What Is Standard Deviation?

Standard deviation is a statistical measure that quantifies how much individual data points deviate from the average (mean) value. In sports betting, it's one of the most powerful tools for understanding risk, measuring performance consistency, and managing your bankroll effectively.

Think of standard deviation as a measure of "spread." If you're analyzing a team's point totals over several games, a low standard deviation means the team scores consistently around the same number of points. A high standard deviation means the team's scoring bounces around wildly—sometimes scoring 70 points, sometimes 110 points. Both situations might have the same average score, but the risk profile is completely different.

The mathematical symbol for standard deviation is σ (sigma) for a population, or s for a sample. It's calculated as the square root of variance, which is why you'll often see these two concepts discussed together.

Why Standard Deviation Matters in Betting

Most casual bettors focus exclusively on expected value (EV)—the average profit or loss they expect from a bet over time. But expected value alone tells only half the story. Standard deviation reveals the other half: how much your actual results will fluctuate around that expected value.

Consider two betting strategies, both with +5% expected value:

• Strategy A: Win or lose roughly £50 on each bet. Steady, predictable results.
• Strategy B: Win or lose anywhere from £500 to £2,000 per bet. Massive swings.

Both have the same EV, but Strategy B has much higher standard deviation. If you don't understand and prepare for this volatility, you might abandon a profitable strategy during a losing streak, or blow your entire bankroll before the law of large numbers kicks in.

Standard deviation helps you:

• Assess true risk beyond just winning percentage
• Size your bets appropriately using the Kelly Criterion or similar methods
• Predict realistic downswings and prepare mentally and financially
• Evaluate consistency of teams, players, and strategies
• Manage bankroll survival during inevitable variance downturns

How Is Standard Deviation Calculated?

The Mathematical Formula

Standard deviation comes in two forms: population standard deviation and sample standard deviation. The difference matters more than most bettors realize.

Why the difference? Sample standard deviation uses (n - 1) instead of n in the denominator—a correction called Bessel's correction. This accounts for the fact that sample data tends to underestimate variability in the true population. For bettors, this is crucial: your past 100 bets are a sample of all possible future bets, so you should use sample SD to avoid underestimating future volatility.

Step-by-Step Calculation Example

Let's walk through a real betting example. Imagine you're analyzing a basketball team's points per game over 10 recent matches:

Scores: 90, 85, 100, 95, 88, 92, 97, 91, 89, 93

Step 1: Calculate the Mean Mean = (90 + 85 + 100 + 95 + 88 + 92 + 97 + 91 + 89 + 93) ÷ 10 = 920 ÷ 10 = 92

Step 2: Find Deviations from the Mean Subtract the mean from each value:

• 90 - 92 = -2
• 85 - 92 = -7
• 100 - 92 = +8
• 95 - 92 = +3
• 88 - 92 = -4
• 92 - 92 = 0
• 97 - 92 = +5
• 91 - 92 = -1
• 89 - 92 = -3
• 93 - 92 = +1

Step 3: Square Each Deviation (We square to eliminate negative signs and emphasize larger deviations)

• (-2)² = 4
• (-7)² = 49
• (+8)² = 64
• (+3)² = 9
• (-4)² = 16
• (0)² = 0
• (+5)² = 25
• (-1)² = 1
• (-3)² = 9
• (+1)² = 1

Step 4: Calculate the Average of Squared Deviations (Variance) Sum = 4 + 49 + 64 + 9 + 16 + 0 + 25 + 1 + 9 + 1 = 178

Since this is sample data (past games representing future potential), use (n - 1): Variance = 178 ÷ (10 - 1) = 178 ÷ 9 = 19.78

Step 5: Take the Square Root Standard Deviation = √19.78 = 4.45 points

What This Means: This team's scores typically vary by about 4.45 points from their average of 92 points. So in a typical game, you'd expect them to score somewhere between 87.55 and 96.45 points (92 ± 4.45), though of course actual results can deviate further.

How Does Standard Deviation Relate to Expected Value and Variance?

The Variance Connection

Variance and standard deviation are mathematically linked: standard deviation is simply the square root of variance. If variance is 19.78, then standard deviation is √19.78 = 4.45.

But why use standard deviation instead of variance? The answer is practical: units matter.

If you're measuring points scored, variance is measured in "squared points"—a meaningless unit. Standard deviation is measured in "points"—the same unit as your original data. This makes standard deviation much more intuitive and useful for decision-making.

The coefficient of variation (CV) is particularly useful for bettors because it lets you compare volatility across different sports or metrics. A team with a mean score of 100 and SD of 10 has CV = 10%. A team with mean 80 and SD of 10 has CV = 12.5%. Despite the same absolute SD, the second team is more volatile relative to its scoring level.

Expected Value and Risk Together

Here's where the real power emerges. Expected value tells you the average outcome. Standard deviation tells you the range of likely outcomes.

Example 1: High EV, Low SD (Ideal)

• Bet type: Favorite moneyline at -200 odds
• Expected Value: +2% per bet
• Standard Deviation: Small (favorites are predictable)
• Outcome: Consistent, steady profits with low volatility

Example 2: High EV, High SD (Risky but Potentially Profitable)

• Bet type: Underdog parlay with +EV
• Expected Value: +5% per bet
• Standard Deviation: Large (underdogs are unpredictable)
• Outcome: Larger swings, but long-term profits if you survive the variance

Example 3: Low EV, Low SD (Mediocre)

• Bet type: Slightly favoured team at -110
• Expected Value: -0.5% per bet
• Standard Deviation: Small
• Outcome: Steady losses over time

Example 4: Low EV, High SD (Disaster)

• Bet type: Random high-odds bets
• Expected Value: -5% per bet
• Standard Deviation: Very large
• Outcome: Unpredictable massive losses

The professional bettor's goal is to find bets with positive EV and manage the SD through proper bet sizing. The Kelly Criterion, a famous formula in betting, does exactly this by balancing EV against SD to determine optimal bet size.

How Do Bettors Use Standard Deviation?

Risk Measurement and Bankroll Management

The most practical application of standard deviation is determining how much to bet. Your bankroll needs to be large enough to survive the inevitable downswings caused by high standard deviation.

The Kelly Criterion is the gold standard for bet sizing. In simplified form:

Bet Size % = (EV × Odds - 1) ÷ Odds

But more importantly, the Kelly Criterion implicitly accounts for standard deviation. A bet with high standard deviation will produce a smaller Kelly percentage than a bet with the same EV but lower SD. This is the formula's way of saying: "High volatility requires smaller bet sizes."

Practical Example:

• Bankroll: £10,000
• Bet with +5% EV and low SD: Kelly suggests 2% of bankroll = £200 per bet
• Bet with +5% EV and high SD: Kelly suggests 0.5% of bankroll = £50 per bet

Why? Because the high-SD bet will produce larger losing streaks. You need more cushion to survive them.

Many professional bettors use fractional Kelly (e.g., 25% of full Kelly or 50% of full Kelly) as additional safety. This reduces volatility even further at the cost of slower long-term growth.

Assessing Consistency and Performance

Standard deviation helps you evaluate whether a team, player, or strategy is actually reliable.

Low SD = Consistent Performance

• A team with low SD in their point totals is predictable
• Good for betting point spreads or totals
• Easier to build a profitable model around them

High SD = Volatile Performance

• A team with high SD is unpredictable
• Might be good for underdog bets (bigger payouts compensate for volatility)
• Harder to model, higher risk

For example, if you're comparing two teams:

• Team A: Average 95 points per game, SD = 3 points
• Team B: Average 95 points per game, SD = 12 points

Both score 95 on average, but Team A is far more consistent. If you're betting on totals, Team A is the safer bet. If you're looking for value in underdog parlays, Team B might offer better odds because the market underestimates their upside potential.

Predicting Outcomes and Managing Downswings

One of the most important uses of standard deviation is understanding the normal distribution and what it tells us about losing streaks.

In a normal distribution (which betting results approximately follow with large sample sizes), approximately:

• 68% of outcomes fall within 1 SD of the mean
• 95% of outcomes fall within 2 SDs of the mean
• 99.7% of outcomes fall within 3 SDs of the mean

This is called the 68-95-99.7 rule or empirical rule.

Real Example: Suppose you have a betting strategy with:

• Expected profit per bet: £10
• Standard deviation per bet: £20
• Number of bets: 100

Your expected total profit = £10 × 100 = £1,000

But thanks to standard deviation, you can predict the likely range:

• 1 SD around the mean = £1,000 ± (£20 × √100) = £1,000 ± £200 = £800 to £1,200
• 2 SDs around the mean = £1,000 ± £400 = £600 to £1,400
• 3 SDs around the mean = £1,000 ± £600 = £400 to £1,600

There's a 68% chance you'll finish between £800 and £1,200. There's a 95% chance you'll finish between £600 and £1,400. And there's a roughly 2.5% chance you'll actually lose money despite a positive EV strategy (falling into the lower tail beyond 2 SDs).

This is why understanding standard deviation is psychologically crucial. When you hit a losing streak, you can calculate whether it's normal variance or a sign that something is wrong with your strategy. A 10-bet losing streak might be perfectly normal given your SD, or it might be a red flag—the math tells you which.

What Are Common Misconceptions About Standard Deviation?

Misconception 1: "High Standard Deviation Always Means Bad Risk"

The Truth: Context matters enormously. High standard deviation is only bad if your expected value is low or negative. High standard deviation with strong positive expected value can be extremely profitable.

Consider underdog bets. They have high standard deviation (you'll lose frequently) but can have positive expected value (when the odds overestimate the favourite's probability of winning). A professional bettor might deliberately seek high-SD bets because the payoffs are larger and the EV is there.

The Kelly Criterion handles this by automatically reducing bet size for high-SD bets. So you can have both high EV and high SD—you just bet smaller amounts on each individual bet.

Misconception 2: "Standard Deviation Is the Same as Risk"

The Truth: Standard deviation measures volatility, not specifically downside risk. It treats upside and downside swings equally.

If you have a strategy that swings between +£500 and -£500, the standard deviation is high. But the upside swings are just as large as the downside swings. Some alternative risk measures, like Value at Risk (VaR) or Conditional Value at Risk (CVaR), specifically focus on downside risk.

In betting, the distinction matters less than in investing (where you care more about losses than gains), but it's worth understanding. A strategy with high SD but symmetric upside/downside is different from a strategy with the same SD but skewed heavily toward losses.

Misconception 3: "I Can Ignore Standard Deviation If I Have Positive EV"

The Truth: Ignoring standard deviation is one of the fastest ways to go broke, even with positive EV.

A strategy with +1% EV and extremely high SD might require a bankroll 100 times larger than a strategy with +1% EV and low SD, just to survive the variance. If your bankroll is too small relative to the SD, you'll hit zero before the law of large numbers kicks in and your EV materializes.

This is why bankroll management and understanding SD go hand-in-hand.

Misconception 4: "Standard Deviation from 10 Games Is Reliable"

The Truth: Standard deviation estimates from small sample sizes are very unreliable. You need dozens or hundreds of data points before SD becomes a trustworthy measure.

A team's SD calculated from 10 games might be 5 points. From 50 games, it might be 8 points. From 100 games, it might be 7.5 points. The sample size matters enormously.

For betting purposes, try to use at least 30-50 data points before trusting your SD calculations. Ideally, 100+.

Misconception 5: "Lower Standard Deviation Means Higher Probability of Profit"

The Truth: Lower SD means more consistent results, but it doesn't guarantee profits. A low-SD strategy with negative EV will consistently lose money.

Standard deviation is about predictability, not profitability. You need both low SD (consistency) and positive EV (profitability) for a successful strategy.

Misconception 6: "Standard Deviation Predicts Which Team Will Win"

The Truth: Standard deviation doesn't predict outcomes; it measures the spread of past outcomes. A team with low SD in their scores doesn't necessarily have a higher probability of winning their next game.

Standard deviation is useful for assessing consistency and managing risk, but it's not a predictive model by itself. You still need to assess team quality, matchups, injuries, and other factors to predict outcomes.

Where Did Standard Deviation Come From?

Historical Development

Standard deviation wasn't invented overnight. It evolved gradually as statisticians sought better ways to measure the spread of data.

Karl Pearson formally defined standard deviation in 1893, building on earlier work by mathematicians like Carl Friedrich Gauss. Pearson's contribution was recognizing that the square root of variance (standard deviation) was more useful than variance itself because it had the same units as the original data.

Throughout the 20th century, standard deviation became the dominant measure of spread in statistics. It became standard in:

• Physics and engineering (1920s-1930s) for quality control
• Finance and economics (1950s-1960s) through Harry Markowitz's portfolio theory
• Gambling and betting (1970s onward) through the Kelly Criterion and modern sports analytics

Today, standard deviation is so universal that it's hard to imagine statistics without it. It appears in financial models, medical research, quality assurance, sports analytics, and countless other fields.

The reason for its dominance is simple: it works. Standard deviation captures the essential information about variability in a single number that's easy to interpret and use for decision-making.

How Does Standard Deviation Differ From Related Concepts?

Standard Deviation vs. Variance

As mentioned earlier, variance is the square of standard deviation. Mathematically, they contain the same information. Practically, standard deviation is superior because:

• Units: Variance is in squared units (e.g., points²), which is meaningless. SD is in original units (points).
• Interpretability: A SD of 4.45 points is intuitive. A variance of 19.8 points² is not.
• Comparability: You can directly compare SDs across different metrics. Variance is harder to interpret.

When to use each:

• Use SD for practical betting decisions and risk assessment
• Use variance for mathematical proofs and theoretical analysis

Standard Deviation vs. Mean Absolute Deviation (MAD)

Mean Absolute Deviation is an alternative measure of spread. Instead of squaring deviations, you take their absolute values and average them.

For example, if a team usually scores 90-100 points but once scored 30 points (due to injuries), SD will be heavily influenced by that outlier. MAD will treat it more neutrally.

For betting purposes, standard deviation is preferred because:

• It's the industry standard (easier to compare with others' analyses)
• It aligns with statistical theory (normal distribution, z-scores, etc.)
• It emphasizes larger deviations, which matter more in betting (a 20-point swing is worse than two 10-point swings)

However, if you're dealing with data that has extreme outliers (e.g., a player who was injured for several games), MAD might give you a more realistic picture of normal variability.

Standard Deviation vs. Other Risk Measures

Professional investors and bettors use several risk measures:

For sports bettors, standard deviation remains the most practical because:

• It's easy to calculate from historical data
• It integrates naturally with the Kelly Criterion
• It aligns with betting theory and research

What Should Bettors Know About Standard Deviation?

Practical Tips for Using SD in Your Betting

1. Recalculate Regularly: Don't use SD from an entire season if you're betting on recent games. Teams change—injuries, trades, form fluctuations matter. Recalculate SD using the last 20-30 games.

2. Use Spreadsheet Software: Excel, Google Sheets, or Python can calculate SD automatically. Use the STDEV.S() function (sample SD) in Excel.

3. Combine with Other Metrics: SD is powerful but not sufficient alone. Combine it with:

   • Expected value calculations
   • Injury reports
   • Head-to-head records
   • Recent form trends
   • Situational factors (home/away, rest days, etc.)

4. Understand Sample Size Limitations: SD from 10 games is unreliable. Aim for at least 30 games. 100+ is ideal.

5. Account for Structural Changes: If a team's key player was injured for part of your data, separate the data into "with player" and "without player" and calculate SD for each.

6. Use Software Tools: Betting analytics platforms, Python libraries (NumPy, Pandas), or specialized betting software can calculate SD automatically and save you time.

Common Pitfalls to Avoid

1. Over-Relying on Recent Data: A team's SD from the last 5 games is almost meaningless. You need history to get reliable estimates.

2. Ignoring Context: A high SD might not indicate inconsistency—it might indicate a team that's improving (trending upward) or declining (trending downward). Separate those trends from random variance.

3. Confusing Correlation with Causation: Just because two teams have similar SDs doesn't mean they'll perform similarly against each other.

4. Forgetting About Regression to the Mean: Extreme performances (very high or very low scores) tend to regress toward the average over time. A team's SD might decrease simply because extreme outliers don't repeat.

5. Betting Too Much on High-SD Outcomes: Even if EV is positive, high SD requires smaller bet sizes. Don't ignore Kelly Criterion guidance.

6. Calculation Errors: Always double-check your calculations, especially when using spreadsheets. One wrong entry can throw off your entire SD estimate.

Frequently Asked Questions

What does a standard deviation of 4.22 mean in sports betting?

It means that, on average, individual results deviate by about 4.22 units (points, goals, etc.) from the mean. If a team's average score is 92 points with SD of 4.22, most games will fall between 87.78 and 96.22 points. It's a measure of consistency—lower SD means more predictable, higher SD means more volatile.

How many bets do I need to see the effects of standard deviation?

You need at least 30 bets to get a rough sense of your strategy's SD. 100+ bets gives you a much more reliable estimate. The larger your sample, the more accurately your actual results will approach your expected value. This is the law of large numbers.

Can I use standard deviation to predict which team will win?

No, not directly. Standard deviation measures the spread of past results; it doesn't predict future outcomes. However, you can use it to assess consistency (low SD = predictable performance) and combine it with other analysis to make better predictions.

Is standard deviation the same as volatility?

Yes, essentially. In betting and finance, "volatility" and "standard deviation" are often used interchangeably. Both refer to how much results fluctuate around the average.

How do I calculate standard deviation in Excel?

Use the formula =STDEV.S(range) for sample standard deviation (which is what you want for betting). For example, if your data is in cells A1:A100, type =STDEV.S(A1:A100). Excel will calculate it automatically.

Why do some bettors ignore standard deviation?

Some bettors don't understand statistics well enough to use SD effectively. Others focus exclusively on short-term results and don't think about long-term variance. Professional bettors always account for SD through proper bankroll management and bet sizing.

What's a "good" standard deviation for a betting strategy?

There's no universal "good" SD—it depends on your bankroll and risk tolerance. However, a strategy with high EV and low SD is ideal. If you have high EV and high SD, you need a larger bankroll and smaller bet sizes to survive the variance. The Kelly Criterion helps you balance EV and SD automatically.

Related Terms

• Variance — The square of standard deviation; measures spread of data
• Expected Value — Average profit or loss per bet over time
• Sharpe Ratio — Risk-adjusted return metric (Expected Value ÷ Standard Deviation)
• Kelly Criterion — Formula for optimal bet sizing based on EV and odds
• Monte Carlo Simulation — Statistical method for modeling variance in betting outcomes
• Bankroll Management — Strategy for protecting capital against variance
• Normal Distribution — Bell curve; basis for understanding SD in betting