Theory
The number for "maybe"
From the CampusPulse survey: 24 of the 60 surveyed students are hostellers. The principal picks one filled form blindly from the pile. How likely is it a hosteller's?
Your instinct already whispers "a bit less than half". Statistics upgrades that whisper to a number: 24/60 = 0.4.
Probability is the mathematics of maybe: a 0-to-1 scale for uncertainty, and the language every distribution, sample and bell curve in this subject speaks from here on.
Theory
A fair lottery drum
Every classical probability question is secretly a lottery drum: all possible outcomes written on identical chits, drum spun, one chit drawn blind.
P(event) = chits you would celebrate ÷ chits in the drum.
A die is a 6-chit drum (3 even chits: P = 3/6). A card deck is a 52-chit drum (4 kings: P = 4/52). The survey pile is a 60-chit drum (24 hosteller chits: P = 0.4). Identical chits is the fine print: the drum must be fair.
Theory
The vocabulary, formally
- Experiment: a repeatable process with an uncertain result (toss a coin, draw a form).
- Sample space S: ALL possible outcomes: for a die, {1, 2, 3, 4, 5, 6}.
- Event: the subset you care about: "even number" = {2, 4, 6}.
Classical probability (equally likely outcomes):
P(E) = favourable outcomes ÷ total outcomes
Always 0 ≤ P ≤ 1: 0 = impossible, 1 = certain, and all outcomes' probabilities sum to 1. (When outcomes are NOT equally likely, probability comes from observed data instead: the empirical approach: relative frequency.)
Theory
Three rules do the heavy lifting
Complement: P(not E) = 1 - P(E). P(no six) = 1 - 1/6 = 5/6.
Addition: P(A or B) = P(A) + P(B) - P(A and B). The subtraction stops double-counting the overlap; for mutually exclusive events (cannot co-occur) the overlap is 0.
Multiplication (independent events): P(A and B) = P(A) × P(B), when one outcome tells you nothing about the other: two coins, P(two heads) = 1/2 × 1/2 = 1/4.
Think first
Work the overlap by hand
One card from a 52-card deck. On paper compute P(king OR heart), using the addition rule: count kings, count hearts, count the overlap. Then tap.
Show the answer
P(king) = 4/52. P(heart) = 13/52. Overlap: the king of hearts, counted in both: 1/52.
P(king or heart) = 4/52 + 13/52 - 1/52 = 16/52 = 4/13.
Sanity check by direct count: 13 hearts plus the 3 non-heart kings = 16 favourable chits. Forgetting the subtraction gives 17/52: the single most common addition-rule error, and the exact trap the next quiz springs.
Quiz
In the survey, 24 of 60 are hostellers and 30 of 60 are in Year 2; 12 students are BOTH. P(hosteller OR Year 2) = ?
- (24 + 30 - 12)/60 = 42/60 = 0.7
- (24 + 30)/60 = 54/60 = 0.9
- 24/60 × 30/60 = 0.2
- 12/60 = 0.2
Show the answer
(24 + 30 - 12)/60 = 42/60 = 0.7
The 12 who are both would be counted twice in 24 + 30, so the addition rule subtracts them once: 42/60 = 0.7. Option B is the double-count. Option C multiplies, which answers a different question (an AND, and only for independent events, which these are not: 12/60 ≠ 0.4 × 0.5). Option D reports only the overlap. Or-questions = add and subtract the overlap, every time.
Watch out
The three probability sins
Fake 50-50s: "either it rains or it does not, so P = 1/2": outcomes must be equally likely for the classical formula; rain is not a fair coin.
The gambler's fallacy: five heads in a row do NOT make tails "due": independent tosses have no memory: the sixth is still 1/2.
Impossible numbers: any probability below 0 or above 1 means an arithmetic error: stop and recheck, free marks saved.
Theory
Why this unit needed probability now
Everything ahead is probability wearing costumes: a distribution (next lesson) is just every outcome listed with its probability; sampling asks how probable it is that your 60-student spoonful misleads you; the bell curve assigns probabilities to ranges of heights. Even Grishu's own quiz analytics ask "what is the probability a student who missed THIS question fails the exam?". The 0-to-1 scale you learned today is the subject's currency.
Summary
Key takeaways
- Experiment → sample space (all outcomes) → event (the subset you care about).
- Classical rule: P = favourable/total, ONLY for equally likely outcomes; scale 0 to 1, total 1.
- Complement: P(not E) = 1 - P(E).
- Addition: P(A or B) = P(A) + P(B) - P(A and B); overlap = 0 for mutually exclusive events.
- Multiplication for independent events: P(A and B) = P(A) × P(B); tosses have no memory.
- Not equally likely? Use observed relative frequency (empirical probability).
- Memory hook: a fair lottery drum: celebrate-chits over total chits.