Mathematics / Preview lesson sketch 05
Probability as evidence
Use distributions to reason about uncertainty without pretending it disappears.
It remains available from the sealed 0.1 preview while the deeper Foundations format is reviewed and expanded. No study-time claim is made.
orient
Why this idea had to exist
Measurements vary, samples omit information, and predictions concern futures we cannot inspect. Probability gives a coherent ledger for uncertainty; statistics connects that ledger to observed evidence.
intuition
Build a picture you can reason with
A distribution is a landscape of plausible outcomes, not one promised answer. Sampling explores the landscape. Conditioning reshapes it after new information rules out or favors some regions.
formalize
Give the intuition a precise edge
For events A and B with P(B)>0, P(A|B)=P(A∩B)/P(B). An expected value is a probability-weighted average. A sample statistic varies across possible samples; its sampling distribution describes that variation.
work through
Follow the decisions, not just the symbols
Suppose 1 percent of devices fail and a test catches 95 percent of failures but falsely flags 5 percent of working devices. Out of 10,000 devices, about 95 true failures and 495 working devices are flagged. A flag therefore implies about 95/590, not 95 percent, chance of failure.
experiment
Change one thing and watch the model answer
Draw repeated samples from a skewed population. Compare the changing sample values with the steadier distribution of sample means as sample size grows.
retrieve
Close the page and reconstruct it
Answer before opening the explanation. Retrieval is evidence only when the answer is produced without a hint.
What does conditioning change?
It changes the reference set of possible outcomes by incorporating the information that the conditioning event occurred.
transfer
Move the idea into a new setting
Write a short evidence note explaining why model accuracy alone can be misleading when one class is rare.
reflect
Leave with a diagnostic habit
Separate what was observed, what process may have produced it, and what remains uncertain. Confidence should follow evidence, not replace it.
Source record
Follow the sketch back.
- Reference
- Introductory Statistics 2e
- Publisher
- OpenStax
- Locator
- Whole-source locator retained from v1; precise locator required before conversion.
- Review
- Pending