Compute Multiple-Choice Elimination Odds

This calculator uses a simplified probability model that assumes questions are independent, guesses among remaining options are equally likely, and there's no negative marking. Real exams may have correlated questions (knowing one helps with related), trick distractors (designed to mislead partial knowledge), varying difficulty, test design differences, or scoring rules that differ from these assumptions. The expected value is an average over many hypothetical attempts—your actual score on any single exam may be higher or lower. Understanding this helps you see when calculator is useful and when real-world factors may affect actual performance.

The model assumes: (1) Each question is independent—answering one doesn't affect your chances on others. (2) Among remaining options after elimination, your guess is equally likely to be any of them. (3) There's no negative marking for wrong answers. (4) Your self-assessment of 'known' vs 'partial knowledge' questions is accurate. Real exams often violate these assumptions: questions may be correlated, trick answers may mislead, difficulty may vary, or negative marking may exist. Understanding assumptions helps you see when calculator is appropriate and when real-world factors may differ.

Yes, mathematically eliminating wrong options before guessing always improves your probability of getting a question correct. On a 4-option question, pure guessing gives 25% odds. Eliminate 1 option → 33% (1/3). Eliminate 2 options → 50% (1/2). However, you must actually eliminate wrong options—if you accidentally eliminate the correct answer, it hurts your chances. Understanding this helps you see why elimination improves odds and why accurate elimination is important.

This model doesn't account for question dependencies (e.g., if you know concept A, you probably know related concepts) or trick answers designed to mislead partially-informed students. Real exam results can differ significantly from this model. For example, trick distractors may look correct to students with partial knowledge, reducing actual probability below model predictions. Use this as a rough planning tool, not a precise prediction. Understanding this helps you see when calculator is appropriate and when real-world complexity may affect results.

Expected score is the average you would get if you took this exact quiz many times under the same conditions. On any single attempt, you might score above or below the expected value due to randomness. The probability distribution (if shown) gives you a sense of how much variation is possible. For example, if expected=11.25, you might score 8, 10, 12, or 14 on different attempts. Understanding expected value helps you see that it's an average, not a guarantee, and why distribution shows variability.

Identify which questions are 'pure guesses' and focus your study time on converting them to 'partial knowledge' or 'known'. Even a small improvement—like being able to eliminate one wrong option—significantly boosts your odds. For example, converting 5 pure guesses (25%) to partial knowledge with 2 eliminated (50%) increases expected score by 1.25 correct. The biggest payoff comes from solid preparation, not optimizing guessing strategies. Understanding this helps you see how to prioritize study and why studying is more important than guessing.

The distribution shows the probability of getting exactly 0, 1, 2, ... N questions correct, given your knowledge breakdown. It's computed using a Poisson-binomial distribution (dynamic programming) that accounts for different probabilities across question groups. For large quizzes (>150 questions), we show a simplified comparison instead due to computational limits. The distribution helps you understand score variability and see that expected value is an average, not a guarantee. Understanding distribution helps you see score range and understand risk.

The probability at least target is calculated from the full probability distribution (if ≤150 questions) using dynamic programming. It gives the chance of reaching target score or higher under the model's assumptions. However, real exams may differ due to question correlations, trick questions, varying difficulty, or test design. The calculation is mathematically correct for the model, but actual probability may differ in practice. Understanding this helps you see when calculation is useful and when real-world factors may affect actual probability.

The calculator uses a single value for 'options eliminated per partial' question, representing an average. If you can eliminate different numbers on different questions, use the average or most common number. For more precise analysis, you could run multiple calculations with different elimination levels and weight them by question counts. Understanding this helps you see how to handle varying elimination levels and why using average is reasonable for most cases.

No. This calculator assumes no negative marking—wrong answers don't reduce your score. If your exam has negative marking (e.g., -0.25 points for wrong answer), the calculator's expected score will be higher than your actual expected score. You would need to adjust calculations manually or use a different model that accounts for negative marking. Understanding this helps you see when calculator is appropriate and when negative marking affects results.