Mathematics Tutor Singapore | H2 Expectation and Variance

Mathematics Tutor Singapore | JC H2 Maths discrete random variables can look like a collection of probabilities until the student asks a fundamental question: what does a numerical outcome mean, and how does its chance contribute to an overall expected result? A probability table, its expected value and its variance describe different aspects of the same model.

Did you know that the expected value of a random variable can be 1.7 even if the only possible actual outcomes are 0, 1, 2 and 3? Expectation is a probability-weighted mean over repeated instances of the model, not a promise that a single trial can produce the mean as an outcome.

This Singapore H2 Mathematics 9758 guide teaches valid discrete distributions, E(X), E(X²), Var(X), linear transformations, missing probabilities and the binomial connection with original calculations. It follows SEAB’s 2026 syllabus, including the explicit exclusion of finding a discrete cumulative distribution function as a required procedure.

Explore H2 discrete random variables, expectation and variance

A random variable assigns numbers to outcomes
A valid probability distribution must add to one
Expected value is a probability-weighted mean
Worked table: mean 1.7 and variance 0.81
Variance measures spread, not the expected outcome
Compute variance using E(X²)−[E(X)]²
What happens to expectation and variance after scaling
Find a missing probability in a distribution
Negative values are allowed for the random variable
E[g(X)] is not generally g(E[X])
Connect a general table to the binomial model
Independent variables and a useful variance extension
Units, interpretation and a realistic application
Nine discrete random-variable mistakes to diagnose
Sixteen original probability-and-expectation questions
A JC1–JC2 probability revision progression
Frequently asked H2 Maths statistics questions
SEAB 2026 H2 syllabus and related Mathematics resources

A random variable assigns numbers to outcomes

Consider a chance process with three possible numerical results. Let X denote the number of suitable items found in a small sample. Before the sample is observed, X is uncertain; after the outcome is known, X takes one particular numerical value.

A discrete random variable takes values from a finite or countable set. We may describe it by a table of values x and associated probabilities P(X=x). Those probabilities tell us which numerical outcomes are more or less likely.

The random variable is not the same as the event. For example, the event “X is at least two” collects several numerical outcomes. If X can be 0,1,2,3, that event contains outcomes 2 and 3 and its probability is the sum of their masses.

A valid probability distribution must add to one

Every probability must lie between zero and one, and the probabilities over all possible values must total exactly one. If an outcome is omitted or counted twice, the table no longer correctly describes a complete random variable.

Value x	P(X=x)	Interpretation
0	0.1	Ten per cent chance
1	0.3	Thirty per cent chance
2	0.4	Forty per cent chance
3	0.2	Twenty per cent chance
Total	1.0	Complete distribution

This original discrete distribution will be used throughout the next worked example.

Here the probability of X≥2 is P(X=2)+P(X=3)=0.4+0.2=0.6. Likewise P(X<2)=0.1+0.3=0.4. Those complementary events add to one.

A student may incorrectly write P(X≥2)=P(X=2) alone. That leaves out the valid value three. The mathematical inequality in the event, not the most prominent row of the table, determines which probabilities to add.

Expected value is a probability-weighted mean

The expectation E(X) is calculated by multiplying each possible value by its probability and adding the products. Symbolically, E(X)=Σ xP(X=x) over all possible values of X. It is a weighted average of the possible outcomes.

For the table above, the contributions are 0×0.1=0, 1×0.3=0.3, 2×0.4=0.8 and 3×0.2=0.6. Adding gives E(X)=1.7.

The expectation is not necessarily a possible observed value. No single trial produces 1.7 if X only takes 0,1,2 and 3. Over many appropriate repeated trials, an average may be close to the expected value, but individual outcomes can vary.

Value x	Probability	xP(X=x)
0	0.1	0
1	0.3	0.3
2	0.4	0.8
3	0.2	0.6
Sum	1.0	1.7

Expectation is weighted by probabilities, not an unweighted average of listed values.

A common incorrect answer is (0+1+2+3)/4=1.5. That calculation assumes the four values are equally likely, which the stated probabilities contradict.

Worked table: mean 1.7 and variance 0.81

Variance describes how spread out the outcomes are around their expected value. One practical formula for a discrete distribution is Var(X)=E(X²)−[E(X)]². We already know E(X)=1.7.

Compute E(X²) by squaring each outcome first and then weighting it by its probability: 0²×0.1=0, 1²×0.3=0.3, 2²×0.4=1.6 and 3²×0.2=1.8. The sum is 3.7.

Then Var(X)=3.7−(1.7)²=3.7−2.89=0.81. The standard deviation is √0.81=0.9.

x	P(X=x)	xP(X=x)	x²P(X=x)
0	0.1	0	0
1	0.3	0.3	0.3
2	0.4	0.8	1.6
3	0.2	0.6	1.8
Totals	1.0	1.7	3.7

The two weighted columns supply mean and second moment.

This check is useful: variance must not be negative. A negative final value often means the student has mixed up E(X²) with [E(X)]² or made a multiplication error.

Variance measures spread, not the expected outcome

Variance can also be written as E[(X−μ)²], where μ=E(X). This formula averages the squared distances between outcomes and the expected value. Squaring ensures outcomes both above and below the mean contribute nonnegative amounts rather than cancelling.

For the worked table, μ=1.7. The distances from this mean are −1.7, −0.7, 0.3 and 1.3 for x=0,1,2,3 respectively. Squaring and weighting gives the same variance 0.81, though the second-moment formula is often quicker.

Outcome x	x−1.7	Squared difference	Weight
0	−1.7	2.89	0.1
1	−0.7	0.49	0.3
2	0.3	0.09	0.4
3	1.3	1.69	0.2

Weighted squared differences give 0.289+0.147+0.036+0.338=0.81.

Standard deviation is measured in the same units as X, while variance has squared units. If X counts objects, variance is expressed in count-squared units as a mathematical measure of spread, not as a number of extra objects actually observed.

Use E(X²)−[E(X)]² correctly

These two expressions are not identical: E(X²) means square each possible outcome first, then average; [E(X)]² means first average, then square that single number. For the example, E(X²)=3.7 while [E(X)]²=2.89.

Their difference is the variance. If a student sets E(X²) equal to [E(X)]² automatically, the computed variance would always become zero, incorrectly asserting that every random variable is constant.

The expectation and variance rules also connect to the identity E[(X−μ)²]=E(X²)−2μE(X)+μ²=E(X²)−μ². This derivation is short and helps a mathematically ready H2 learner understand why the formula works.

Scaling and shifting a random variable

Suppose a transformed random variable is Y=2X+5. Each value of X is doubled and then increased by five. Expectation behaves linearly: E(Y)=2E(X)+5=2×1.7+5=8.4.

Variance measures spread, so adding five to every outcome moves the distribution without changing its spread. Doubling every deviation doubles the standard deviation and multiplies the variance by four. Thus Var(Y)=2²Var(X)=4×0.81=3.24.

Quantity	Transformation rule	Worked result
E(2X+5)	2E(X)+5	8.4
Var(2X+5)	4Var(X)	3.24
Standard deviation of 2X+5	2SD(X)	1.8
Var(X+5)	Var(X)	0.81

A fixed shift changes mean but does not change variance.

A frequent mistake writes Var(2X+5)=2Var(X)+5, as though variance follows the expectation rule. It does not. A variance depends on squared deviations, and the constant shift cancels out when deviations from the shifted mean are calculated.

Find a missing probability in a distribution

Suppose X takes values 0,1,2 with probabilities k,2k,3k respectively. The probabilities must add to one, giving 6k=1, hence k=1/6.

Now E(X)=0×k+1×2k+2×3k=8k=4/3. The second moment is E(X²)=0+1×2k+4×3k=14k=7/3.

Variance is 7/3−(4/3)² = 7/3−16/9 =21/9−16/9=5/9. All three probabilities are nonnegative and sum to one, so this is a valid distribution.

x	Probability	With k=1/6
0	k	1/6
1	2k	1/3
2	3k	1/2
Sum	6k	1

Solve the total-probability condition before calculating expectations.

A random variable may take negative values

A discrete random variable is not always a count. It can represent a change in a balance, temperature or score, and therefore may take negative values. Suppose X equals −1 with probability 0.2, zero with probability 0.5 and two with probability 0.3.

Its expected value is (−1)(0.2)+0(0.5)+2(0.3)=−0.2+0.6=0.4. The second moment is 1×0.2+4×0.3=1.4, so variance is 1.4−0.4²=1.24.

The negative possible outcome does not make variance negative. Variance is still an average squared distance from the mean.

E[g(X)] is not generally g(E[X])

For a function g of the random variable, E[g(X)] is calculated by weighting each transformed outcome g(x). It is generally not equal to applying g to E(X).

For g(x)=x² in our main example, E[g(X)]=E(X²)=3.7. But g(E[X])=(1.7)²=2.89. These differ exactly by Var(X)=0.81. The rule E(aX+b)=aE(X)+b is a special linearity property, not permission to move every function outside an expectation.

This distinction is useful in later statistics and modelling. Squaring, exponentiating or applying another nonlinear operation can change how outcomes contribute to a mean.

Connect a general table to a binomial distribution

The binomial distribution is a particular discrete model. If X counts successes in n independent trials, each with the same success probability p and only two outcome types per trial, X can be modelled as B(n,p), provided the conditions fit.

For X~B(5,0.2), the expected number of successes is E(X)=np=5×0.2=1. Its variance is np(1−p)=5×0.2×0.8=0.8.

The exact probability of two successes is P(X=2)=C(5,2)(0.2)²(0.8)³=10×0.04×0.512=0.2048. The expectation one does not mean exactly one success must occur in every block of five trials.

Binomial question	Working	Answer
Mean	np=5×0.2	1
Variance	np(1−p)	0.8
P(X=2)	C(5,2)(0.2)²(0.8)³	0.2048
E(X²)	Var(X)+[E(X)]²	1.8

A binomial distribution is one special model within discrete random variables.

This method should not be applied when the number of trials is random or the success chance changes from trial to trial. For example, drawing balls without replacement from a small bag can violate the independence or constant-p conditions. The model must be justified before substituting n and p.

Independent variables: a useful variance extension

Expectation is linear even when two random variables are not independent: E(X+Y)=E(X)+E(Y), provided the expectations exist. For variance, independence matters. If X and Y are independent with finite variances, then Var(X+Y)=Var(X)+Var(Y).

In general, dependent variables have a covariance term. This is why adding two variances without checking independence can give a wrong result. In this guide the rule is a contextual extension; the current H2 syllabus’s specific examination requirements should govern how deeply combinations of random variables are assessed.

A simple special case avoids ambiguity: Y=2X+5 is certainly dependent on X, but its variance follows the affine transformation rule, Var(Y)=4Var(X), not Var(X)+Var(5) or an independence shortcut.

Units, interpretation and a realistic application

Suppose X is the number of tasks completed during a short time window under a specified random model. An expectation of 1.7 tasks means the probability-weighted average per such window is 1.7. It does not require any individual window to contain a fractional task.

If someone compares two processes only by their expected task counts, they may miss important differences in variability. Two processes could have the same mean but one may be much more unpredictable. Variance and standard deviation help describe that dispersion.

A probability model is conditional on its assumptions and data. Real workloads may exhibit dependence or changing probabilities across time. Using the mathematical table responsibly includes acknowledging those limitations instead of treating a theoretical expectation as a promise.

Nine discrete random-variable mistakes

First wrong move	Likely concept gap	Useful next question
Probabilities do not total one	Incomplete distribution	Are all outcomes included?
Uses simple average of values	Probability weights ignored	Are values equally likely?
Says 1.7 must occur in one trial	Expectation confused with outcome	Is 1.7 in the support?
Calculates E(X²) as [E(X)]²	Nonlinearity confused	Square before weighting
Variance is negative	Second moment arithmetic wrong	Check weighted square column
Writes Var(2X+5)=2Var(X)+5	Scaling variance incorrectly	How do deviations change?
Applies binomial without constant p	Model conditions overlooked	Are trials independent with fixed chance?
Adds variances for dependent variables	Independence assumed	What links the outcomes?
Reports sample average as exact E(X)	Empirical vs theoretical mean confused	Where did probabilities come from?

The tutor’s next lesson should fix the first invalid relationship.

A student might calculate E(X) correctly but be unable to interpret it. Another can explain the expectation but makes an arithmetic error squaring a negative possible value. Good tutoring separates statistical meaning and calculation technique.

Sixteen original discrete probability questions

1. Check a distribution

Probabilities 0.1,0.3,0.4,0.2 sum to one.

Check every probability is feasible and the distribution totals one. Identify whether the requested output is a probability, mean, variance, standard deviation or transformed expectation before substituting formulas.

2. Event from the table

For X=0,1,2,3 with those probabilities, P(X≥2)=0.4+0.2=0.6.

3. Another event

P(X<2)=0.1+0.3=0.4.

4. Main expected value

E(X)=0(0.1)+1(0.3)+2(0.4)+3(0.2)=1.7.

5. Second moment

E(X²)=0+0.3+1.6+1.8=3.7.

6. Main variance

Var(X)=3.7−1.7²=0.81.

7. Standard deviation

SD(X)=√0.81=0.9.

8. Linear expected value

E(2X+5)=2(1.7)+5=8.4.

9. Linear transformed variance

Var(2X+5)=4(0.81)=3.24.

10. Constant shift variance

Var(X+5)=Var(X)=0.81.

11. Missing k

Probabilities k,2k,3k yield 6k=1 and k=1/6.

12. Mean from k table

For values 0,1,2 and probabilities k,2k,3k, E(X)=4/3.

13. Variance from k table

The same distribution gives Var(X)=5/9.

14. Signed random variable

For values −1,0,2 and probabilities 0.2,0.5,0.3, E(X)=0.4 and Var(X)=1.24.

15. Binomial moment

If X~B(5,0.2), E(X)=1 and Var(X)=0.8.

16. Binomial exact event

For X~B(5,0.2), P(X=2)=0.2048.

A JC1–JC2 probability revision progression

Session	Concept	Independent evidence
1	Sample spaces and variable values	Distinguishes X from event
2	Valid distribution tables	Probabilities sum to one
3	Expectation as weighted mean	Computes E(X)
4	E(X²) and variance	Does not confuse square orders
5	Standard deviation	Uses correct units
6	Affine transformations	Distinguishes E and Var rules
7	Unknown probabilities k	Uses total-probability constraint
8	Binomial model	Checks trial conditions
9	Mixed questions without headings	Selects method independently
10	Delayed test and mistake review	Explains without tutor prompts

Repair the earliest recurring error before repeating a full Paper 2.

For students who can press calculator keys but cannot explain why expected value is weighted, return to a small four-row table and ask what each possible result contributes. For students who understand the concept but stumble with algebra, practise signed products and squares.

Frequently asked H2 Maths statistics questions

What is a discrete random variable?

A numerical quantity determined by a random experiment that takes values from a finite or countable set.

How do I know whether a probability table is valid?

Each probability must be between zero and one and all probabilities over possible outcomes must sum to one.

What does expected value mean?

It is the probability-weighted mean over the model, not necessarily an outcome possible in a single trial.

Why might E(X) be 1.7 when X is always an integer?

A mean of repeated integer outcomes need not be an integer.

What is the difference between E(X²) and [E(X)]²?

One averages squared outcomes; the other squares the weighted average. Their difference is variance.

Why does adding five not change variance?

Every outcome and the mean shift equally, so deviations from the mean remain the same.

Why does multiplying X by two multiply variance by four?

Every deviation doubles, and variance averages squared deviations.

When can I use the binomial formula np and np(1−p)?

When the random variable truly counts successes in a fixed number of independent trials with a common success probability.

Is finding the cumulative distribution function required in H2 9758?

The 2026 syllabus expressly excludes finding the CDF of a discrete random variable.

Can expectation guarantee what will happen next?

No. It describes a probability-weighted mean, not a guaranteed immediate outcome.

2026 SEAB syllabus and related Mathematics routes

The official SEAB 2026 A-Level syllabus index links the H2 Mathematics 9758 specification. Section 6.2 covers discrete random variables, probability distributions, expectations, variances and the binomial distribution, and excludes finding a discrete random variable’s cumulative distribution function.

Continue through the Mathematics Learning Hub. For an earlier probability prerequisite, read Secondary 4 Probability and Tree Diagrams. For the later choice between distributions, see JC H2 Binomial versus Normal Distributions. To connect sampling evidence with statistical inference, use JC H2 Hypothesis Testing and p-Values.

The unchanged Secondary 1 Mathematics Tutor Clementi small-group reference remains the original teaching benchmark. For a parent–student consultation, contact eduKate Singapore with one distribution table and the pupil’s first attempt at its variance calculation.

Mathematics Tutor Singapore | JC H2 Discrete Random Variables, Expectation and Variance