Every bettor tracks profit and loss, but very few actually measure how volatile that profit and loss is from bet to bet. Standard deviation is the single number that tells you how wild your swings really are, independent of whether you’re winning or losing overall.
Loading calculator...
This tool takes a list of your individual bet results — in currency or in units/points — and calculates the mean, variance, and standard deviation of that sample. It’s the same statistical foundation used to build risk-adjusted metrics like Sharpe and Sortino ratios for a betting bankroll.
Once you know your standard deviation, you can size your bankroll properly, set realistic expectations for losing streaks, and stop mistaking normal variance for a broken strategy.
📊 How to Use the Standard Deviation Calculator
Paste your bet-by-bet results into the data field — one number per bet, separated by commas, spaces, or new lines. Each number represents the net profit or loss on a single wager, in whatever unit you track (dollars, units, or points).
Use at least 20-30 results if possible. Small samples (under 10 bets) produce a standard deviation that barely reflects your true long-run variance.
Choose Sample standard deviation if this is a partial log you’re still adding to (this is the correct choice almost all of the time), or Population only if the data set represents every bet you will ever count for this analysis, with nothing more to add.
🔢 Calculator Fields Explained
Bet Results – the raw list of profit/loss figures for each individual wager, entered as plain numbers (negative for losses, positive for wins).
Calculation Type – Sample (divides by n-1) or Population (divides by n). Sample is the statistically correct choice for almost every real-world bettor’s log.
Value Type – whether your figures represent currency amounts or abstract betting units/points.
Currency – the currency symbol used to format results, shown only when Value Type is set to Currency.
💰 Understanding the Results
| Result Field | What It Tells You |
|---|---|
| Data Points (n) | How many bets are in your sample – more data means a more reliable estimate |
| Mean (Average) | Your average profit/loss per bet across the sample |
| Variance | The average squared deviation from the mean – the raw building block behind standard deviation |
| Standard Deviation | The typical size of your bet-to-bet swing, in the same units as your original data |
| Coefficient of Variation | Standard deviation expressed as a percentage of your mean – useful for comparing volatility across different stake levels |
| Min / Max / Range | Your single worst result, single best result, and the gap between them |
The headline number is standard deviation itself. A small standard deviation relative to your average stake means your results cluster tightly around your mean outcome bet after bet.
A positive mean does not guarantee a smooth ride. You can be profitable overall and still experience brutal short-term drawdowns if your standard deviation is high relative to your average stake.
Standard deviation measures the size of your swings, not whether you’re winning or losing. Those are two separate questions, and conflating them is the most common misread of this metric.
📐 Calculation Formulas
| Metric | Sample Formula | When to Use |
|---|---|---|
| Sample Variance | Σ(x – mean)² ÷ (n – 1) | Your log is a subset of a larger, ongoing betting history |
| Population Variance | Σ(x – mean)² ÷ n | Your log is the complete, final data set with nothing more to add |
| Standard Deviation | √Variance | Always, once variance is known |
| Coefficient of Variation | (Std Dev ÷ Mean) × 100 | Comparing volatility across different stake sizes or bankrolls |
Dividing by n-1 instead of n for a sample might look like a small tweak, but it corrects a real bias: without it, a partial sample would systematically understate the true variance of your full betting history.
The n-1 adjustment is called Bessel’s correction. It exists specifically because a sample’s own mean is calculated from the same data, which slightly shrinks the apparent spread unless corrected.
In practice, almost every bettor should default to the sample formula, since very few people are ever analyzing a truly closed, final data set.
📝 Practical Examples
Example 1 – Flat staking, moderate variance. A bettor places 25 wagers at a flat $50 stake with results ranging from -$50 to +$95. The calculator returns a mean of +$8.20 and a standard deviation of roughly $42, giving a coefficient of variation over 500%, since the mean sits so close to zero.
Example 2 – Aggressive parlay staking. A parlay bettor logs 15 results swinging between -$100 and +$400 due to occasional big multi-leg hits. The standard deviation comes out far higher than Example 1’s, even though the overall profit might look similar on paper.
Comparing two bettors’ standard deviations side by side, at similar average stakes, tells you which strategy is actually riskier to run – regardless of which one happened to finish ahead this month.
Example 3 – Matched betting log. A matched bettor logging 40 small, mostly consistent qualifying-and-free-bet results sees a very tight standard deviation relative to the mean, reflecting the low-variance nature of that strategy by design.
Example 4 – Small sample trap. A bettor with only 5 results calculates a standard deviation that looks reassuringly low. With fewer than 10-15 results, that number is far too unstable to trust for real bankroll planning.
Example 5 – Unit-based staking log. A bettor tracking results in units rather than currency enters results like 1.5, -1, -1, 2.2, -1, 0.8. Switching Value Type to Units keeps the output in the same scale as their staking plan, avoiding a currency conversion error.
💡 Tips & Best Practices
Log every single result, including small losses and pushes, rather than cherry-picking only the bets you remember. A partial log skews both the mean and the standard deviation in ways that are hard to predict.
Track results in the same unit you actually stake in. Mixing currency amounts from different stake sizes into one data set inflates variance for reasons that have nothing to do with your actual edge.
Recalculate your standard deviation periodically as your log grows, rather than treating an early estimate as fixed. Early-sample estimates are noisy and tend to settle as more data accumulates.
Use the coefficient of variation, not raw standard deviation alone, when comparing volatility across different bankroll sizes or stake levels. A $50 standard deviation means something very different at a $10 average stake than at a $500 average stake.
Pair this calculator’s output with a drawdown calculation. Standard deviation tells you typical swing size; drawdown analysis tells you how bad a realistic bad stretch could actually look.
Reviewing standard deviation alongside your win rate and average odds gives a far more complete risk picture than any single number in isolation.
Consider segmenting your log by bet type (straight bets vs parlays vs system bets) and running each segment through this calculator separately, since blending very different risk profiles into one number can hide what’s really driving your volatility.
- Keep a running spreadsheet with a timestamp per bet, not just the final number
- Note stake size alongside each result so you can normalize later if it changes
⚠️ Common Mistakes to Avoid
Using Population Instead of Sample Data
Choosing population standard deviation for an ongoing, partial log is one of the most frequent errors bettors make with this tool.
Population mode will always report a smaller number than sample mode on the same data – if a user picks it just because the result “looks better,” they’re quietly understating their own real variance.
Unless your data set is genuinely complete and permanently closed, sample mode is the statistically honest choice.
Mixing Stake Sizes Without Normalizing
Entering results from a period when you staked $10 per bet alongside results from a later period staking $100 per bet produces a standard deviation that reflects your staking changes more than your actual betting risk.
Convert everything to a consistent unit – either fixed currency at one stake level, or units/points relative to your stake – before running the calculation.
If your staking has genuinely changed over time, it’s often more honest to run separate calculations for each staking era rather than one blended figure.
Trusting a Tiny Sample
A standard deviation calculated from 5-8 bets can look deceptively tidy, purely by chance, even for a genuinely high-variance strategy.
Treating a small-sample standard deviation as a reliable long-run figure is the costliest mistake on this list, since it leads directly to under-sizing a bankroll for the volatility that eventually shows up.
Ignoring the Mean Entirely
Standard deviation only describes spread, not direction. A bettor fixated purely on a low standard deviation number can overlook that their mean result is negative.
Always read standard deviation next to the mean, never as a standalone verdict on whether a strategy is worth continuing.
🎯 When to Use This Calculator
Use this tool any time you want to quantify how bumpy a particular betting strategy actually is, rather than relying on gut feel about “good” or “bad” runs. It’s especially useful before increasing stake size, since higher stakes scale your standard deviation right alongside your potential profit.
A strategy’s average return tells you where you’re likely to end up; its standard deviation tells you how rough the road will be getting there.
It’s also a natural companion tool when comparing two different systems or bet types side by side, since raw profit totals alone can hide very different underlying risk profiles.
🔗 Related Calculators
Drawdown Calculator, Drawdown Recovery Calculator, Sharpe Ratio Calculator, Sortino Ratio Calculator, ROI Calculator, Confidence Interval Calculator, Variance Calculator.
📖 Glossary
Standard Deviation – a measure of how spread out a set of results is around its average value.
Variance – the average of the squared differences between each result and the mean; standard deviation is its square root.
Mean – the arithmetic average of all results in a data set.
Sample – a subset of a bettor’s total results, typically an ongoing or partial log.
Population – a complete, final data set with no further results to be added.
Bessel’s Correction – the n-1 adjustment used in sample variance to correct for bias.
Coefficient of Variation – standard deviation expressed as a percentage of the mean, useful for comparing volatility across different scales.
Volatility – the general tendency of results to swing widely rather than stay close to an expected value.
Bankroll – the total funds a bettor has allocated for wagering activity.
Drawdown – the decline from a bankroll’s peak value to a subsequent low point.
Unit/Point Staking – a staking system where bets are sized relative to a base unit rather than fixed currency amounts.
Flat Staking – a strategy where every bet uses the same fixed stake amount.
❓ Frequently Asked Questions
What does a “good” standard deviation look like for betting results?
There’s no universal good number, since it depends entirely on your average stake and strategy type. What matters is the coefficient of variation relative to strategies of a similar style.
For example, a matched betting log might show a coefficient of variation under 50%, while a parlay-heavy log could easily exceed 300% due to the occasional large multi-leg payout.
Should I use sample or population standard deviation?
Sample standard deviation is correct for almost every real bettor’s situation, since your log is virtually always a partial, ongoing record rather than a permanently closed data set.
Only switch to population mode if you are deliberately analyzing one specific, finished period – such as one completed betting season – with no intention of adding to it later.
How many results do I need before the number becomes meaningful?
Most analysts consider at least 20-30 results a reasonable minimum for a rough estimate, with 50+ giving a noticeably more stable figure.
Below roughly 10 results, the standard deviation can swing dramatically just from adding or removing a single unusual result.
Can I use this for units/points instead of currency?
Yes – switch Value Type to Units and the calculator will treat your entries as abstract staking units rather than currency, which keeps the output consistent with a points-based staking plan.
This is the more common choice for bettors who deliberately track results independent of any specific stake size in dollars or another currency.
Why is my coefficient of variation showing “N/A”?
The coefficient of variation divides by the mean, so if your average result across the sample is exactly zero, that calculation is undefined and the field shows N/A instead.
A mean of exactly zero on a real result log is rare, and usually signals either a very small sample or a coincidental break-even stretch.
Does a high standard deviation mean my strategy is bad?
Not by itself – some legitimately profitable strategies (like parlay or accumulator betting) are inherently high-variance by design, and a high standard deviation there simply reflects that structure.
What matters more is whether that volatility is properly matched to your bankroll size and your own tolerance for extended losing stretches.
⚖️ Legal Disclaimer
This calculator is provided for informational and educational purposes only. It does not constitute financial, investment, or gambling advice, and results should not be interpreted as a guarantee of future performance. Gambling involves risk of financial loss. Please gamble responsibly and within your means, and seek independent professional advice if you have concerns about your betting behavior or finances.
Been tracking my crypto casino bets for about 8 months now and finally ran this calculator on my Provably Fair logs. My standard deviation came out to 2.8x my average bet size, which honestly shocked me. I thought I was playing tight, but the swings are massive. The thing is, no-KYC platforms don’t give you these tools built in, so you have to do the math yourself. Switched all my wallets to hardware storage after seeing how volatile my results actually are. Makes me want to be way more careful about position sizing. Bitcoin volatility plus betting variance is a brutal combo if you’re not disciplined. The calculator here beats anything I found on the sketchy offshore sites.
That’s a genuinely insightful observation about the 2.8x ratio. Most recreational bettors don’t do this cross-check, so you’re already ahead of the curve. Your point about combining crypto volatility with betting variance is mathematically sound, especially if you’re holding positions rather than converting to stablecoin immediately after wins. One thing worth considering: if you’re logging 8 months of data, you’ve got a solid sample size for reliable standard deviation estimates. If you haven’t already, try segmenting your logs by bet type or game category and recalculating separately—sometimes you’ll find one subset has dramatically different volatility than your overall portfolio. That can help you identify which games or bet structures are creating the wildest swings. Regarding hardware storage and position sizing: those are exactly the right moves. Standard deviation becomes actionable once you know the actual number. Some players use it to set daily loss limits as a multiple of their calculated stdev, which creates a statistical guardrail rather than an arbitrary dollar amount.
Thanks for that breakdown. I haven’t segmented by game type yet, that’s a solid idea. Most of my volume is on the Provably Fair dice games, but I do hit some slot-style stuff too. Makes sense that they’d have different variance profiles. I’ll split the data and see what shakes out.