Probability questions often appear easy because the numbers are small. The difficult part is not multiplying two fractions. It is deciding whether multiplication is appropriate, whether the fractions still describe the right event, and whether an earlier outcome changes a later one. A neat calculation with a mistaken event model is still a wrong answer.
This lesson develops the method-selection skill in G3 Mathematics K310 for the 2027 Singapore–Cambridge Secondary Education Certificate (SEC). The official SEAB G3 directory confirms the subject code, while the K310 syllabus includes simple combined events, possibility diagrams, tree diagrams and addition or multiplication of probabilities for mutually exclusive and independent events. It describes two Mathematics papers, each 90 marks and 2 hours 15 minutes. Essential working must be shown.
This volume does not duplicate Vol 0178 on basic probability trees, Vol 0363 on recovering missing counts, or Vol 0359 on cumulative frequency. Those are useful owners for their respective skills. Here the decision is earlier: before calculating, determine exactly what ‘or’, ‘and’, ‘given’ and ‘after’ mean in the event the question describes.
Use the Secondary 4 Mathematics guide when you need a fuller route through combined events, the PSLE learning guide when fractions and simple chance are unstable, and the G3 SEC learner hub for the overall examination path. A student who can compute 3/5 × 2/5 but does not know which event those fractions refer to needs model repair, not more speed drills.
The most useful habit is to write one line in words before doing any arithmetic: ‘I am finding the chance that …’. Finish the sentence precisely. Then decide whether the events can overlap, whether they are separate stages, whether the first action changes what is available, and whether all possible outcomes have been counted.
Two rules deserve special attention. For events A and B that cannot occur together, P(A or B) = P(A) + P(B). For independent events, P(A and B) = P(A) × P(B). They answer different questions. Mutually exclusive does not mean independent. Except in special zero-probability cases, two non-empty mutually exclusive events cannot be independent because one happening rules out the other.
An exam question might use a bag, spinner, two dice, a selection of students, or the outcomes of two tests. Each context is only a surface. The mathematical engine is the event structure, and the skill is to recognise that structure while showing the required working.
The four sentences that prevent wrong methods
Start with ‘A or B’. Does this mean either event, including the possibility that both occur? In ordinary probability language, the union of events includes outcomes belonging to A, to B, or to both. If the sets overlap, adding their individual probabilities counts the overlap twice. The correction is P(A or B) = P(A) + P(B) − P(A and B). The simpler sum is reserved for mutually exclusive events.
Next read ‘A and B’. This often calls for a combined-event calculation, but you must ask whether the second probability is unchanged or depends on the first event. Independent trials such as two separate fair coin tosses allow multiplication of their fixed probabilities. Drawing from a bag without replacement changes the contents of the bag and therefore the second fraction.
Now read ‘at least one’. Its complement is ‘none’. When there are several independent attempts, finding P(none) and subtracting from 1 can be more efficient than listing every success path. But the complement method still depends on describing the sample space correctly.
Finally read ‘given that’ or ‘after’. The words indicate a restriction or a stage condition. The denominators must come from the outcomes still possible under that information. A student should write the reduced pool explicitly instead of continuing with the original total out of habit.
Mutually exclusive is about overlap
On one throw of a standard fair die, the event ‘the number is 1’ and the event ‘the number is 6’ cannot happen together. Their intersection is empty, so the chance of either is 1/6 + 1/6 = 1/3. This is a simple sum because the two successful outcomes belong to different places in the sample space.
Now compare ‘even’ and ‘greater than 3’ on one die. Even outcomes are {2,4,6}; greater than 3 gives {4,5,6}. Both include {4,6}. Adding 3/6 and 3/6 gives 1, but the union is {2,4,5,6}, four outcomes out of six. Subtract the 2/6 overlap: 3/6 + 3/6 − 2/6 = 4/6 = 2/3.
The correct question is not whether the two descriptions sound different. It is whether an individual elementary outcome can satisfy both descriptions. A single die result of 4 is both even and greater than 3. Its membership in two sets does not mean there are two separate die results.
Independent is about whether a result changes another chance
Two separate tosses of a fair coin are independent in the ideal model. The outcome of the first toss does not alter the chance of heads on the second. Thus P(two heads) = 1/2 × 1/2 = 1/4. The tree has four equally likely ordered outcomes: HH, HT, TH and TT.
In contrast, when you draw two cards without replacement from a small bag, the first selection changes the composition for the second selection. The conditional fraction must use the remaining counts. If a bag has three red and two blue counters, the probability of red then blue is 3/5 × 2/4 = 3/10, not 3/5 × 2/5.
Independence is a property of the probability relationship, not merely a claim that actions happened at different times. Two stages may be dependent because the first physically changes what is available. Two variables measured at the same time may also have a statistical relationship. Use the mechanism and the model given by the question.
Tree diagrams are accounting devices, not decoration
A probability tree records the stages and the probability at each branch. Outgoing branches from a node must add to 1 when they cover every possible outcome of that stage. Multiply along a complete path to obtain that ordered outcome’s probability; add across disjoint complete paths when the target event has several ways to occur.
Suppose the question asks for one of each colour in two draws without replacement. ‘Red then blue’ and ‘blue then red’ are different ordered paths, and together they represent the unordered result ‘one red and one blue’. For the bag of three red and two blue counters, the two probabilities are 3/5 × 2/4 and 2/5 × 3/4, giving 3/10 + 3/10 = 3/5.
A tree may be unnecessary for a single event, but it becomes valuable when branch denominators change or the required outcome can follow several ordered paths. Keep the branch labels beside their fractions and write the final event as a sentence so that a reader can see what was counted.
Use a sample-space check at the end
Every valid probability lies between 0 and 1 inclusive. The probabilities of all mutually exclusive, exhaustive final outcomes must sum to 1. If your result is negative or greater than 1, something has gone wrong in the construction or arithmetic. If two complementary events do not sum to 1, look for an omitted or double-counted outcome.
Sanity checks can also expose answers that are numerically possible but implausible. Drawing two red counters without replacement should be impossible if the bag contains only one red counter. Getting two identical numbers on a pair of fair dice has probability 6/36, not 1/2. These checks bring meaning back into arithmetic.
The aim is not to memorise a giant formula sheet. It is to build a reliable small vocabulary: sample space, event, union, intersection, complement, independent, without replacement, ordered path. Each term gives a question you can ask of the situation.
How to earn marks when the method is difficult
Show the event first, then the calculation. In a tree problem, a small clean diagram with branch fractions can make the method readable, even if a later multiplication error occurs. In a union question, display the overlap term rather than jumping from two fractions to a total that cannot be checked.
A calculator can check arithmetic but it cannot tell you whether ‘at least one’ includes two successes, whether colours are returned to the bag, or whether a survey subgroup is the denominator. A good student learns to use the calculator after making the model, not before.
Under exam pressure, write a one-sentence method justification where it clarifies a choice: ‘The draws are without replacement, so the second denominator is 4’; or ‘The events are mutually exclusive, so their probabilities can be added’. These explanations are short and tightly linked to the working.
Worked studio: thirty method-selection decisions
Case 01: One roll, either endpoint
Task. A fair six-sided die is rolled once. Find the probability of 1 or 6.
Solution. The events are mutually exclusive. There are two successful outcomes from six equally likely outcomes; P = 1/6 + 1/6 = 1/3.
Why this method. The word ‘or’ does not automatically justify addition; here addition works because 1 and 6 cannot occur on the same roll.
Before reading the answer, underline the event words. Decide whether you are selecting a single result, travelling down a multi-stage path, or collecting several complete paths. This simple annotation is a stronger habit than hunting for a formula by keyword. The same numerical fractions can be used correctly or incorrectly depending on what they represent.
Case 02: Even and high numbers
Task. A die is rolled once. Find the probability that the number is even or greater than 3.
Solution. Even {2,4,6} and greater than 3 {4,5,6} overlap at {4,6}. P = 3/6 + 3/6 − 2/6 = 4/6 = 2/3.
Why this method. An answer of 1 double-counts outcomes 4 and 6.
To explain your method, give the denominator a name. Is it the whole sample space, the remaining counters, or the outcomes inside a subgroup? Many wrong answers are not failures of multiplication. They are failures to know which population the fraction describes at that moment.
Case 03: Odd or prime
Task. On one fair die, find the probability that the result is odd or prime.
Solution. Odd outcomes {1,3,5} and primes {2,3,5} overlap at {3,5}. The union is {1,2,3,5}, giving 4/6 = 2/3.
Why this method. Do not treat 1 as prime; it is odd but not a prime number.
Check the result from the other direction when possible. A complement, complete list of outcomes or sum of final tree-path probabilities often catches an error quickly. Use that check to test the model rather than merely recalculating the same mistaken expression.
Case 04: Two independent coins
Task. A fair coin is tossed twice. Find the probability of two heads.
Solution. For independent fair tosses, P(HH) = (1/2)(1/2) = 1/4.
Why this method. Write the ordered outcome HH before the multiplication. The first result does not change the next.
Compare this case with a nearby alternative. Would returning an object change the second fraction? Would changing ‘or’ to ‘and’ change the required set? Would an overlap make a direct sum invalid? A student who can answer those three questions has understood more than the isolated number.
Case 05: Exactly one head
Task. Two fair coins are tossed. Find the probability of exactly one head.
Solution. The mutually exclusive ordered outcomes are HT and TH, each probability 1/4, so the answer is 1/2.
Why this method. The phrase ‘exactly one’ excludes HH as well as TT.
In examination writing, show enough steps for the reasoning to be visible. Write the qualifying paths or outcome sets, then the arithmetic, then the final probability. The aim is not maximum length. It is to make the method auditable and to remove ambiguity about the event you counted.
Case 06: At least one head
Task. Two fair coins are tossed. Find the probability of at least one head.
Solution. The complement is TT with probability 1/4. Thus P(at least one head) = 1 − 1/4 = 3/4.
Why this method. At least one includes exactly one and two heads; stopping after HT and TH misses HH.
Before reading the answer, underline the event words. Decide whether you are selecting a single result, travelling down a multi-stage path, or collecting several complete paths. This simple annotation is a stronger habit than hunting for a formula by keyword. The same numerical fractions can be used correctly or incorrectly depending on what they represent.
Case 07: Two sixes
Task. A fair die is rolled twice. Find the probability of getting a six both times.
Solution. Each independent chance is 1/6, so the ordered pair (6,6) has probability 1/36.
Why this method. Do not add 1/6 and 1/6 for an ‘and’ event.
To explain your method, give the denominator a name. Is it the whole sample space, the remaining counters, or the outcomes inside a subgroup? Many wrong answers are not failures of multiplication. They are failures to know which population the fraction describes at that moment.
Case 08: At least one six
Task. A fair die is rolled twice. Find the probability of at least one six.
Solution. The complement is no six on either roll: (5/6)^2 = 25/36. The answer is 1 − 25/36 = 11/36.
Why this method. The two success events overlap when both rolls are six; direct addition without correcting the overlap overcounts.
Check the result from the other direction when possible. A complement, complete list of outcomes or sum of final tree-path probabilities often catches an error quickly. Use that check to test the model rather than merely recalculating the same mistaken expression.
Case 09: With-replacement colours
Task. A bag holds 3 red and 2 blue counters. A counter is selected, returned, and a second is drawn. Find red then blue.
Solution. Returning the first counter keeps the contents unchanged. P(R then B) = 3/5 × 2/5 = 6/25.
Why this method. The problem explicitly restores the bag; do not reduce the denominator.
Compare this case with a nearby alternative. Would returning an object change the second fraction? Would changing ‘or’ to ‘and’ change the required set? Would an overlap make a direct sum invalid? A student who can answer those three questions has understood more than the isolated number.
Case 10: Without-replacement colours
Task. Use the same bag, but do not return the first selected counter. Find red then blue.
Solution. The chance of red first is 3/5. After that draw, 2 blue remain among 4 total, so P(R then B) = 3/5 × 2/4 = 3/10.
Why this method. The denominator 4 is essential; copying the earlier fraction would model a different experiment.
In examination writing, show enough steps for the reasoning to be visible. Write the qualifying paths or outcome sets, then the arithmetic, then the final probability. The aim is not maximum length. It is to make the method auditable and to remove ambiguity about the event you counted.
Case 11: One of each in any order
Task. From a bag of 3 red and 2 blue counters, draw two without replacement. Find one of each colour.
Solution. Add the ordered paths RB and BR: (3/5)(2/4) + (2/5)(3/4) = 3/10 + 3/10 = 3/5.
Why this method. A single path gives only one order. The target event includes two distinct orders.
Before reading the answer, underline the event words. Decide whether you are selecting a single result, travelling down a multi-stage path, or collecting several complete paths. This simple annotation is a stronger habit than hunting for a formula by keyword. The same numerical fractions can be used correctly or incorrectly depending on what they represent.
Case 12: Two red counters
Task. From 3 red and 2 blue, two counters are drawn without replacement. Find the chance of both red.
Solution. P(RR) = 3/5 × 2/4 = 6/20 = 3/10.
Why this method. The second red numerator falls because the first red counter has been removed.
To explain your method, give the denominator a name. Is it the whole sample space, the remaining counters, or the outcomes inside a subgroup? Many wrong answers are not failures of multiplication. They are failures to know which population the fraction describes at that moment.
Case 13: No blue counters
Task. From the same bag, two counters are drawn without replacement. Find the probability of no blue.
Solution. No blue means RR in this two-colour bag, so P = 3/10.
Why this method. It is easier to name the only qualifying path than to list rejected paths.
Check the result from the other direction when possible. A complement, complete list of outcomes or sum of final tree-path probabilities often catches an error quickly. Use that check to test the model rather than merely recalculating the same mistaken expression.
Case 14: At least one blue
Task. From 3 red and 2 blue, select two without replacement. Find the probability of at least one blue.
Solution. The complement is RR, probability 3/10. Hence P = 1 − 3/10 = 7/10.
Why this method. The complement method prevents a common error: missing BB or one of the mixed-colour orders.
Compare this case with a nearby alternative. Would returning an object change the second fraction? Would changing ‘or’ to ‘and’ change the required set? Would an overlap make a direct sum invalid? A student who can answer those three questions has understood more than the isolated number.
Case 15: Two blues
Task. From 3 red and 2 blue, draw two without replacement. Find the probability both are blue.
Solution. P(BB) = 2/5 × 1/4 = 1/10.
Why this method. Only one blue remains after the first; writing 2/4 for the second branch is wrong.
In examination writing, show enough steps for the reasoning to be visible. Write the qualifying paths or outcome sets, then the arithmetic, then the final probability. The aim is not maximum length. It is to make the method auditable and to remove ambiguity about the event you counted.
Case 16: Impossible second red
Task. A bag has 1 red and 4 blue counters. Two are drawn without replacement. Find two red.
Solution. After a red first draw, no red remains; the probability is zero.
Why this method. The arithmetic 1/5 × 0/4 = 0 agrees with the physical impossibility. Checking meaning is stronger than trusting an isolated decimal.
Before reading the answer, underline the event words. Decide whether you are selecting a single result, travelling down a multi-stage path, or collecting several complete paths. This simple annotation is a stronger habit than hunting for a formula by keyword. The same numerical fractions can be used correctly or incorrectly depending on what they represent.
Case 17: Spinner and die
Task. A fair two-colour spinner has equal sectors and a fair die is rolled independently. Find blue on the spinner and 5 on the die.
Solution. P = 1/2 × 1/6 = 1/12, under the independent fair-device model.
Why this method. Keep the devices separate. The spinner colour does not remove a die face.
To explain your method, give the denominator a name. Is it the whole sample space, the remaining counters, or the outcomes inside a subgroup? Many wrong answers are not failures of multiplication. They are failures to know which population the fraction describes at that moment.
Case 18: An overlapping card label
Task. One card is picked at random from cards numbered 1–10. Find ‘multiple of 2 or multiple of 5’.
Solution. Multiples of 2 are {2,4,6,8,10}; multiples of 5 are {5,10}; overlap {10}. P = (5+2−1)/10 = 6/10 = 3/5.
Why this method. The number 10 belongs to both descriptions but is only one card.
Check the result from the other direction when possible. A complement, complete list of outcomes or sum of final tree-path probabilities often catches an error quickly. Use that check to test the model rather than merely recalculating the same mistaken expression.
Case 19: A survey subgroup
Task. In a class of 40 pupils, 18 join art club and 12 join science club, with 5 in both. Find the chance a random pupil joins either.
Solution. The union count is 18+12−5=25. Probability =25/40=5/8.
Why this method. Subtract overlap once, and keep the denominator as the whole class when the pupil is selected from the whole class.
Compare this case with a nearby alternative. Would returning an object change the second fraction? Would changing ‘or’ to ‘and’ change the required set? Would an overlap make a direct sum invalid? A student who can answer those three questions has understood more than the isolated number.
Case 20: Neither of two clubs
Task. Use the same 40 pupils and 25 in at least one of two clubs. Find the probability of neither.
Solution. There are 40−25=15 in neither. The probability is 15/40=3/8.
Why this method. The result complements 5/8 and is a useful check.
In examination writing, show enough steps for the reasoning to be visible. Write the qualifying paths or outcome sets, then the arithmetic, then the final probability. The aim is not maximum length. It is to make the method auditable and to remove ambiguity about the event you counted.
Case 21: A changed selection pool
Task. An organisation first selects a volunteer from 7 trained and 3 untrained people. Without replacing that person, it selects another. Find two trained volunteers.
Solution. The first chance is 7/10; after one trained volunteer is selected, the second is 6/9. Product = 42/90=7/15.
Why this method. The word ‘another’ plus no replacement matters. Do not assume repeated independent selections.
Before reading the answer, underline the event words. Decide whether you are selecting a single result, travelling down a multi-stage path, or collecting several complete paths. This simple annotation is a stronger habit than hunting for a formula by keyword. The same numerical fractions can be used correctly or incorrectly depending on what they represent.
Case 22: Return before repeating
Task. Use the same group, but suppose the first volunteer is returned to the selection pool before the second draw.
Solution. Each probability of trained is 7/10; P(two trained) = 49/100.
Why this method. The only change from the prior case is replacement. The changed answer demonstrates why reading the mechanism matters.
To explain your method, give the denominator a name. Is it the whole sample space, the remaining counters, or the outcomes inside a subgroup? Many wrong answers are not failures of multiplication. They are failures to know which population the fraction describes at that moment.
Case 23: Order and a fixed final colour
Task. A bag holds 4 green and 1 yellow token. Two are drawn without replacement. Find green then yellow.
Solution. P(GY) = 4/5 × 1/4 = 1/5.
Why this method. The first green event leaves the single yellow among four tokens; confirm that both branches are feasible.
Check the result from the other direction when possible. A complement, complete list of outcomes or sum of final tree-path probabilities often catches an error quickly. Use that check to test the model rather than merely recalculating the same mistaken expression.
Case 24: At least one yellow
Task. With 4 green and 1 yellow, draw two without replacement. Find at least one yellow.
Solution. At most one yellow is possible. The complement is both green: (4/5)(3/4)=3/5. Thus answer is 2/5.
Why this method. The outcome cannot include two yellow tokens, because the bag has only one.
Compare this case with a nearby alternative. Would returning an object change the second fraction? Would changing ‘or’ to ‘and’ change the required set? Would an overlap make a direct sum invalid? A student who can answer those three questions has understood more than the isolated number.
Case 25: An erroneous sum
Task. A candidate writes P(A or B) = P(A)+P(B) for A=’even die result’ and B=’prime die result’.
Solution. Even {2,4,6}, prime {2,3,5}, overlap {2}. The union has five results, so answer is 5/6.
Why this method. The assumption of exclusivity is false, and the correct subtraction term is 1/6.
In examination writing, show enough steps for the reasoning to be visible. Write the qualifying paths or outcome sets, then the arithmetic, then the final probability. The aim is not maximum length. It is to make the method auditable and to remove ambiguity about the event you counted.
Case 26: An erroneous product
Task. A learner multiplies 3/5 × 3/5 for two red draws from 3 red and 2 blue without replacement.
Solution. The second red chance, after a red first draw, is 2/4. Correct product is 3/5 × 2/4 = 3/10.
Why this method. Name the changed pool in a short phrase; that one phrase explains the entire correction.
Before reading the answer, underline the event words. Decide whether you are selecting a single result, travelling down a multi-stage path, or collecting several complete paths. This simple annotation is a stronger habit than hunting for a formula by keyword. The same numerical fractions can be used correctly or incorrectly depending on what they represent.
Case 27: A total probability test
Task. A two-draw bag experiment has final ordered paths RR=3/10, RB=3/10, BR=3/10, BB=1/10. Are the probabilities exhaustive?
Solution. The four path probabilities sum to 1. They describe every possible ordered pair of colour outcomes for the given bag.
Why this method. If a tree’s final paths do not sum to 1, look for missing branches or incorrect fractions.
To explain your method, give the denominator a name. Is it the whole sample space, the remaining counters, or the outcomes inside a subgroup? Many wrong answers are not failures of multiplication. They are failures to know which population the fraction describes at that moment.
Case 28: An exam explanation in one line
Task. A pupil finds 7/10 for at least one blue from a 3-red 2-blue bag, but gives no working.
Solution. A compact explanation is: P(at least one blue)=1−P(RR)=1−(3/5)(2/4)=7/10.
Why this method. The written event and complement establish that the result follows from the model rather than a lucky guess.
Check the result from the other direction when possible. A complement, complete list of outcomes or sum of final tree-path probabilities often catches an error quickly. Use that check to test the model rather than merely recalculating the same mistaken expression.
Case 29: Choosing a sensible check
Task. A result for a simple probability event is 9/8. The pupil says this is plausible because there were nine successful cases.
Solution. A probability cannot exceed 1; 9/8 signals a wrong denominator, overlap count or other mistake.
Why this method. Rebuild the sample space before adjusting arithmetic. A cosmetic change to the fraction can conceal the original modelling error.
Compare this case with a nearby alternative. Would returning an object change the second fraction? Would changing ‘or’ to ‘and’ change the required set? Would an overlap make a direct sum invalid? A student who can answer those three questions has understood more than the isolated number.
Case 30: A diagnostic on independence
Task. A pupil argues two mutually exclusive non-empty events are independent because they occur on one roll.
Solution. If A occurs, B cannot occur; the occurrence changes the status of the other. They cannot be independent when both have positive probability.
Why this method. Ask separate questions: can both happen together, and does one event change the likelihood of the other? They are different tests.
In examination writing, show enough steps for the reasoning to be visible. Write the qualifying paths or outcome sets, then the arithmetic, then the final probability. The aim is not maximum length. It is to make the method auditable and to remove ambiguity about the event you counted.
Build a small decision table in your own handwriting
Write five headings down a page: one event; either event; both events; at least one; and repeated stages. Beside each, write an original example, one line about the event structure, and one correct calculation. Include a case with overlap, an independent pair of events and a draw without replacement. This forces you to decide the method before you see a formula.
Next hide the worked solutions, repeat six cases in random order and explain why the method works. If you can perform the arithmetic but cannot say whether two events overlap or whether a denominator changes, return to that specific concept. If you can explain the event structure but make small fraction errors, practise that arithmetic separately. Diagnosis saves time.
For reference, the official 2027 Mathematics K310 syllabus determines the examined content and paper arrangement. This teaching material uses original practice questions and does not reproduce official SEC examination questions. Continue through the G3 SEC learner guide when you want the English–Mathematics–Science route.
