University College London · FACULTY OF STATISTICS

SECU0006 Chap.1 Uncertainty and Probability

- one subject, every graph, every model, every mark
6 Chapters2-page Bible
Our own words - no uploaded lecturer files
Updated for this semester
Chapter 1 of 6 · SECU0006

Uncertainty and Probability

Uncertainty and Probability introduces events, complements, unions, intersections and conditioning as tools for reasoning about uncertain real-world phenomena. It emphasises defining the population and denominator before calculating, distinguishing independence from mutual exclusivity, and comparing competing explanations when evidence is rare.

Frequency tables and probability trees provide independent checks on conditional calculations.

In this chapter

What this chapter covers

  • 01

    Events and sample spaces

  • 02

    Complements and unions

  • 03

    Conditional probability

  • 04

    Independence

  • 05

    Probability trees

  • 06

    Evidence under alternatives

Worked example · free

Update a screening probability

Q [4 marks]. A rule flags 90 true cases and 88 other cases. What share of all flagged cases are true? The four marks shown are an original revision allocation, not a published university scheme. This revision weighting was authored for this guide and is not an official marking scheme. The mark allocation for this probability drill is not an official university marking scheme.
The branches show a screening flag followed by a review decision; multiply along a path and add disjoint paths that reach the same outcome.The branches show a screening flag followed by a review decision; multiply along a path and add disjoint paths that reach the same outcome.startflagged 0.20clear 0.80review 0.70no review 0.30review 0.10no review 0.90
The branches show a screening flag followed by a review decision; multiply along a path and add disjoint paths that reach the same outcome.
  • 1Total flagged is 178.
  • 2Compute 90/178 = 0.506.
  • 1Interpret for the stated population.
The conditional share is about 50.6%. The marks are original practice weighting.
Sia tip — Name the conditioning group before choosing the denominator.
Glossary

Key terms

Sample space
The set of outcomes considered possible in a model.
Independence
A relationship where learning one event does not change the probability of another.
Joint event
The event that two stated conditions both occur.
FAQ

Uncertainty and Probability FAQ

Why does conditioning change a probability?

It restricts the denominator to cases where the stated condition is already known, so the comparison group changes. Recalculate after changing the conditioning set to see how strongly the result depends on that choice.

Are disjoint events independent?

Not usually. If one disjoint event occurs, the other cannot occur, so its conditional probability becomes zero. Distinguish both ideas by writing the relevant conditional probability rather than relying on the event labels.

How can I check a probability tree?

Confirm branches from each node sum to one, multiply along paths and add only disjoint terminal paths. A frequency table should reproduce the same terminal totals and provides a useful independent check.

Study strategy

Assessment move

Draw the sample space and write events in words before using symbols. Solve each conditional problem twice, once with a tree or formula and once with a frequency table. Check complements, branch totals and probability bounds. Practise reversing the condition so the denominator difference becomes automatic. In evidence questions, compare the observation under more than one explanation and keep dependence assumptions explicit.

Build a weekly retrieval set around event direction. Start with complements and unions, then move to conditional probabilities represented as tables and trees. For every answer, write the population, event and conditioned group before any arithmetic. Practise pairs in which the same counts support P(A given B) and P(B given A), because confusing those directions is a common conceptual failure.

Change an independence assumption and explain which multiplication step must change. When using repeated trials, decide whether shared causes make outcomes dependent. Check every final probability against zero and one, reconstruct terminal totals and translate the result back into a natural frequency. For evidence problems, list at least two competing hypotheses and compare how expected the observation would be under each.

Avoid treating a rare event as proof of a preferred cause. Keep a short error log with categories for reversed conditioning, double-counted overlap, false independence and an omitted complement. Rework each error with new numbers until both the formula and the frequency-table route agree. In timed revision, alternate a concise single-answer calculation with a short explanation of why the denominator is correct.

Finish by stating which assumption would most change the result and what information could test it. Add a second pass devoted to explanation rather than arithmetic. Take a completed probability problem and remove the numbers, then describe which event is conditioned upon, why the branch probabilities may be multiplied and which paths may be added. Reintroduce different numbers only after the logic is stable.

Create two-by-two tables from verbal scenarios and read every conditional proportion in both directions. Use natural frequencies to check Bayes-style updates, because counts expose denominator reversals more clearly than symbols alone. When an independence claim appears, write the equality it requires and ask whether the setting supplies shared causes.

Distinguish mutually exclusive events from independent ones with a counterexample. For unions, shade overlapping sets before using the addition rule. For at-least-one questions, calculate the complement and explain why it is simpler. During timed practice, allocate a brief checking phase to branch totals and probability bounds.

In short answers, name the consequence of a violated assumption rather than saying only that the model is imperfect. Review the supplied evidence case for its reasoning structure, but use fresh contexts and numbers in practice. The target skill is a transparent path from uncertainty to a probability statement whose population, condition and limitations are all visible.

End the weekly review by explaining one conditional result aloud without symbols, then reconstruct the symbols and verify both representations use the same denominator. Use the probability checklist to design another conditional scenario and confirm that denominator direction remains correct without prompts.

Working through Uncertainty and Probability in SECU0006? Sia is AskSia’s AI Statistics tutor — ask any SECU0006 Uncertainty and Probability question and get a clear, step-by-step explanation grounded in how SECU0006 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.

A+Everything unlocked
Unlocks this Bible + all 32 of your subjects at University College London, plus 1,000+ more Bibles across the AskSia library.
Sia - your SECU0006 tutor, unlimited, worked the way the exam marks it
The full 2-page Bible + practice bank with worked solutions
Chrome extension - sync your LMS so Sia knows your deadlines
Bilingual EN / Chinese on every Bible and every Sia answer
$0.99 Trial
30-day money-back · cancel in one tap · how it works