Expected Value and Variance

So far, we have learned how to describe random variables and their probability distributions.

Now we answer two very important questions:

• What value do we expect on average?
• How much do values spread around that average?

These questions are answered using Expected Value and Variance.

---

Why Expected Value and Variance Are Important

In real life, we rarely care about one outcome.

We care about:

• Long-term average results
• Risk and uncertainty
• Stability vs volatility

Expected value and variance quantify these ideas.

---

Expected Value (Intuition)

The expected value is the long-run average outcome of a random experiment.

It does NOT mean the value must actually occur.

Think of it as the “center of gravity” of a probability distribution.

---

Expected Value of a Discrete Random Variable

For a discrete random variable X, the expected value is defined as:

E(X) = Σ [ x · P(X = x) ]

This means:

• Multiply each value by its probability
• Add all results

This formula is extremely important for exams.

---

Example: Expected Value (Coin Toss)

Let X = number of heads in one fair coin toss.

X	P(X)
0	1/2
1	1/2

Expected value:

E(X) = (0 × 1/2) + (1 × 1/2) = 0.5

On average, half the tosses are heads.

---

Expected Value in Daily Life

Examples:

• Average marks in exams
• Average waiting time
• Average profit or loss

Expected value helps in planning and forecasting.

---

Expected Value in Business & Economics

Businesses use expected value to:

• Evaluate investments
• Assess expected profit
• Compare risk vs reward

Decisions are made using averages.

---

Expected Value in Data Science & ML

In data science:

• Mean of data = expected value
• Loss functions compute expected loss

Training models minimizes expected error.

---

Variance (Intuition)

Two distributions can have the same average but behave very differently.

Variance measures how spread out the values are around the mean.

Higher variance → more uncertainty.

---

Why We Don’t Use Simple Deviation

If we add deviations from the mean, positive and negative values cancel out.

To avoid this, we:

• Square deviations

This leads to variance.

---

Variance of a Discrete Random Variable

Variance is defined as:

Var(X) = Σ [ (x − μ)² · P(X = x) ]

where μ = E(X).

This formula must be memorized conceptually.

---

Alternative Variance Formula (Very Important)

A commonly used shortcut:

Var(X) = E(X²) − [E(X)]²

This form is faster in calculations.

---

Example: Variance (Simple Case)

Let X = number of heads in one fair coin toss.

We know:

• E(X) = 0.5
• E(X²) = (0² × 1/2) + (1² × 1/2) = 0.5

Variance:

Var(X) = 0.5 − (0.5)² = 0.25

---

Standard Deviation

Variance is measured in squared units.

To bring it back to original units, we take the square root.

Standard Deviation = √Variance

This is easier to interpret.

---

Visual Understanding (Important)

Imagine two distributions:

• Same mean
• Different spreads

The wider one has higher variance.

Graphs make this very clear.

---

Expected Value vs Variance

Measure	What it tells
Expected Value	Average outcome
Variance	Spread / risk

---

Applications in Competitive Exams

Exams test:

• Expected value formula
• Variance shortcuts
• Interpretation of spread

Conceptual clarity saves time.

---

Applications in Machine Learning

Machine learning uses:

• Expected loss minimization
• Variance to measure model stability
• Bias-variance tradeoff

This lesson is foundational for ML theory.

---

Common Mistakes to Avoid

• Thinking expected value must occur
• Forgetting to square deviations
• Confusing variance with standard deviation

Always interpret results carefully.

---

Practice Questions

---

Quick Quiz

---

Quick Recap

• Expected value = average behavior
• Variance = spread or risk
• Standard deviation = √variance
• Used in exams, business, and ML

With expected value and variance mastered, you are now ready to study Normal Distribution, the most important distribution in statistics.