What is a good standard deviation for poker?

For NLH 6-max, typical standard deviations range from 75-120 BB/100. 100 BB/100 is a reasonable default. PLO players typically see 150-200 BB/100. Lower standard deviation reflects playing style rather than better results.

How many hands do I need to prove my winrate?

At 100k hands with a 5 BB/100 winrate and 100 BB/100 std dev, there is still a meaningful probability of showing a loss. Most players need 300k-500k hands to statistically confirm a positive winrate at typical win rates.

What is risk of ruin in poker?

Risk of ruin is the probability that your bankroll reaches zero before you can replenish it. The formula accounts for your edge (winrate), variance (standard deviation), and bankroll size. A common target is 5% risk of ruin.

What is a good winrate in poker?

5+ BB/100 is a strong winrate at most stakes. 2-3 BB/100 is a solid small-stakes regular. Anything under 1 BB/100 is easily erased by rake and is functionally breakeven or losing in most games.

What causes high standard deviation in poker?

Playing style (aggression, 3-bet frequency), pot size relative to stack depth, and game format. PLO runs 150 to 200 BB/100 against 75 to 120 for NLH. Deep-stack play and high-variance bluffing lines also increase standard deviation.

How does rake affect my variance?

Rake reduces your effective winrate without changing your standard deviation. A 5 BB/100 player paying 3 BB/100 in rake has an effective winrate of 2 BB/100. Lower effective winrate means longer downswings, higher bankroll requirements, and a lower probability of showing a profit over any given sample.

Cash Game Variance Calculator

Enter your win rate and standard deviation to see confidence bands, downswing odds, bankroll requirements and whether your results prove you are a winner.

Enter Game Parameters

Winrate in BB / 100

Observed winrate in BB / 100 (optional)

Standard deviation BB / 100

Number of hands

Standard deviation guide

NLH full ring	60–80 BB/100
NLH 6-max	75–120 BB/100
PLO full ring	100–140 BB/100
PLO 6-max	120–160 BB/100

Find this stat in PokerTracker or HEM.

Expected Value

70% Confidence

95% Confidence

Best Run

Worst Run

Sample Paths

Your Result

Variance in Numbers

Winrate
Standard deviation
Hands
70% confidence interval
95% confidence interval
Probability of loss after hands


Minimum bankroll for <5% risk of ruin

Rake at Your Stake

Stake, 6-max NLHE Room you play

Room	Rake after rakebackBB/100	Rake savedBB/100	Win rateBB/100	Chance of losingover 100k hands

Win rate at each room is the rate you entered plus the rake saved. Every room and format is in the rake calculator.

100k hands

Cumulative Profit

Drawdown from Peak

The blue line tracks cumulative profit over the median simulated run. The red shading shows drawdowns: how far below your all-time high you dropped at any given point. Use the slider to zoom in on shorter stretches. If the red shading is deep and sustained, you need a larger bankroll to survive the swings.

Downswing	Probability

Current bankroll BB

Kelly Stake Sizing

Optimal stake selection for your bankroll under the Kelly Criterion. Half-Kelly is the practical default. It gives up 25% of theoretical growth for 50% less variance. That buffer matters because your edge estimate is never as precise as the math assumes.

Rolled to playmax safe stake at your stated winrate	--
Kelly fractionoptimal % of bankroll per buy-in	--
Half-Kelly buy-insrecommended bankroll depth	--

Stake	Kelly utilization	Recommendation

Growth rate comparison

Strategy	Buy-ins required	Growth / 100 hands	Drawdown risk

For shareable URLs and method notes, see the standalone Kelly Criterion Calculator.

Risk of Ruin Simulator

Monte Carlo simulation of 10,000 bankroll paths plus the Malmuth analytical formula as a cross-check. Drawdowns matter more than the headline RoR for most players: a 5% RoR bankroll can still drop 40 to 50 buy-ins before recovering. Required bankroll by risk level is included in the output.

Risk of ruin

Analytical RoRMalmuth formula	--
Monte Carlo RoR10,000 simulations	--
Probability of profit	--

Maximum drawdown

Required bankroll by risk level

Risk level	Required BB	Buy-ins (100bb)	Buy-ins (20bb)

Profit distribution

Expected profit	--
Median profitp50 of simulations	--
Worst case5th percentile outcome	--

For full percentile tables and shareable URLs, see the standalone Cash Bankroll Simulator.

Enter Your Actual Results

Hands played

Total BB won/lost

Standard deviation

Enter your hands played and total BB won or lost, then click Analyze.

Reading the Confidence Intervals

What the Bands Mean

The green bands on the simulation chart are confidence intervals, the range of outcomes you should expect given the win rate and standard deviation you entered. A 95% confidence interval means that 95 out of every 100 simulated runs ended within that band at any given hand count, while the remaining 5 landed outside it, above or below.

The narrower 70% band shows where the majority of outcomes cluster: run the same sample 100 times and 70 finish inside it. The region between the 70% and 95% bands is where the unlucky and lucky runs live, and anything outside the 95% band is a statistical outlier at your entered win rate.

Higher Confidence Levels Widen the Range

The more certain you want to be that the true outcome lands inside an interval, the wider the interval has to be. A 70% interval accepts a 30% chance the result falls outside, a 95% interval accepts only 5%, and shrinking that miss rate means extending the boundaries to cover more of the distribution's tails.

In concrete terms: if your win rate is 5 BB/100 and you play 100,000 hands, the 70% interval might span +200 BB to +800 BB, while the 95% interval for the same inputs might span -100 BB to +1,100 BB. The expected value is identical in both cases; the wider interval simply captures more of the distribution, extreme outcomes included.

Reading Your Own Results Against the Bands

When your actual results fall inside the 95% band, the simulation is saying: this outcome is consistent with your entered win rate. It shows only that variance alone can explain where you landed. When your results fall outside the 95% band, the model is flagging a mismatch. Either you ran exceptionally hot or cold, or the win rate you entered does not reflect your actual edge.

Confidence intervals widen as standard deviation increases and narrow as hand count increases. A player with 100 BB/100 standard deviation at NL50 6-max needs four times the sample to achieve the same interval width as a player with 50 BB/100 standard deviation at NL10 full ring. High-variance formats therefore need larger samples before your win rate means anything. Run the bankroll simulator alongside this tool for the same reason.

Standard Deviation Sets the Spread

Win rate determines where the center of the distribution sits. Standard deviation determines how wide the bands spread. Two players with identical win rates but different standard deviations will produce confidence intervals of very different widths. The aggressive player who 3-bets wide and runs big bluffs will have a standard deviation of 90 to 120 BB/100 and correspondingly wide confidence bands. The tight grinder who plays a low-variance style might sit at 50 to 70 BB/100 with much narrower bands.

If you do not know your standard deviation, most tracking software reports it. In 6-max NLHE, 75 to 100 BB/100 is a reasonable default. For full ring, 55 to 75 BB/100. For PLO, 120 to 160 BB/100. Entering the wrong standard deviation will distort the confidence intervals more than entering the wrong win rate.

Methodology

1,000 Monte Carlo simulations per calculation. The simulation walks a random path in 100-hand increments using normally distributed steps centered on your entered win rate with variance scaled to your entered standard deviation. Confidence bands are empirical percentiles (2.5th/15th/85th/97.5th) across all 1,000 runs at each step. The normal distribution assumption holds well for cash games where each hand is an independent trial with bounded outcomes.