of the ten-hour paths completed—not a verified 95% guarantee. At $250, two-hour risk is far lower than ten-hour risk. Do not automatically apply these results to other games or multiple lines.

Why mean loss and standard deviation do not give “95% survival”

Mean loss and standard deviation help show the scale of possible swings, but they do not replace a calculation of whether money runs out along the way.

Let n be the number of hands, b the full stake, and r RTP expressed as a decimal. For independent hands:

• expected loss: n × b × (1 − r);
• standard deviation of the total result: b × 4.42 × √n.

Standard deviation measures the spread of results; variance is its square. Using a $1.25 stake, 99.54% RTP, and 600 hands per hour gives:

These monetary figures use the rounded RTP and standard deviation.

For ten hours, 34.50 + 1.645 × 427.96 gives $738.50, or about $739. This is only an approximate terminal-loss threshold under a normal-distribution assumption. It is not a bankroll with a 95% probability of completing the session.

There are two reasons. First, a path can run out of money before a later hypothetical recovery; an ending result does not reveal that earlier exhaustion. Second, rare royal flushes make the distribution asymmetric, and a normal approximation can poorly describe its tails. The $739 figure is neither a financial recommendation nor a confidence interval for your actual cash balance.

A royal flush does not arrive on a schedule

A royal flush remains rare, and a long drought does not increase its probability on the next independent hand.

In this model, its per-hand probability is 0.000024758268, or approximately one royal flush per 40,390.55 hands on average. That average interval is not a payout deadline.

If p is the probability of a royal flush on one hand and n is the number of independent hands, the probability of at least one is 1 − (1 − p)^n.

The final row is an unconditional calculation for a specified number of independent hands. It does not account for exhausting a bankroll earlier, establish a bankroll requirement, or recommend playing that many hands.

At a $0.25 denomination, a 4,000-credit payout is $1,000 gross, including the wager. Do not budget for it as a scheduled replenishment.

What cashback and positive expected value change

Cashback can change expected value, but only when its actual terms match the calculation. Positive expected value does not mean low risk or guaranteed net profit.

The Wizard of Odds risk-of-ruin table, updated August 3, 2026, defines one betting unit for 9/6 Jacks or Better as five credits and credits cashback after every hand.

For example, at 0.5% cashback, the table gives 65,928 full-stake units for 5% unlimited-horizon risk of ruin. At a $0.25 denomination, that is $82,410, because each full stake is $1.25.

This is not an ordinary session budget, a no-cashback calculation, or “a 95% chance of lasting ten hours.” It illustrates how different fixed-session and unlimited-horizon questions are.

Deferred cashback, conditional free play, and hotel or meal comps cannot be inserted into this model as cash credited after every hand. Do not increase your budget merely because an advertised reward appears to make expected value positive.

How to check your decisions in the trainer

Use the video poker trainer to compare the expected value of hold decisions. It does not calculate a bankroll, risk of ruin, or the result of a future cash session.

The trainer supports Jacks or Better presets 9/6, 8/5, 7/5, and 6/5. For a given hand, its exact solver considers all 32 hold choices and possible draws from the 47 unseen cards.

To practice:

1. Select the 9/6 paytable.
2. Look at the current five cards.
3. Select the cards you want to keep, or leave all cards unselected.
4. Click “Check my hold.”
5. Compare your decision’s expected value with the optimal value.
6. Click “Show optimal” to see the recommended cards.
7. Move to a new hand and repeat.

English video poker trainer comparing gross expected values of 0.316 and 0.486 for 10♣ 4♦ J♠ K♥ 8♦

Local trainer interface: 9/6 paytable and a hold-decision example for one hand.

For the displayed 10♣ 4♦ J♠ K♥ 8♦ hand, keeping no cards gives expected value 0.316, while the optimal choice, J♠ K♥, gives 0.486. Both values are expected gross payouts per full stake, not win probabilities.

The rounded difference is 0.17 full-stake units. At $1.25 per hand, that is about $0.21 in expected error cost each time this particular decision occurs. It is not a guaranteed loss on that hand or an amount you can simply subtract from the game’s overall RTP.

The cards may change after you request a new hand. The trainer helps analyze decisions; the link does not promise to reproduce this particular hand.

Why multiple lines and extra features need a different model

For multi-line play, calculate the total cost of the deal and do not use the single-line risk table without a new model.

The basic cost calculation is:

total wager = number of lines × stake per line + feature cost

When lines share an initial hand and the same hold decision, their results are dependent. Even separate virtual draw decks do not remove the dependence created by the shared starting hand. You cannot simply multiply the hand count by the number of lines and insert it into the single-line table.

Extra features and progressive payouts can change the price, payout distribution, and optimal decisions. Check the exact game model. The Dream Card strategy guide examines a feature-cost example. Another variant, such as Deuces Wild, needs its own strategy; the 9/6 Jacks or Better results do not transfer to it.

A practical session plan and recordkeeping routine

Set your budget and time limit before playing, then stop when either limit is reached. The plan controls spending; it does not change RTP.

1. Set an affordable loss limit. Write down the total trip budget and the separate session loss limit.
2. Check the game. Record the complete paytable, denomination, full stake, and royal flush payout conditions.
3. Estimate duration. Hand count depends on time and speed. The 600 hands per hour used here is a modeling assumption, not a promised pace.
4. Match risk to your budget. Use the simulation table only when its assumptions fit. Otherwise, choose a lower stake, less time, or no play.
5. Practice decisions in the trainer. Do not treat published optimal-strategy RTP as your own result if you make strategy mistakes.
6. Follow the session stop rule. Stop at the time or loss limit. Do not use an ATM, debt, or a refill to chase losses.

Afterward, record the date, game, complete paytable, denomination, full stake, hands played, time, starting balance, ending balance, and net result. Track cashback separately.

A deposit is not a win. If you made additional deposits, calculate the playing result as ending balance + withdrawals − starting balance − additional deposits, separately identifying cashback and other credits.

The guiding rule is simple: a suitable bankroll is not an amount that lets you continue at any cost. It is a loss boundary chosen in advance—not one you expand to chase a recovery.