Skip to sample
Synlecto / Public samples

Simple Properties of Probabilities

John Tsitsiklis · MIT OpenCourseWare

About this guide and its source

Derive useful probability rules and recognise when addition is valid.

These curated examples are not untouched extension output. Reviewed against the public lecture; use the source as the authority.

6 concepts7 flashcards5 quiz questions1 mission

Before you start: Know sample spaces, events as sets, union, intersection and complement. The lecture starts from non-negativity, P(Ω) = 1 and addition for two disjoint events.

Covers the complete short lecture: complements, the upper bound of 1, the empty set, finite disjoint unions and finite sets of outcomes. It does not cover conditional probability, independence or infinite unions. New practice scenarios are Synlecto exercises.

Links return to source sections, not exact sentence timestamps. Editorial edition: 2026-11-09.

Original MIT lecture (new tab) · Official transcript (new tab)

Try this curated guide without signing in. Practice resets when you leave or reload. Nothing is saved to an account.

Chapter 01

Starting points

Understand the set relationship before using the formula.

Start with complements

Read Aᶜ as “not A”, A ∪ B as “A or B, including either”, and A ∩ B as “both A and B”. For complements, no outcome is in both sets and every outcome is in one of them. Disjointness licenses addition; covering Ω supplies the total of 1. Keep these two reasons separate.

Derive two consequences

The proof has a route: first derive P(Aᶜ) = 1 − P(A); then combine that identity with non-negativity to obtain P(A) ≤ 1; finally apply it to Ω to obtain P(∅) = 0. If you cannot explain a later result, return to the earlier identity it uses.

Check disjointness before adding

Before adding event probabilities, check whether a single outcome could belong to more than one event. For three or more events, every pair must be disjoint. Having no outcome shared by all of them is a weaker condition and does not prevent double counting.

Chapter 02

Concept map

Lecture concept map

Explore the ideas, then return to the lecture to check the detail and context.

Use the up and down arrow keys to move between concepts. Press Home or End to jump to the first or last concept.

Concept 1 of 6

Focused concept: Axioms and derived rules. Concept 1 of 6.

Focused concept

Axioms and derived rules

Concept 1 of 6

Definition

The lecture begins with three available rules: probabilities are non-negative; the whole sample space has probability P(Ω) = 1; and probabilities add for two disjoint events. Its purpose is to derive further rules from these assumptions. In each proof, distinguish a fact about sets from a probability axiom and from an algebraic rearrangement.

Example

The upper bound P(A) ≤ 1 is proved rather than added as a separate axiom.

Common confusion

Every intuitive probability rule needs a separate axiom.

Check the distinction

Show how a proposed rule follows from the existing axioms before adding assumptions.

Chapter 03

Compare

Similar-looking expressions can require different conditions.

Disjoint versus complementary

Disjoint events cannot occur together. Complementary events also cover the whole sample space. Synlecto example: in Ω = {a, b, c}, the events {a} and {b} are disjoint, but {b} is not the complement of {a}; that complement is {b, c}. Identify the missing set before choosing a formula.

One outcome versus its singleton event

An outcome s is one element; {s} is the event containing that element. Probabilities formally apply to events, so P({s}) is the precise notation. The lecturer then permits P(s) as shorthand for it. This is a notation change, not a different probability rule.

Axiom versus result

“P(A) ≥ 0” is available as an axiom. “P(A) ≤ 1” must be derived using the complement relation and non-negativity of P(Aᶜ). Naming the desired result is not a reason for it. Try annotating each proof line with “set fact”, “axiom” or “algebra”.

Chapter 04

Process

Reconstruct the proofs rather than memorising disconnected formulas.

Complement identity

The complement rule is not a new assumption. Reconstruct it from two facts about sets and the probability axioms, keeping the reason beside each step.

  1. Cover the whole space

    Ω = A ∪ Aᶜ. Every outcome either belongs to A or does not, so the two sets together contain all possible outcomes.

  2. Check that addition is allowed

    A ∩ Aᶜ = ∅. No outcome is counted twice, so the additivity axiom gives P(A ∪ Aᶜ) = P(A) + P(Aᶜ).

  3. Use normalisation, then rearrange

    P(Ω) = 1, so P(A) + P(Aᶜ) = 1. Subtract P(A) to obtain P(Aᶜ) = 1 − P(A). Coverage and disjointness have different jobs in this proof.

Try it: which step fails if B is merely disjoint from A?

You can add P(A) and P(B) to obtain P(A ∪ B), but you cannot replace the union with Ω without knowing that the two events cover every outcome. Their sum need not be 1.

Upper bound and empty set

These are two separate consequences of the complement identity. Neither requires outcomes to have equal probabilities, nor does the upper-bound proof require the event A to be finite.

  1. Obtain the upper bound

    P(A) = 1 − P(Aᶜ). Non-negativity says P(Aᶜ) ≥ 0, so subtracting it cannot give more than 1. Together with P(A) ≥ 0, this establishes 0 ≤ P(A) ≤ 1.

  2. Identify the complement of Ω

    All possible outcomes are already in Ω. There are none outside it, so Ωᶜ = ∅.

  3. Apply the identity to that event

    P(Ω) + P(Ωᶜ) = 1 becomes 1 + P(∅) = 1. Subtract 1 to obtain P(∅) = 0.

Try it: why does “P(A) is non-negative” not prove that it is at most 1?

A number such as 1.2 is non-negative but greater than 1. The missing argument is that P(A) = 1 − P(Aᶜ), where the subtracted probability is non-negative. That relation connects the lower-bound axiom to the total of 1.

Three-event union

The lecture extends a two-event axiom by grouping events. Suppose A, B and C are pairwise disjoint. Treat A ∪ B as one event, so the problem has the two pieces required by the axiom. Every justification still depends on the stated disjointness.

  1. Check the grouped pair

    A ∪ B is disjoint from C: an outcome in both would have to lie in A and C, or in B and C, contradicting pairwise disjointness.

  2. Apply the axiom twice

    P((A ∪ B) ∪ C) = P(A ∪ B) + P(C). Because A and B are also disjoint, replace P(A ∪ B) with P(A) + P(B). The result is P(A) + P(B) + P(C).

  3. See how the finite extension works

    For a fourth disjoint event D, group A ∪ B ∪ C as one event and add D. Repeating this grouping gives the rule for any finite number of pairwise disjoint events. This lecture does not justify extending the argument to infinitely many events.

Chapter 05

Applied examples

Keep the conditions alongside the arithmetic.

The lecture’s coin

The lecture gives a two-outcome coin model with P(heads) = 0.6. Tails is the complement of heads, so its probability is 1 − 0.6 = 0.4. The reasoning depends on the two outcomes covering the model, not on an assumption that a coin must be fair.

  1. State the model

    The possible outcomes are heads and tails. In this model, not-heads is exactly tails.

  2. Choose the complement rule

    P(tails) = 1 − P(heads). This is justified because the two events are disjoint and exhaustive.

  3. Calculate and check

    The result is 0.4. It is non-negative, and 0.6 + 0.4 = 1. Neither outcome was assumed equally likely.

A finite event

Write a finite event as a union of distinct singleton events. Because one outcome cannot be two different outcomes, these singleton events are disjoint. Add their assigned probabilities; counting outcomes alone is not enough.

  1. Specify the givens

    Synlecto worked example: Ω = {a, b, c, d}, with singleton probabilities 0.1, 0.2, 0.3 and 0.4. These are supplied exercise values, not numbers from the lecture.

  2. Represent the requested event

    For A = {a, c}, write A = {a} ∪ {c}. The singleton events have no shared outcome, so finite additivity applies.

  3. Add the weights, not the counts

    P(A) = 0.1 + 0.3 = 0.4. The complement is {b, d}, with probability 0.2 + 0.4 = 0.6. The two results add to 1. Using 2/4 would incorrectly assume equal outcome probabilities.

Try it: what is P({b, c}) in this exercise?

0.2 + 0.3 = 0.5. The two singleton events are disjoint. This happens to equal two out of four here, but that coincidence does not make equal-outcome counting a valid general method.

Why an empty three-way intersection is not enough

Synlecto worked example: Ω = {1, 2, 3, 4}, with probability 0.25 assigned to each outcome. Let A = {1, 2}, B = {2, 3} and C = {4}. No outcome belongs to all three events, but the events are not pairwise disjoint.

  1. Find the shared outcome

    Outcome 2 belongs to both A and B. Although A ∩ B ∩ C = ∅, A ∩ B = {2}. The condition for adding all three event probabilities fails.

  2. Expose the double count

    P(A) + P(B) + P(C) = 0.5 + 0.5 + 0.25 = 1.25. Outcome 2 contributed 0.25 twice. A probability above 1 is a warning that the calculation cannot be right.

  3. Calculate with distinct outcomes

    The union is {1, 2, 3, 4} = Ω. Add each distinct singleton once to obtain 0.25 + 0.25 + 0.25 + 0.25 = 1. No overlapping-event formula is needed for this finite example.

Try it: would direct addition become valid if B were changed to {3}?

Yes. A = {1, 2}, B = {3} and C = {4} are now pairwise disjoint. Their probabilities are 0.5, 0.25 and 0.25, whose sum is 1. The justification is the repaired set relationship, not simply that the arithmetic now gives a plausible number.

Chapter 06

Common confusions

Common traps, with the distinction to keep in mind.

Axioms and derived rules

Common confusion: Every intuitive probability rule needs a separate axiom.

Check the distinction: Show how a proposed rule follows from the existing axioms before adding assumptions.

Complementary events

Common confusion: Any two disjoint events have probabilities that add to 1.

Check the distinction: Complements have both properties: no overlap and complete coverage. These justify replacing their union with Ω and adding their probabilities to get 1. Disjointness alone permits addition but does not guarantee a sum of 1; other outcomes may remain.

Probabilities lie between 0 and 1

Common confusion: The upper bound follows just from probabilities being non-negative.

Check the distinction: Non-negativity alone does not impose an upper bound. The derivation also uses the complement identity, which depends on normalisation and additivity.

The empty event

Common confusion: The empty set and the whole sample space have the same probability because both are special sets.

Check the distinction: They are complements with very different roles: P(Ω) = 1 and P(∅) = 0.

Add disjoint events

Common confusion: No outcome common to all three events is enough to justify adding all three probabilities.

Check the distinction: Every pair must be disjoint. Two events can overlap even when the intersection of all three is empty.

Add distinct individual outcomes

Common confusion: A finite event’s probability is always its number of outcomes divided by the total number of outcomes.

Check the distinction: This lecture’s rule adds the assigned probabilities. It does not assume all outcomes have equal probability.

Chapter 07

Decision points

Choose a rule by its conditions.

Do I know the complement?

Use 1 − P(A) when the requested event is exactly Aᶜ, not merely some event disjoint from A.

Can I add these probabilities directly?

Check every pair for overlap. If the events overlap, this lecture’s disjoint-addition rule is not applicable as stated.

Am I counting or weighting outcomes?

For distinct outcomes, add their individual probabilities. Do not silently replace them with equal weights.

Chapter 08

Checklist

Explain the reason for each step out loud or in your own notes.

Prove the complement identity

Name the set relation, disjointness and normalisation used.

Recover both bounds

Identify the lower bound as an axiom and derive the upper bound using the complement.

Justify finite addition

Explain the grouping step for three events and why distinct singleton events can be added.

When you come back

Try these from memory in a later session. These are suggested prompts, not a scheduled reminder.

Prove the complement identity without looking at the formula first.

Check your explanation

Start with a partition of Ω into A and Aᶜ; use disjoint additivity and P(Ω) = 1.

Explain why three events can have no common outcome and still fail pairwise disjointness.

Check your explanation

For example, A = {1, 2}, B = {2, 3}, C = {4}. No outcome belongs to all three, but 2 belongs to both A and B. With probability 0.25 on each of 1 to 4, adding the three event probabilities gives 1.25 because 2 is counted twice; the union has probability 1. Explain both the set error and its numerical consequence.

Try the flashcards
Simple Properties of Probabilities · Curated sample | Synlecto