When to Move Up in Stakes: Bankroll, Sample Size and the Shot Rule

Moving up is where most bankrolls get broken. Not at the stake where a player is comfortable, but in the week after a heater, when the next level looks beatable and the bankroll "almost" covers it.

The decision gets easier once you split it into three questions you can answer with numbers: does the bankroll cover the new stake, how sure are you that you win where you are, and what happens if it goes wrong.

Key takeaways

• Move up when the bankroll covers the new stake under your rule, calculated with a lower win rate than your current one.
• You will never be sure of your win rate before moving up: knowing it within ±2 bb/100 takes close to 800,000 hands. The bankroll exists to make that uncertainty affordable.
• A shot is a fixed budget at the higher stake with a written exit on both sides. Losing it is not a verdict on your game.
• A 10 buy-in shot takes about 12,000 hands to settle on average, around 400 hours live.

Condition 1: the bankroll covers the new stake

Start from the rule you already use. If your bankroll plan holds the risk of ruin at 5%, the question is whether your bankroll is big enough for that at the next stake, not at this one.

Two details change the answer. The first is that a bigger stake means bigger buy-ins: a bankroll of 60 buy-ins at NL50 is 30 at NL100. The second is that your win rate at the new stake will almost certainly be lower, because the players are better. Size the move with that lower win rate.

For a player who wins 3 bb/100 at a standard deviation of 90 bb/100, a 5% risk of ruin takes about 41 buy-ins. If the same player expects 2 bb/100 at the next stake, the requirement becomes about 61 buy-ins: half again as much, for a win rate one point lower. The poker bankroll calculator gives the figure for your own inputs and risk level, and poker bankroll management explains where it comes from.

Condition 2: you win where you are now

It would be convenient to wait until you are certain you beat your current stake. You will not get there. The 95% range around an observed win rate is ±1.96 × standard deviation / √(hands / 100), and at 90 bb/100:

• after 50,000 hands, a 3 bb/100 result means a true win rate somewhere between about −5 and +11 bb/100;
• narrowing the range to ±2 bb/100 takes close to 800,000 hands.

So the honest version of this condition is weaker: you have been winning over a meaningful sample, tens of thousands of hands rather than a few sessions, and nothing in your own numbers, such as one stake or one format carrying all the profit, says the result is a fluke. Downswing or leak covers how to read your results while the range is still wide.

What makes it safe to act without certainty is condition 1. The bankroll is sized so that being wrong about your win rate costs you a stake, not the whole roll.

Condition 3: a written exit

Before the first hand at the new stake, decide the bankroll level at which you move back. Use buy-ins of the new stake, so the line does not drift. One workable version uses two risk levels: play the stake while the bankroll holds the 5% figure, and move down if it drops below the 10% figure. For the 2 bb/100 player above, that means playing NL100 with 61 buy-ins or more and moving down below 47.

Moving down is part of the plan, not a failure of it. The players who go broke moving up are mostly the ones who decided where to stop after they started losing.

Taking a shot

Often the bankroll is between stakes: comfortably enough for the current one, not yet enough for the next. A shot lets you play the next stake anyway, with a strict limit on what it can cost.

1. Set the budget in buy-ins of the new stake. Five to ten is typical.
2. Check what is left if the shot fails. The rest of the bankroll must still cover your current stake under your rule.
3. Set both lines. Back down when the budget is gone. Stay up when the bankroll reaches the full requirement for the new stake.
4. Play your normal volume and game selection. A shot played tired or at the worst tables tests the wrong thing.

For example, a player with $3,500 at NL50 has 70 buy-ins there. A 10 buy-in shot at NL100 is $1,000. If it fails, $2,500 is still 50 buy-ins of NL50, enough under most rules. If it succeeds, the bankroll grows towards the NL100 requirement and the shot turns into the player's new stake.

What a shot can and cannot tell you

A shot ends at one of two lines, and which one it hits first is mostly luck. The table runs a shot that ends after losing 10 buy-ins or winning 10, for players of different true skill at the new stake:

Outcome of a 10 buy-in shot with a 10 buy-in target

True win rate at the new stake	Wins the shot first	Loses the shot first	Average hands to settle
-2 bb/100	38%	62%	12,000
0 bb/100	50%	50%	12,000
1 bb/100	56%	44%	12,000
3 bb/100	68%	32%	12,000
5 bb/100	77%	23%	11,000

A genuine 3 bb/100 winner loses this shot about a third of the time, and a player who loses 2 bb/100 at the new stake wins it more than a third of the time. Suppose you start with no idea whether you are a 3 bb/100 winner or a break-even player at the new stake, even odds. Winning the shot moves those odds to only about 58 to 42. Losing it moves them to about 39 to 61. Either way, you know a little more, not a lot more.

That is the real purpose of a shot. It is not a test you pass or fail; it is a way to keep playing the higher stake for as long as the bankroll can afford it, while the sample that actually answers the question builds up.

A bigger target makes the shot longer and the result more informative, at the cost of more time at risk:

Outcome of a 10 buy-in shot with a 20 buy-in target

True win rate at the new stake	Wins the shot first	Loses the shot first	Average hands to settle
-2 bb/100	19%	81%	22,000
0 bb/100	33%	67%	25,000
1 bb/100	42%	58%	25,000
3 bb/100	59%	41%	25,000
5 bb/100	73%	27%	24,000

Live players: same maths, far fewer hands

Live games deal about 30 hands an hour, so everything above takes longer. The 10 buy-in shot that settles in roughly 12,000 hands is about 40 hours online at four tables of 75 hands, and about 400 hours live. That is several months of regular play for one attempt at the next stake.

Live stakes also jump in bigger steps. Going from 1/2 to 2/5 multiplies the big blind by two and a half, so the bankroll requirement grows by the same factor. Plan live shots in buy-ins of the new game and expect them to take a season, not a weekend.

Tournaments: move up by average buy-in

Tournament players move up by raising their average buy-in, and the same three conditions apply with bigger numbers. Tournament results swing so much that a shot of a few dozen tournaments says almost nothing: a 10% ROI player in fields of 100 to 200 is still behind after 50 tournaments about 43% of the time, against 50% for a break-even player.

A sensible tournament shot is therefore a small share of the schedule, for example one higher buy-in a week instead of a separate block, with the same written rule for stopping. The poker bankroll calculator shows how many buy-ins each field size needs.

A checklist before you move up

1. Does the bankroll cover the new stake under my rule, with a lower win rate than I have now?
2. Have I been winning at my current stake over tens of thousands of hands, across formats and venues rather than in one hot spot?
3. Have I written down the bankroll level at which I move back?
4. If this is a shot: how many buy-ins of the new stake, and is the rest enough for my current stake?
5. Will I play my normal volume and game selection at the new stake?

Every answer depends on numbers you can only get by tracking: your results at each stake, your hours, and your bankroll over time. PokerRoll records them from the sessions you log, live or online, so these questions have answers instead of impressions. The Poker Bank also works through how long climbing the stakes takes at a given win rate and volume.