Home/Notes/Probability & Statistics
Engineering Exam Notes

Probability & Statistics Notes

Questions

~2 questions per paper

Difficulty

Medium

Importance

High yield for HPCL, BHEL, and NTPC technical sections

Overview

Probability and Statistics form the foundation of analytical decision-making in engineering disciplines, frequently appearing in PSU technical assessments. Mastering this topic allows candidates to efficiently handle data analysis, reliability modeling, and error distribution problems common in large-scale infrastructure projects.

Probability Distributions

Distributions model how data points or events are spread across a range, which is critical for reliability testing and quality control in PSUs. Understanding when to apply discrete versus continuous models is the key differentiator for high-scoring candidates.

  • Binomial: P(X=k) = nCk * p^k * q^(n-k)
  • Poisson: P(X=k) = (lambda^k * e^-lambda) / k!
  • Normal: Z = (X - mu) / sigma
  • Normal properties: Area under curve is 1, mean=median=mode
  • Poisson requirement: np must be constant as n approaches infinity

Central Tendency and Variance

These metrics describe the behavior of a dataset, providing the 'typical' value and the dispersion of data points. Examiners often frame simple numerical problems around calculating mean, median, and variance for ungrouped data sets.

  • Mean (Arithmetic): Sum of observations / n
  • Variance: E[X^2] - (E[X])^2
  • Standard Deviation: Square root of variance
  • Empirical relationship: Mode = 3(Median) - 2(Mean)
  • Coefficient of Variation: (SD / Mean) * 100

Correlation and Regression

Correlation measures the strength of the linear relationship between two variables, while regression allows for predictive modeling. These concepts are frequently tested through calculation-heavy MCQs that test your ability to derive lines of best fit.

  • Pearson correlation coefficient (r) ranges from -1 to 1
  • Regression line of Y on X: Y - Y_bar = r * (sigmaY / sigmaX) * (X - X_bar)
  • r = 0 implies no linear correlation
  • Regression coefficients are independent of change of origin
  • Slope of regression line is b = r * (sigmaY / sigmaX)

Formula Sheet

Variance = Sigma(x - mu)^2 / N

P(A U B) = P(A) + P(B) - P(A n B)

Correlation r = Cov(X,Y) / (sigmaX * sigmaY)

Exam Tip

Always verify if a probability problem specifies 'with replacement' or 'without replacement', as this determines whether you use Binomial or Hypergeometric logic.

Common Mistakes

  • Confusing the standard deviation formula for populations versus samples (dividing by n instead of n-1).
  • Applying the Binomial distribution formula when the number of trials 'n' is very large instead of switching to Poisson approximation.
  • Neglecting to square the standard deviation when asked for the variance in a calculation problem.

More Revision Notes

Ready to test yourself?

Play topic-wise Probability & Statistics questions in Aspirant Arcade — gamified MCQ practice.

Download Free