Statistics, Probability & Linear Programming

1. Core Concept

Statistics quantifies the spread (dispersion) of data around a central value. Probability mathematically measures the likelihood of uncertain events using set theory and conditioning. Linear Programming (LP) optimizes a linear objective function subject to linear constraints, typically solved using the graphical "Feasible Region" method.

2. Key Formulas

1. Statistics (Dispersion)

Mean Deviation (MD): MD(x̄) =
∑|x_i - x̄|n

(Average of absolute distances from mean).
Variance (σ²): σ² =
∑x_i²n
- (x̄)²
(Mean of squares minus square of mean).
Shortcut Formula: σ² = E(X²) - [E(X)]²
Grouped Data (Step Deviation):
σ² = h² × [
∑f_id_i²N
- (
∑f_id_iN
)²]
Where d_i = (x_i - A)/h
Standard Deviation (σ): σ = √(Variance)
Coefficient of Variation (CV): CV =
σx̄
× 100
(Lower CV = More Stable).

Addition Rule: P(A ∪ B) = P(A) + P(B) - P(A ∩ B)

Complement Rule: P(A′) = 1 - P(A) (Used for "at least one")

Conditional: P(A|B) =

P(A ∩ B)P(B)

Multiplication: P(A ∩ B) = P(A) · P(B|A)

Total Probability: P(A) = ∑ P(E_i) · P(A|E_i)

Bayes’ Theorem:
P(E_i|A) =

P(E_i) · P(A|E_i)∑ P(E_j) · P(A|E_j)

Random Variable & Expectation:
E(X) = ∑ x_i P(x_i) (Mean)
Var(X) = E(X²) - [E(X)]²

3. Important Results

Change of Origin & Scale (Statistics)

Variance is independent of a change of origin (adding/subtracting a constant like x + a) but dependent on a change of scale (multiplying/dividing like ax).

If y_i = ax_i + b, then:
σ_y² = a²σ_x² and σ_y = |a|σ_x

Event Types (Probability)

Independent Events: The occurrence of one doesn't affect the other.
P(A ∩ B) = P(A) · P(B) and P(A|B) = P(A).
Mutually Exclusive (ME): Cannot happen at the same time.
A ∩ B = ∅ &implies; P(A ∩ B) = 0.
Exhaustive Events: Cover all possible outcomes in sample space S.
E₁ ∪ E₂ ∪ … ∪ E_n = S &implies; ∑ P(E_i) = 1.

LP Feasible Region

The intersection of all constraint half-planes forms the Feasible Region. The optimal solution (Maximum or Minimum of Z) always occurs at a corner point (vertex) of this feasible region.

Solving LP Graphically (Step-by-Step)

State the Objective: Identify Z = ax + by.
Plot Constraints: Treat inequalities as equations (lines).
Identify Feasible Region: Shade the common intersection area.
Identify Corner Points: Find coordinates of all vertices.
Evaluation: Calculate Z value at each corner point.
Optimization: Pick the vertex giving Max/Min Z value.

4. Conceptual Insights

Variance Intuition: Variance is the "average squared distance" from the mean. Squaring serves two purposes: it removes negative signs and it penalizes extreme outliers much more heavily than small deviations.
Bayes’ Theorem Logic: Think of it as "Reverse Probability." You already know the final result (A has occurred) and you want to trace backward to find the most likely original cause (E_i).
LP Corner Point Theorem: If the feasible region is bounded, both Max and Min absolutely exist. If unbounded, Max/Min may or may not exist (you must check the half-plane ax + by > Z_max).

5. Common Mistakes

Variance vs SD: Students constantly forget to square the Standard Deviation when the formula or question specifically asks for Variance (or forget to square-root when asked for SD).
Independent vs ME: They are completely different concepts! Independence is about probability formulas (P(A∩B) = P(A)P(B)); Mutually Exclusive is about set intersection (P(A∩B) = 0).
Bayes’ Denominator: Ensure the denominator sum completely accounts for ALL mutually exclusive and exhaustive paths to event A.
Inequality Signs in LP: Shading the wrong side of a line graph. Always test with the origin (0,0)—if it genuinely satisfies ax + by ≤ c, shade towards the origin.
Unbounded Region Trap: In an unbounded region, a maximum might not mathematically exist. You must verify if the half-plane ax + by > Z_max has any common points with the feasible region.

6. Example Applications

Statistics: Variance of Natural Numbers

Finding the variance of the first n natural numbers.

Insight: σ² =

n² - 112

.
(This is a standard direct formula that saves massive time on the IAT).

Probability: Bayes' Theorem

Drawing balls from two bags. Bag 1 (3R, 2W), Bag 2 (2R, 4W). One ball is drawn and is Red. Find the probability it came from Bag 1.

Insight: Use Bayes’ Theorem. Let E₁ = Bag 1, E₂ = Bag 2, A = Red. Calculate P(E₁|A) directly using the formula.

LP Optimization

Maximize Z = 3x + 4y subject to x + y ≤ 4, x, y ≥ 0.

Insight: Corner points of the polygon formed are (0,0), (4,0), (0,4). Evaluate Z at each: Z(0,0)=0, Z(4,0)=12, Z(0,4)=16. Max is 16 at (0,4).

7. IAT Exam Focus Points

The Gold Strategy: Which Probability Formula?

"Probability of A given B": Conditional Probability.
"At least/At most/Neither": Use Complement Rule P(A′) = 1 - P(A).
"A occurs but results from different causes": Total Probability Theorem.
"Tracing back to the cause" (e.g., Bag 1 given Red): Bayes' Theorem.
"Sequence of independent trials": Multiplication Rule / Binomial.

Bayes' Theorem: A massive staple for IISER IAT. Practice classical "Disease Testing" accuracy problems or "Urn/Ball" probability paths.
Variance/SD Properties: Tricky properties questions like "If every observation is multiplied by 2 and increased by 5, what is the new SD?" (Answer: New SD = 2 × Old SD. The +5 shift does nothing).
Independent Events: Using the shortcut P(A ∪ B) = 1 - P(A′)P(B′) for solving "at least one" problems rapidly.
LP Corner Points: Finding the optimal value for a bounded region. Be careful interpreting "Unbounded" regions where the maximum infinity might not mathematically exist within the domain.
Conditional Probability Interpretation: Word problems often try to trick you into swapping P(A|B) and P(B|A). Carefully identify which event is the actual "given" condition.

8. Practice Mock Test

Ready to test your knowledge?

Take a quick 15-question assessment specifically designed for Statistics, Probability & LP. Challenge yourself with IAT-level questions.

Start Practice Mock