2/5 poker bankroll and buy-ins by live stake

One number per stake, not one number for poker

Ask how big a poker bankroll should be and you will usually hear a single figure, 20 or 30 buy-ins. The honest answer is a schedule, not a number. The buy-ins you need at a given stake are set by two inputs: the net win rate you can actually hold in that game, and the size of your swings. The swings barely change as you climb, roughly 80 to 120 BB/100 standard deviation for live full ring at every rung, while the win rate thins out. Same variance against a smaller edge means more buy-ins.

Many players cannot run this calculation because a buy-in and cash-out log does not contain the inputs. Without a net win rate per stake, any buy-in count is a guess borrowed from someone else's game. The rest of this page shows what the number should be once you have real inputs, and what to assume until you do.

The maths that sets the count

The standard risk-of-ruin model gives the bankroll needed as the standard deviation squared, times the natural log of one over your acceptable risk, divided by twice your win rate. That is the same formula behind the bankroll calculator on this site, so every figure here can be checked against it.

Worked example, using the calculator's default inputs: net win rate 4 BB/100, standard deviation 90 BB/100, acceptable risk of ruin 5%. Step 1: square the standard deviation, 90 × 90 = 8,100. Step 2: multiply by ln(1/0.05), which is about 3.0, giving 8,100 × 3.0 = 24,300. Step 3: divide by twice the win rate, 24,300 ÷ 8 = 3,037 BB. Step 4: divide by 100 BB per buy-in, which lands at roughly 30 buy-ins. At 2/5, where a 100 BB buy-in is about £500, that is roughly £15,000.

Note what drives the result: the win rate sits in the denominator. Halve the edge to 2 BB/100 and the same sums demand about 61 buy-ins. Nothing about the game changed except how much of it you keep.

Buy-ins by win rate

Run the same formula across the win rates live players actually hold and the pattern is stark. The table fixes standard deviation at 90 BB/100, the default the site calculators use, and shows buy-ins needed at a 5% and a 1% risk of ruin, plus the expected depth of a normal downswing from the downswing calculator's model.

Under this model, a smaller assumed positive edge requires more capital for the same modelled ruin probability. That relationship does not establish the edge or prescribe a bankroll. A characteristic drawdown scale is not an expected deepest downswing, and a real loss can exceed it.

Rake rewrites the answer at 1/2

At 5/10 the rake cap is a rounding error on most pots. At 1/2 it is the toughest opponent at the table. A fixed cap takes a far larger share of a £60 pot than a £600 one, which is why the same rake structure is nearly invisible at high stakes and brutal at low ones.

Put numbers on it. A live 1/2 game deals roughly 27 hands an hour nine-handed, so an average player drags about three pots an hour. If the average rake on a dragged pot is £4, that is £12 an hour paid to the room, or 6 BB/hr at a £2 big blind. A player whose gross edge is 12 BB/hr keeps 6: half the win rate went to the rake before the dealer was tipped. Add a £1 tip per dragged pot and another 1.5 BB/hr is gone.

Now connect that to the bankroll formula. The win rate sits in the denominator, so halving the net edge doubles the buy-ins needed at the same risk of ruin. This is the trap at 1/2: the game feels soft, the gross graph looks healthy, and the honest buy-in count is higher than the table feel suggests, not lower. Log net of rake and tips or the input to every calculation on this page is wrong.

The ladder, stake by stake

The table below compares illustrative buy-in budgets by stake. It assumes a chosen positive edge for the model; a winning log alone does not establish that edge, and the figures do not guarantee safety at any stake.

The shape of the ladder matters more than any single row. From 1/2 to 2/5 the buy-in count barely moves, because the rake relief roughly offsets the tougher field, but the money required rises 2.5x. From 2/5 to 5/10 both move: the field strength jumps more than the stakes do, the edge thins, and the safe count climbs to 40 buy-ins or more. The 5/10 row is not just double the 2/5 row, it is a different risk profile.

The gates: moving up and moving down

A per-stake bankroll only works with rules at the boundaries. Moving up has two gates, and both must clear. Gate one is the buy-in count: divide your roll by the next stake's buy-in and check the risk of ruin at that count is one you can accept, which is exactly the check the move-up readiness calculator runs.

Gate two is a proven win rate at your current stake over a real sample. Live poker deals 25 to 30 hands an hour, so even 500 hours is only 12,500 to 15,000 hands; be honest about how wide the error bars still are before you promote yourself. A roll built by one upswing satisfies gate one and fails gate two, and that combination is how most failed shots start.

Moving down needs one gate, set in advance. Pick the buy-in count at which you drop a stake, write it down while you are calm, and treat it as binding. A sensible trigger sits at the bottom of your floor range: if 2/5 needs 25 buy-ins minimum, dropping to 1/2 when the roll falls below that keeps risk of ruin controlled while you rebuild. The players who bust are rarely the ones with the wrong number. They are the ones who renegotiated the number mid-downswing.

Where the inputs come from

The model needs a win-rate assumption and a compatible standard deviation. Dated session records with stake, playing time and reconciled results can support estimates; short or changing samples remain uncertain. Cash-out already reflects costs paid from the stack, so only additional costs should be subtracted. A list of cash totals without hours cannot establish an hourly rate.

Once the log exists, the hourly rate calculator turns it into BB/hr so stakes compare honestly, and the win rate confidence calculator shows whether your edge is statistically clear of zero yet or still inside the noise. Size the roll from those outputs, set the floors and triggers per stake, and the bankroll question stops being a debate and becomes arithmetic you re-run as the sample grows.

Moving up in stakes

Moving up means committing to a higher stake once both your bankroll and your win rate support it. Done on a hot streak instead of a proven edge, it quietly hands the money back.

Two gates, not one

A larger bankroll can reduce modelled ruin risk when the assumptions hold, but no buy-in count guarantees that variance will not exhaust it. A winning sample at one stake does not establish an edge at a different stake. Set spending limits independently of any calculator output.

Take a shot, not a leap

Ring-fence a fixed number of buy-ins for a shot at the higher stake, decide your drop-back point in advance, and treat it as an experiment. If the shot money goes, you return to your stake with information, not a dented roll.

Dropping back is not failure

A pre-set rule to move down when your buy-in count falls is discipline. The players who survive stake jumps are the ones who move down without ego when the maths says to.

Cash vs tournament bankroll

Cash and tournaments are different games for bankroll purposes. A roll that is healthy for 2/5 NL would be reckless for a tournament schedule, because tournament variance is in another league.

Why tournaments need more

Tournament prize pools are top-heavy: you cash a fraction of the time and most of your profit comes from rare deep runs. That produces enormous variance and long dry spells, so tournament players need many more buy-ins, often 100+ for a regular schedule, than a cash player needs for the same comfort.

ROI, not hourly

Cash is measured in hourly and BB/hr; tournaments in ROI, profit as a percentage of total buy-ins. A solid tournament ROI is small and slow to confirm because the sample of cashes is tiny relative to the events played.

Keep the rolls separate

Mixing live cash and tournament money into one graph hides two very different win rates and risk profiles. Separate bankrolls let each tell its honest story, and stop a tournament dry spell from making your cash game look broken.

Moving down in stakes

Moving down means stepping back to a lower stake to protect your bankroll when the roll dips. The players who survive stake jumps are the ones who move down on a rule, before ego gets a vote.

A rule, set before you need it

Decide the buy-in count at which you drop down, in advance, while you are calm. A pre-set trigger, for example dropping from 5/10 to 2/5 when your roll falls below a fixed number of 5/10 buy-ins, takes the decision away from tilt and fatigue.

It is protection, not punishment

A smaller stake reduces the money exposed in a buy-in; the games, edge and variability can also change. Moving down does not guarantee recovery, and taking a break or stopping remains an option. Decide within a pre-set spending policy rather than trying to win a loss back.

Coming back up

Move back up on the same two gates you would for any move up: the buy-in count and a win rate you trust at the lower stake. Rebuild the roll, then climb, rather than jumping straight back the moment you feel better.

Shot-taking, done right

Shot-taking is a small, ring-fenced attempt at a stake above your usual, with the budget and the exit decided before you sit. Done well it costs little and teaches a lot; done on a whim it just hands the roll back.

Ring-fence the money

Set aside a fixed number of buy-ins you are willing to lose on the shot, separate from your core roll. When that money is gone, the shot is over. This caps the downside and stops a bad run at the higher stake from denting the bankroll you actually live on.

Decide the exit before you sit

Write down the drop-back rule in advance: if the shot money goes, you return to your normal stake with information, not a chase. Deciding the exit while calm is the whole point, because mid-session you will always find a reason to stay.

Treat it as an experiment

A shot is a test of whether you can beat the higher game, not a promotion. Mark hands and review the shot afterwards. Even a losing shot is useful if it shows you where the tougher field punished you.

Live vs online bankroll

Live and online cash are the same game with different maths for bankroll purposes. Hand volume, variance and how you measure your edge all differ, so the same roll is not equally safe across both.

Volume changes everything

Online you can play hundreds of hands an hour across tables, so your sample grows fast and your win rate stabilises sooner. Live you see a fraction of that, so your sample is always smaller than it feels and short-term results mislead for longer.

Different swings, different metrics

Online short-handed play often runs a higher standard deviation than live full-ring, which demands more buy-ins for the same comfort. Live leans on BB/hr and hours; online leans on BB/100 and exact hand counts. Do not compare the two as if the numbers mean the same thing.

Keep the rolls readable

If you play both, track them separately so a live dry spell does not make your online game look broken, and so each win rate is measured against its own sample and variance.

Bankroll for 1/2 NL live cash

At 1/2 NL a 100 big-blind buy-in is around £200. The edge is usually large against a soft field, but the rake is the quiet story: at this stake a capped rake takes a big share of every pot, so the real win rate is smaller than the table feels.

25 to 40 buy-ins, roughly £5,000 to £8,000.

Rake is the main character

A fixed rake cap is a much bigger slice of a 1/2 pot than a 5/10 one. The impact depends on the actual structure, pots and game. A cash-out-based result already reflects pot rake; an attributed add-back is not a forecast of no-rake play. That means the buy-in count you need is not as low as the soft games suggest.

A sensible count

A serious 1/2 player usually wants something like 25 to 40 buy-ins, roughly £5,000 to £8,000, depending on how well they actually beat the rake. Cash-outs minus all buy-ins already reflect pot rake and stack-paid tips or fees. Deduct only additional costs not already included, and label the hourly cost/time basis.

It is a launchpad, not a ceiling

1/2 is where most live players start and where record-keeping habits form. Build the roll, prove a real win rate net of rake, and use the move-up gates before stepping into 2/5.

Bankroll for 2/5 NL live cash

2/5 NL, with a buy-in around £500, is the standard serious live cash stake. The rake bites less than at 1/2, the edges are real for a studying player, and the swings are large enough that the buy-in count matters.

25 to 40 buy-ins, roughly £12,500 to £20,000.

The semi-pro stake

2/5 is a bigger step in money at risk: a typical buy-in is two and a half times the 1/2 buy-in. The rake is a smaller share of the pot than at 1/2, so a studying player can hold a solid win rate, but the fields are a step tougher and game selection starts to matter more.

A sensible count

A common range is 25 to 40 buy-ins, roughly £12,500 to £20,000. Sit towards the higher end if your sample is thin or your swings run large, and use the calculator to size it for a risk of ruin you can accept.

Hold here before climbing

Prove a real win rate at 2/5 over a meaningful sample before taking shots at 5/10. The jump in field strength is bigger than the jump in stakes, so a shot taken on an upswing rather than a proven edge usually hands the money back.

Bankroll for 5/10 NL live cash

5/10 NL buys in for around £1,000 to £1,500. The rake is a small share of the pot here, but the field is tougher, so your edge is usually thinner, and a thinner edge against the same swings demands more buy-ins.

40 or more buy-ins, roughly £40,000 to £60,000.

Thinner edge, more buy-ins

At 5/10 the recreational players are fewer and the regs are stronger, so the win rate you can hold is usually smaller than at 2/5. The standard deviation does not shrink to match, so the safe buy-in count rises even though the rake is no longer the problem.

A sensible count

A common floor is 40 or more buy-ins, roughly £40,000 to £60,000. Be honest about your edge here rather than assuming the win rate that beat 2/5 carries up, because it usually does not, at least not at first.

Take the shot, do not leap

Most players reach 5/10 through ring-fenced shots, not a single jump. Set a shot budget and a drop-back point, prove the edge, and move up permanently only when both the buy-in count and a real win rate clear the bar.