IB Middle Years Programme · Mathematics MYP Year 3

Probability and Outcomes

Cover illustration for Probability and Outcomes (Mathematics MYP Year 3).
MYP · Mathematics MYP Year 3

Probability and Outcomes

24 min readStandardCore strand of the Statistics & Probability unit — assessed through unit tests and MYP Criterion A/D investigation tasks; recurs in every subsequent probability unit as the foundation for compound and conditional work.

Probability is careful counting dressed up as prediction. Every question in this topic reduces to comparing two numbers: the theoretical fraction you calculate before you roll, spin or draw, and the experimental fraction you calculate after. Get comfortable listing outcomes without missing any — the rest (trees, grids, Venn diagrams) is just bookkeeping around that one skill.

Overview

This chapter has exactly two families of question: those asking you to predict a probability from counting outcomes (theoretical), and those asking you to measure one from recorded data (experimental). Everything else — dice, coins, spinners, trees, Venn diagrams, sample space grids — is just the setting the counting happens in.

  • The probability scale (0 to 1) gives you the vocabulary: impossible, unlikely, even chance, likely, certain.
  • Single events (one die roll) become combined events (two dice, spinner + die) once you're tracking more than one action at once.
  • Listing outcomes systematically — as a set, a grid, or a tree — is the skill every later calculation depends on.

The shape of the chapter

Command terms this topic actually uses

Command termWhat it demandsAOMark-earning move
CalculateObtain a numeric answer, showing the working/formula used.Knowing and applyingMark is for the correct method AND the final value — a bare answer with no formula shown can lose the method mark.
ExplainGive a reason connected to the maths, not just restate the situation.CommunicatingVague answers ('it's rare', 'it's not fair') earn nothing without a specific mathematical or physical reason.
ConstructBuild the actual diagram or table completely and correctly.Applying mathsMarked on structure/completeness — one missing cell or branch loses marks even if later arithmetic is fine.
ComparePut two values in the same format and state which is bigger and by roughly how much.CommunicatingComparing raw counts instead of probabilities (e.g. '6 blue > 3 blue') scores zero — must normalise first.
JustifyGive a clear verdict (correct/incorrect, true/false) backed by a calculated value.ReasoningA calculation with no stated verdict, or a verdict with no calculation, only earns partial credit.
ApplyUse a given formula on new numbers.Knowing and applyingMarks for correct substitution into the stated formula — using a different (unstated) method can cost marks even if numerically correct.
StateGive a fact directly, no working required.KnowledgeNo explanation expected or rewarded — but the value must still be exact (e.g. 1/6, not 0.17).

Key point

Every probability answer is a number between 0 and 1 (or 0% and 100%) — never negative, never above 1. If your answer breaks that rule, you've miscounted the sample space or mixed up favourable/total.

Theoretical vs Experimental Probability

  • Theoretical probability is calculated before anything happens, purely by counting the sample space — no die needs rolling.
  • Experimental (empirical) probability is measured after actually running trials: frequency of the event divided by total trials performed.
  • The Law of Large Numbers says experimental probability tends towards theoretical probability as the number of trials increases — but it never has to match exactly, and small samples can look wildly 'unfair' even from a perfectly fair object.
  • A single unusual run (three 6s in a row, or 8 heads out of 10 tosses) is evidence, not proof — you need many trials before concluding anything about fairness or bias.
Theoretical probability
P(event) = number of favourable outcomes ÷ total number of possible outcomes, assuming all outcomes are equally likely.
Experimental probability
P(event) = number of times the event occurred ÷ total number of trials performed.
Bias
A systematic tendency for an outcome to occur more or less often than the theoretical value predicts, usually caused by a physical flaw in the object generating the outcomes.

Theoretical probability

— number of favourable outcomes
— total number of possible outcomes
Memorise

Experimental probability

Predicted count of an event, not what you'll necessarily observe

Worked example

Q1 — is the die actually unfair?

[2]
A player rolls a fair six-sided die three times and gets a 6 every time, claiming the die must be unfair. (a) Explain how theoretical probability shows rolling three 6s in a row is still possible with a fair die. [1] (b) Explain one limitation of using only three rolls to decide whether a die is fair. [1]

Worked example

Q5 — does the coin get 'more fair' with more tosses?

[2]
A biased coin is tossed 10, 20, 50, and 100 times, with 4, 9, 24 and 48 heads recorded respectively. (a) Calculate the experimental probability of heads for each number of tosses, as a decimal. [2] (b) Compare the experimental probabilities to the theoretical probability of 0.5. Explain what happens to the difference as the number of tosses increases. [part]

Common mistake

Treating one small experimental result as the 'true' probability and ignoring the theoretical calculation entirely — e.g. stating P(rolling a 3) = 7/50 from a bar chart instead of the theoretical 1/6.

Consequence: Loses the mark on 'state the theoretical probability' questions (like Q8a), because examiners want the exact fraction 1/6 stated, not the observed frequency.

Fix: Underline the keyword before calculating: 'theoretical' → use the formula on the sample space; 'experimental'/'from the data'/'recorded' → use the table.

Examiner tip

When a question gives a frequency table and asks for 'expected frequency', that means theoretical probability × number of trials — never a value read straight off the chart. Q8(b): expected frequency of a 3 in 50 rolls = (1/6)×50 ≈ 8.3, and it will NOT match the observed frequency of 7 exactly — that mismatch is exactly what Q8(c) asks you to explain.

Frequently confused

Concept AConcept BThe distinction that matters
Theoretical probabilityExperimental probabilityTheoretical is reasoned out before any trial from equally likely outcomes; experimental is measured from real data after trials — they only have to be equal in the limit of infinitely many trials, never in a small sample.

Key takeaways

  • Theoretical: P(E) = n(E)/n(S), calculated with no experiment needed.
  • Experimental: P(E) = frequency/trials, calculated only from real recorded data.
  • More trials → experimental probability drifts closer to theoretical (Law of Large Numbers), but a small sample proves nothing about fairness.
  • Expected frequency = theoretical probability × number of trials — a prediction, not a guarantee.

Single and Combined Events

  • A single event involves one action with one outcome recorded — rolling one die, tossing one coin, drawing one marble.
  • A combined event involves two or more actions (or one action checked against two conditions) considered together — rolling two dice and adding the sum, or a spinner AND a die outcome checked at once.
  • For combined events built from independent actions, multiply the individual probabilities: P(A and B) = P(A) × P(B) — this is the multiplication rule.
  • When 'either/or' connects two outcomes that can't both happen at once, add the probabilities instead: P(A or B) = P(A) + P(B).
Independent events
The outcome of one event has no effect on the probability of the other, e.g. two separate die rolls.
Mutually exclusive events
Events that cannot both happen at the same time, e.g. rolling a 2 and rolling a 5 on the same single die roll.
Combined event
An outcome built from two or more single events considered together as one unit, e.g. 'spinner shows even AND die shows a factor of 6'.

Multiplication rule — valid only for independent events

Memorise

Addition rule — valid only for mutually exclusive events

Memorise

Worked example

Q6 — marbles: single event and comparing bags

[3]
A bag contains 4 red, 3 blue and 5 green marbles. (a) Calculate the total number of marbles. [1] (b) Apply the theoretical probability formula to find P(blue), as a simplified fraction. [2] (c) A second bag holds 6 blue marbles out of 16 total. Compare the probability of drawing blue from each bag. [part]

Frequently confused

Concept AConcept BThe distinction that matters
Independent eventsMutually exclusive eventsIndependent = one outcome doesn't affect the other's probability, and they CAN happen together (use ×); mutually exclusive = the two outcomes physically cannot both occur at once (use +) — mixing these up means adding when you should multiply, or vice versa.

Mnemonic

AND means multiply (×), OR means add (+)

For combined events, decode the connecting word first: 'AND' chains two conditions that both must happen (multiply their probabilities); 'OR' offers alternatives (add them, provided they can't overlap).

Common mistake

Adding P(A) and P(B) when the question says 'AND' (or multiplying when it says 'OR' between mutually exclusive outcomes).

Consequence: The arithmetic can be perfectly correct but answer the wrong question entirely — full method marks are lost because the operation itself is what's being assessed.

Fix: Circle the connecting word before doing any arithmetic. AND → ×. OR (mutually exclusive) → +.

Key takeaways

  • Single event = one action, one outcome recorded.
  • Combined event = two or more actions treated as one unit — decode AND/OR before choosing an operation.
  • P(A and B) = P(A)×P(B) only for independent events; P(A or B) = P(A)+P(B) only for mutually exclusive events.
  • Comparing probabilities always needs a common format (matching decimals or fractions) — never compare raw favourable counts.

Probability with Dice, Coins and Spinners

  • These are the 'equally likely outcomes' machines of the course: a fair die has 6 equally likely faces, a fair coin has 2 equally likely sides, a spinner with equal sectors has as many equally likely landing zones as it has sectors.
  • 'Fair' and 'equal sectors' are doing real mathematical work in the wording — that assumption is exactly what lets you use n(favourable)/n(total) at all.
  • Once wear, weighting, or unequal sector sizes are involved, theoretical probability calculated this way stops being trustworthy — recorded experimental frequencies become the more reliable guide to what's really happening.
  • Spinners with unequal-sized sectors are NOT '1 divided by the number of sectors' — you need each sector's angle (or area) as a fraction of the whole circle (360°).

Dice, coin and spinner — the three standard probability generators

Standard fair generators: a six-sided die, a two-sided coin, and an 8-sector spinner with labelled sectors.

123456HTBonusLose102050100200500

Only needed when a spinner's sectors are NOT equal in size

For equal sectors this simplifies back to 1 ÷ number of sectors.

Worked example

Q2 — a worn spinner

[2]
A game-show spinner has 8 equal sectors: Bonus, Lose, 10, 20, 50, 100, 200, 500. The axle has worn unevenly, favouring certain sectors. Bonus appeared 5 times in 80 recorded spins. (a) Calculate the theoretical probability of landing on Bonus. [1] (b) Explain why the theoretical probability may not accurately reflect the spinner's real behaviour. [1]

Real-world application

Casinos and game-show producers routinely test their dice, roulette wheels and spinners by recording thousands of experimental outcomes — any statistically significant drift from the theoretical probability triggers a physical inspection, exactly the reasoning tested in Q2.

Key takeaways

  • Dice: 6 equally likely faces; coin: 2 equally likely sides; spinner: equally likely only if sectors are equal in size/angle.
  • Unequal-sector spinners need angle/360° (or area/whole), not 1 ÷ number of sectors.
  • Wear, weighting or damage break the 'fair' assumption — theoretical probability then stops matching reality, and experimental data becomes the more reliable evidence.

Using Tree Diagrams and Venn Diagrams (Introductory)

  • A tree diagram maps combined/sequential events branch by branch: each branch is labelled with its probability, and every complete path from start to end represents one possible combined outcome.
  • Rule for reading a tree: multiply the probabilities along a single path (AND — narrowing to one exact sequence) to get that outcome's probability; add the probabilities of different complete paths (OR — separate routes) that all satisfy what you want.
  • A Venn diagram shows how two events overlap inside a rectangle representing the universal set (everything possible); circles represent each event, and the overlapping region is the intersection (both happening at once).
  • At this level, Venn diagrams are mostly given with actual counts placed in each region; probability is then that region's count divided by the total in the universal set.
Tree diagram
A branching diagram showing all possible sequences of outcomes for combined events, with probabilities labelled on each branch.
Venn diagram
A diagram using overlapping circles inside a rectangle (universal set) to show which outcomes belong to one event, both events, or neither.
Intersection (A ∩ B)
Outcomes belonging to both event A and event B at once — represented by the overlapping region of a Venn diagram.

Tree diagram for two coin tosses

First toss branches into H (1/2) and T (1/2); each of those branches again into H (1/2) and T (1/2), giving four paths each with probability 1/4.

HT1/2HHHTTHTT1/4

Venn diagram — two overlapping events

Universal set U contains two overlapping events, Set A and Set B, with counts shown in each region.

USet ASet B12583

Choosing the right diagram

DiagramBest used when…What you read off it
Sample space gridTwo single events combined together, both with a small number of equally likely outcomes (e.g. spinner + die)Every ordered pair of outcomes at a glance — count favourable cells over total cells
Tree diagramSequential events (one after another), especially with different numbers of stages or unequal branch probabilitiesMultiply along one path for AND; add across separate complete paths for OR
Venn diagramTwo events with real overlap in membership (people/items belonging to more than one group)Count in the intersection, union, or 'neither' region directly, divided by the universal set total

Worked example

Two coins, tree diagram — 'exactly one head'

Two fair coins are tossed together. Use a tree diagram to find P(exactly one head).

Common mistake

On a tree diagram, adding branch probabilities along a single path instead of multiplying (e.g. writing 1/2 + 1/2 for two branches in a row).

Consequence: Produces probabilities that don't shrink as you go deeper into the tree — sometimes even exceeding 1, an immediate red flag, and full method marks are lost.

Fix: Along ONE path = multiply (AND, narrowing down to one exact sequence). Across DIFFERENT complete paths = add (OR, separate routes to the wanted result).

Examiner tip

Introductory Venn diagram questions almost always give you actual counts in each region first, not probabilities. Read the totals in every region (only A, only B, both, neither) before touching any formula, and always sanity-check they sum to the total given for the universal set.

Key takeaways

  • Tree diagrams: multiply along one path (AND), add across separate complete paths giving the wanted result (OR).
  • All the path probabilities in a complete tree diagram must sum to 1 — a quick self-check for arithmetic errors.
  • Venn diagrams show overlap: intersection = both, union = either/both, complement = neither — always confirm region counts sum to the universal set total.
  • Choosing tree vs Venn vs sample space grid depends on whether events are sequential, overlapping-membership, or simply two combined single events.

Introduction to Probability and the Probability Scale

  • Probability is a number between 0 and 1 (equivalently 0% to 100%) measuring how likely an event is — 0 means impossible, 1 means certain, 0.5 means an even chance.
  • Every probability value can be written as a fraction, decimal or percentage — any format is fine unless the question specifies one, but fractions should always be simplified fully.
  • An outcome is one possible result of a trial (e.g. 'rolling a 4'); an event is the outcome or collection of outcomes you're actually interested in (e.g. 'rolling an even number' covers outcomes 2, 4 and 6).
  • The basic formula for equally likely outcomes only works because every outcome in the sample space is assumed equally likely — that assumption (a fair die, coin or spinner) is the whole reason the formula is valid.
Outcome
A single possible result of a trial.
Event
An outcome, or set of outcomes, that satisfies some stated condition.
Sample space (S)
The complete list of all possible outcomes of a trial; n(S) is the total count of outcomes in it.
Probability scale
A number line from 0 (impossible) to 1 (certain) used to represent and compare the likelihood of events.

The probability scale

0 = impossible, 0.5 = even chance, 1 = certain, with example events positioned along the scale.

00.250.50.751ImpossibleUnlikelyEven chanceLikelyCertain

Basic probability formula for equally likely outcomes

Memorise

Only valid when the object generating outcomes is fair/unbiased — every outcome must be equally likely.

Worked example

Q7 — why does 'fair' matter?

[2]
A fair six-sided die has faces numbered 1 to 6. (a) Calculate the theoretical probability of rolling a 4, as a fraction. [1] (b) Explain why the die being fair is necessary for the theoretical probability formula to apply. [1]

Frequently confused

Concept AConcept BThe distinction that matters
OutcomeEventAn outcome is one single specific result (e.g. 'rolling a 4'); an event is often a broader condition satisfied by several outcomes (e.g. 'rolling an even number' = outcomes 2, 4 or 6) — always check whether the question wants one specific result or a whole category.

Key takeaways

  • Probability scale runs 0 (impossible) to 1 (certain); 0.5 is an even chance — 'unlikely'/'likely' sit either side of 0.5.
  • P(event) = n(event)/n(S) — valid ONLY when every outcome is equally likely (a fair object).
  • Outcome = one specific result; event = the outcome or group of outcomes satisfying the condition asked about.
  • Fractions, decimals and percentages are interchangeable formats for the same probability — simplify fractions fully.

Listing Outcomes and Sample Space Diagrams

  • Listing outcomes systematically is the foundation skill under everything else: for one die, S = {1, 2, 3, 4, 5, 6} — miss one out and every probability calculated afterwards is wrong.
  • For combined events (two dice, spinner + die, coin + die), the sample space grows multiplicatively: n(S) = n(first event) × n(second event) — a 4-sector spinner with a 6-face die gives 4×6 = 24 total equally likely combined outcomes, never 4+6 = 10.
  • A sample space diagram (two-way table/grid) is the clearest way to list every combined outcome without missing or repeating any: one event's outcomes go along the top, the other's down the side, and each cell is one ordered pair.
  • Once the grid is built, any probability question becomes pure counting: count the cells satisfying the condition, divide by the total number of cells n(S).

Sample space grid: spinner (1–4) × die (1–6)

Spinner outcomes 1–4 across the top, die outcomes 1–6 down the side; each of the 24 cells is one ordered pair (spinner, die).

123456SpinnerDie

Worked example

Q3 — spinner and die sample space

[6]
A spinner has four equal sections numbered 1, 2, 3, 4. A fair six-sided die is rolled at the same time. (a) Construct a sample space table showing all possible outcomes. [2] (b) Calculate P(even on spinner AND factor of 6 on die), using your sample space. [2] (c) A student claims this probability is 1/4. Justify whether the student is correct. [2]

Sample space grid: spinner (1–4) × die (1–6) — answer to Q3(a)

Spinner \ Die123456
1(1,1)(1,2)(1,3)(1,4)(1,5)(1,6)
2(2,1)(2,2)(2,3)(2,4)(2,5)(2,6)
3(3,1)(3,2)(3,3)(3,4)(3,5)(3,6)
4(4,1)(4,2)(4,3)(4,4)(4,5)(4,6)

Common mistake

Adding the number of outcomes from each event instead of multiplying to find the total sample space size (e.g. 4+6=10 instead of 4×6=24 for a spinner and a die).

Consequence: Every subsequent probability calculated from a wrong-sized sample space is wrong, even if the favourable-outcome counting is otherwise perfect — this is a total-method error, not a small slip.

Fix: Total combined outcomes = n(first event) × n(second event), always. Only use addition when combining separate favourable counts WITHIN an already correctly sized sample space.

Examiner tip

'List' or 'construct a sample space' is marked on completeness and structure — a table missing even one cell, or two events' outcomes not clearly written as an ordered pair, loses marks even if your later probability arithmetic is flawless.

Key takeaways

  • List systematically: for one die, {1, 2, 3, 4, 5, 6} — missing an outcome breaks every probability calculated afterwards.
  • Combined sample space size = n(first) × n(second), never n(first) + n(second).
  • A sample space grid turns any combined-event probability question into counting cells over total cells.
  • 'Justify' answers need the correct calculated value AND an explicit verdict on whether the given claim is right or wrong.

Quick Reference

Probability and Outcomes — everything on one page

Formulas

  • Theoretical probability — equally likely outcomes only
  • Experimental probability from recorded data
  • A prediction, not guaranteed to match observed data
  • Independent combined events only
  • Mutually exclusive events only
  • Size of a combined sample space — never add the two counts

Key facts

Probability of an impossible event
0
Probability of a certain event
1
Even chance
0.5 = 1/2 = 50%
Sum of all probabilities in a sample space
1

Mnemonics

  • AND = ×, OR = +Decode the connecting word before choosing to multiply or add for combined events.

Don't confuse

  • Theoretical probability vs Experimental probability: Calculated by reasoning (before trials) vs measured from data (after trials) — they converge only over many trials.
  • Independent events vs Mutually exclusive events: Can happen together, don't affect each other (×) vs cannot both happen at once (+).
  • Outcome vs Event: One specific result vs a category that can include several outcomes.

Command terms

Calculate:
Show the formula, substitute values, give a numeric final answer — usually a simplified fraction or decimal.
Explain:
Give a reason tied to the maths or physical situation, not a restated fact.
Construct:
Build the actual diagram/table fully and correctly — marked on completeness, not just the final number.
Justify:
State a clear verdict (correct/incorrect) AND back it with the correctly calculated value.
Compare:
Convert both values to the same format and explicitly state which is bigger/smaller and by how much.

Exam checklist

  • Have I identified whether this needs theoretical (formula) or experimental (data) probability?
  • Is the sample space size found by multiplying, not adding, event outcome counts?
  • Have I decoded AND (×) vs OR (+) correctly for combined events?
  • Is my final probability between 0 and 1, and fully simplified if a fraction was requested?
  • For 'justify' questions, have I stated a clear verdict, not just a number?