π Lecture 34 β FACTORIALS, PERMUTATIONS AND COMBINATIONS
π Overview: This lecture covers the review of exponential smoothing from Lecture 33, introduces the concept of factorials, and explains permutations and combinations. Understanding these concepts is essential for probability calculations, counting problems, and statistical analysis.
ποΈ Topics Covered
The lecture begins with a review of exponential smoothing, including the forecast formula, alpha values, Mean Square Error (MSE), and the Excel Exponential Smoothing Tool. It then introduces factorials with definitions and examples, followed by the fundamental counting principle (multiplication rule). Finally, permutations are covered in detail, including the formula nPr, permutations of n objects taken r at a time, permutations with repeated objects, and the Excel PERMUT function.
π Lecture Summary
Review Lecture 33 - Exponential Smoothing
The lecture begins by revisiting exponential smoothing from Lecture 33. Using Ξ± = 0.3, the forecast for week 3 is calculated as: Forecast week 3 = week 2 forecast + Ξ± Γ (week 2 actual sale β week 2 forecast) = 4500 β 0.3 Γ 500 = 4350. The overestimate is reduced by 30% of the error margin of 500. This method is called Exponential Smoothing and Ξ± is the Smoothing Constant.
The general rule for obtaining a forecast is: Let A = Actual and F = Forecast. Then: F(t) = F(t-1) + Ξ±(A(t-1) β F(t-1)) = Ξ±A(t-1) + (1-Ξ±)F(t-1). By substituting F(t-1) = Ξ±A(t-2) + (1-Ξ±)F(t-2), we get F3 = Ξ±[A(t-1) + (1-Ξ±)A(t-2)] + (1-Ξ±)Β²F(t-2). Continuing this substitution yields: F(t) = Ξ±[A(t-1) + (1-Ξ±)A(t-2) + (1-Ξ±)Β²A(t-3)] + (1-Ξ±)F(t-3).
π‘ Why this matters: Exponential smoothing gives more weight to recent observations and less weight to older ones, making it a powerful forecasting tool.
The accepted criterion for evaluating forecasts is Mean Square Error (MSE). MSE is found by squaring all errors, including the present one, and dividing by the number of periods included. A sign of a good forecast is when MSE stabilizes. Generally, alpha between 0.1 and 0.3 performs best.
Excel Exponential Smoothing Tool
The Exponential Smoothing Tool in Excel can be used to perform this analysis. The dialog box includes:
- Input Range: The cell reference for the data (a single column or row with 4+ cells).
- Damping factor: The exponential smoothing constant (default is 0.3). Values of 0.2 to 0.3 are reasonable smoothing constants, indicating the current forecast should be adjusted 20-30% for error in the prior forecast. Larger constants yield faster response but can produce erratic projections; smaller constants can result in long lags.
- Labels: Select if the first row/column contains labels.
- Output Range: Reference for the upper-left cell of the output table.
- Chart Output: Select to generate an embedded chart for actual and forecast values.
- Standard Errors: Select to include a column with standard error values.
Factorials
Factorials are defined for natural numbers (1, 2, 3,...). For example, Five Factorial is: 5! = 5Β·4Β·3Β·2Β·1 = 1Β·2Β·3Β·4Β·5 = 120. Similarly, Ten Factorial = 1Β·2Β·3Β·4Β·5Β·6Β·7Β·8Β·9Β·10 = 3,628,800.
In general: n! = n(n-1)(n-2)...3Β·2Β·1 or n! = n(n-1)!.
π Definition β Factorial: The product of all positive integers from 1 to n, denoted by n!.
π Formula: n! = n(n-1)(n-2)! or n! = n(n-1)! Example 1: 8!/5! = 8Β·7Β·6Β·5!/5! = 8Β·7Β·6 = 336 Example 2: 12!/9! = 12Β·11Β·10Β·9!/9! = 12Β·11Β·10 = 1320 Example 3: 10!8!/9!5! = (10Β·9!)(8Β·7Β·6Β·5!)/(9!5!) = 10Β·8Β·7Β·6 = 3360
Ways (Fundamental Counting Principle)
If operation A can be performed in m ways and operation B in n ways, then the two operations can be performed together in mΒ·n ways.
Example: A coin can be tossed in 2 ways. A die can be thrown in 6 ways. A coin and a die together can be thrown in 2Β·6 = 12 ways.
Permutations
An arrangement of all or some of a set of objects in a definite order is called a permutation.
π Definition β Permutation (nPr): The number of ways to arrange r objects selected from n distinct objects where order matters.
π Formula: nPr = n!/(n-r)!
Example 1: There are 4 objects A, B, C, D. Permutations of 2 objects A & B: AB, BA. Permutations of 3 objects A, B, C: ABC, ACB, BCA, BAC, CAB, CBA.
Example 2: Number of permutations of 3 objects taken 2 at a time = 3P2 = 3!/(3-2)! = 3Β·2 = 6 = AB, BA, AC, CA, BC, CB.
Example 3 - Movie Problem: You have 6 movies showing at 2:00, 4:00, and 6:00 pm. How many ways can you watch 3 different movies? P(6, 3) = 6!/(6-3)! = 6!/3! = 6Β·5Β·4Β·3!/3! = 120 ways.
Example 4 - Lottery: Suppose there are 100 numbers (00 to 99), you choose 5 numbers in specific order without replacement. Number of ways: P(100,5) = 100!/(100-5)! = 9,034,502,400 ways. Your chance of winning is 1 in 9,034,502,400.
Example 5 - Spy Problem: A spy has 8 cards, needs 5 in correct order. Number of permutations: P(8,5) = 8!/3! = 6720. Chance of correct sequence: 1/6720 = 0.00015 = 0.015%.
Permutations of n Objects with Repetitions
Number of n permutations of n different objects taken n at a time: nPn = n!/(n-n)! = n!/0! = n!/1 = n!
For permutations of n objects where nβ are alike of one kind, nβ are alike of another kind, and nβ are alike:
π Formula: n!/(nβ!nβ!...nβ!)
Example: How many permutations can be formed from the word STATISTICS? S=3, A=1, T=3, I=2, C=1 nPr = 10!/(3!1!3!2!1!) = 10Β·9Β·8Β·7Β·6Β·5Β·4Β·3!/(3!3!2!) = 50,400 permutations.
Excel PERMUT Function
The Excel PERMUT function returns the number of permutations for a given number of objects selected from number objects.
Syntax: PERMUT(number, number_chosen)
- Number: An integer describing the number of objects
- Number_chosen: An integer describing the number of objects in each permutation
Equation: P(n, r) = n!/(n-r)!
Example: For a lottery with 3 numbers from 0-99 (100 possibilities): PERMUT(100, 3) = 100!/(100-3)! = 970,200 possible permutations.
β Key Takeaways
In exponential smoothing, the forecast is updated by adding Ξ± multiplied by the error (difference between actual and forecast) to the previous forecast, with Ξ± (smoothing constant) typically between 0.1 and 0.3. Factorials (n!) are products of all positive integers from 1 to n, and n! = n(n-1)!. The fundamental counting principle states that if operation A has m ways and B has n ways, then both together have mΒ·n ways. Permutations (nPr) are arrangements where order matters, calculated as n!/(n-r)!, and permutations with repeated objects use the formula n!/(nβ!nβ!...nβ!). The Excel PERMUT function can calculate permutations automatically.
π§ Quick Revision Questions
- Calculate the exponential smoothing forecast for week 3 if week 2 forecast is 4500, actual sale is 5000, and Ξ± = 0.3.
- What is the value of 5! and how is it calculated?
- If you have 8 books and want to arrange 3 on a shelf, how many different arrangements are possible?
- How many permutations can be formed from the word "MISSISSIPPI"?
- In Excel, which function calculates permutations, and what is its formula?
π Lecture 35 β COMBINATIONS ELEMENTARY PROBABILITY PART 1
π Overview: This lecture introduces combinations as arrangements where order does not matter, contrasting them with permutations through clear examples. It then transitions into elementary probability, covering foundational concepts like sample space, events, and the key rules (OR rule and AND rule) for calculating probabilities, with applications to business decision-making.
ποΈ Topics Covered
The lecture begins by reviewing combinations, their formula, and the critical difference from permutations, illustrated with examples like team selection and ball picking. It then covers important combinatorial results and their application in binomial expansion. The second half introduces fundamental probability concepts: sample space, events, theoretical and empirical probability, the OR rule for mutually exclusive events, the AND rule for joint events, and the concept of exhaustive events, all demonstrated with practical business and manufacturing examples.
π Lecture Summary
COMBINATIONS
Combinations are arrangements of objects where the order in which they are arranged does not matter. The number of combinations of n objects taken r at a time is given by a specific formula.
π Definition β Combination: An arrangement of objects without caring for the order in which they are arranged.
π Formula: $nCr = \frac{n!}{r!(n-r)!}$ where n is the total number of objects and r is the number of objects selected at a time.
π Example: Number of combinations of 3 different objects A, B, C taken two at a time = $3!/2!(3-2)! = 6/2 = 3$. These combinations are: AB, AC, and BC.
COMBINATIONS EXAMPLES
Example 3: In how many ways a team of 11 players be chosen from a total of 15 players?
- $n=15, r=11$
- $15C11 = \frac{15!}{11!(15-11)!} = \frac{15!}{11!(4!)} = \frac{15.14.13.12.11!}{11!(4.3.2.1)} = \frac{15.14.13.12}{24} = 1365$ Ways
Example 4: There are 5 white balls and 4 black balls. In how many ways can we select 3 white and 2 black balls?
- $5C3 \times 4C2 = \frac{5!}{3!(5-3)!} \times \frac{4!}{2!(4-2)!} = \frac{5!}{3!2!} \times \frac{4!}{2!2!} = 10 \times 6 = 60$ ways
Example 5: If a committee of 3 people is to be selected from among 5 married couples so that the committee does not include two people who are married to each other, how many such committees are possible?
- Solution:
- Total ways of picking 3 out of 10 people = $10C3 = 120$
- Ways with a married couple: A set can have any of the five married couples (5 ways) x The third person can be any one of the remaining eight (8 ways) = $5 \times 8 = 40$
- Number of combinations without any married couples = $120 - 40 = 80$
RESULTS OF SOME COMBINATIONS
These are important for simplifying calculations, especially in Binomial Expansion.
- $nC0 = nCn = 1$ (e.g., $4C0 = 4C4 = 1$)
- $nC1 = nCn-1 = n$ (e.g., $4C1 = 4C3 = 4$)
- $nCr = nCn-r$ (e.g., $5C2 = 5C3$)
BINOMIAL EXPANSION
An expression consisting of two terms joined by + or β sign is called a Binomial Expression. The right-hand side of expanded binomials are called Binomial Expansions.
π Definition β Binomial Expansion: The expansion of a binomial expression (like $(x+y)^n$) where the coefficients can be written in combinatorial notation.
π Example: $(x+y)^5 = 5C0.x^5 + 5C1x^4y + 5C2x^3y^2 + 5C3x^2y^3 + 5C4xy^4 + 5C5y^5$ which simplifies to $x^5 + 5x^4y + 10x^3y^2 + 10x^2y^3 + 5xy^4 + y^5$.
Calculation of Binomial Expansion Coefficients:
- Coefficient of first and last term is always 1.
- Coefficient of any other term = (coefficient of previous term) * (power of x from previous term) / number of that term.
- π Example:
- First term = $x^5$, Last term = $y^5$
- Second coefficient = $5/1 = 5$
- Third coefficient = $5*4/2 = 10$
- Fourth coefficient = $10*3/3 = 10$
- Fifth coefficient = $10*2/4 = 5$
PROJECT DEVELOPMENT MANAGERβS PROBLEM
This problem introduces a business context where probability is needed.
- A toy manufacturer has two choices: abandon or risk new development.
- There is a 40% chance of a TV series (leading to 12,000 unit sales).
- Without the series, demand may be 2,000 units.
- A rival has a 50% chance of bringing a similar toy, which would reduce sales to 8,000 units.
- The question is how to tie these probabilities to financial results.
Sample space, Event
The set of collection of all possible outcomes of an experiment is called the sample space. Each possible outcome of an experiment is called an event. An event is a subset of the sample space.
π Example: All six faces of a die make a sample space. Rolling the dice and getting the number 1 is an event.
Probability
Probability is the numerical measure of the chance that an uncertain event will occur. The probability that event A will occur is usually denoted by p(A).
π Definition β Probability: For any event A, $0 \le p(A) \le 1$. $p(A) = 1$ means certain, $p(A) = 0$ means impossible.
PROBABILITY EXAMPLE 1
How can we make assessment of chances? Look at a simple example: A worker out of 600 gets a prize by lottery. What is the chance of any one individual, say Rashid, being selected?
- Solution: Chance = $1/600$. This is an aβ priori method of finding probability as we can assess the probability before the event occurred.
- $p(\text{Rashid is selected}) = 1/600$
PROBABILITY EXAMPLE 2
When all outcomes are equally likely, aβ priori probability is defined as:
- $p(\text{event}) = \frac{\text{Number of ways that event can occur}}{\text{Total number of possible outcomes}}$
- π Example: If out of 600 persons 250 are women, then $p(\text{woman}) = 250/600$.
PROBABILITY - EMPIRICAL APPROACH
In many situations, there is no prior knowledge to calculate probabilities. This is the experimental or empirical approach.
π Formula: $p(\text{event}) = \frac{\text{Number of times event occurs}}{\text{Total number of experiments}}$
π‘ Why this matters: Larger the number of experiments, more accurate the estimate. Experimental probability approaches theoretical probability as the number of experiments becomes very large.
OR RULE
This rule calculates the probability of either event A or event B (or both) happening.
Condition for Or Rule: A and B must be mutually exclusive (cannot happen at the same time).
π Formula: $p(A \text{ or } B) = p(A) + p(B)$
π Example: If a dice is thrown, what is the chance of getting an even number or a number divisible by 3?
- $p(\text{even}) = 3/6$
- $p(\text{div by 3}) = 2/6$
- $p(\text{even or div by 3}) = 3/6 + 2/6 = 5/6$... but the number 6 is not mutually exclusive (it is both even and divisible by 3).
- Correct answer: The events {2,4,6} and {3,6} share the element {6}. The correct union is {2,3,4,6} = $4/6$.
AND RULE
This rule calculates the probability of both event A and event B happening.
π Formula: $p(A \text{ and } B) = p(A) \times p(B)$
π Example: In a factory, 40% workforce are women. 25% females are in management grade. 30% males are in management grade. What is the probability that a worker selected is a woman from management grade?
- Solution: Assume total workforce = 100.
- $p(\text{woman}) = 0.4$
- $p(\text{management} | \text{woman}) = 0.25$
- $p(\text{woman and Management grade}) = p(\text{woman}) \times p(\text{management} | \text{woman}) = 0.4 \times 0.25 = 0.1$ or 10%
SET OF MUTUALLY EXCLUSIVE EVENTS
To cover all possibilities between mutually exclusive events, add up all their probabilities. The probabilities of all these events together add up to 1.
π Formula: $p(A) + p(B) + p(C) + ... + p(N) = 1$
EXHAUSTIVE EVENTS
If A happens or A does not happen, then A and βnot Aβ are Exhaustive Events.
π Formula: $p(A \text{ happens}) + p(A \text{ does not happen}) = 1$
π Example: $p(\text{you pass}) = 0.9$, then $p(\text{you fail}) = 1 - 0.9 = 0.1$.
EXAMPLE1 - EXHAUSTIVE EVENTS
A production line uses 3 machines. The chance that the 1st machine breaks down in any week is $1/10$, the 2nd is $1/20$, and the 3rd is $1/40$. What is the chance that at least one machine breaks down in any week?
- Solution: $p(\text{at least one not working}) + p(\text{all three working}) = 1$
- $p(\text{at least one not working}) = 1 - p(\text{all three working})$
- $p(\text{all three working}) = p(1\text{st working}) \times p(2\text{nd working}) \times p(3\text{rd working})$
- $p(1\text{st working}) = 1 - 1/10 = 9/10$
- $p(2\text{nd working}) = 19/20$
- $p(3\text{rd working}) = 39/40$
- $p(\text{all working}) = 9/10 \times 19/20 \times 39/40 = 6669/8000$
- $p(\text{at least 1 not working}) = 1 - 6669/8000 = \mathbf{1331/8000}$
APPLICATION OF RULES
A firm has the following rules: When a worker comes late there is a $1/4$ chance he is caught. First time he is given a warning. Second time he is dismissed. What is the probability that a worker who is late three times is not dismissed?
-
Solution: Using the AND Rule to find probabilities of each branch:
- $p(\text{1C and 2C})$ (Dismissed) = $1/4 \times 1/4 = 1/16 = 4/64$
- $p(\text{1C and 2NC and 3C})$ (Dismissed) = $1/4 \times 3/4 \times 1/4 = 3/64$
- $p(\text{1NC and 2C and 3C})$ (Dismissed) = $3/4 \times 1/4 \times 1/4 = 3/64$
- $p(\text{1C and 2NC and 3NC})$ (Not Dismissed) = $1/4 \times 3/4 \times 3/4 = 9/64$
- $p(\text{1NC and 2C and 3NC})$ (Not Dismissed) = $3/4 \times 1/4 \times 3/4 = 9/64$
- $p(\text{1NC and 2NC and 3C})$ (Not Dismissed) = $3/4 \times 3/4 \times 1/4 = 9/64$
- $p(\text{1NC and 2NC and 3NC})$ (Not Dismissed) = $3/4 \times 3/4 \times 3/4 = 27/64$
-
Using the OR Rule for all "Not Dismissed" scenarios:
- $p(\text{not dismissed}) = 9/64 + 9/64 + 9/64 + 27/64 = \mathbf{54/64}$
-
$p(\text{caught})$ using OR Rule:
- $p(\text{caught}) = 4/64 + 3/64 + 3/64 = \mathbf{10/64}$
-
Using Exhaustive events to verify:
- $p(\text{not caught}) = 1 - p(\text{caught}) = 1 - 10/64 = \mathbf{54/64}$ (Matches!)
β Key Takeaways
The fundamental distinction between combinations (order doesn't matter, formula $nCr = n!/(r!(n-r)!)$) and permutations (order matters) is critical for solving selection problems. The lecture establishes that probability is always between 0 and 1, and can be found theoretically (a priori) or empirically. The two most important operational rules are the OR rule ($p(A \text{ or } B) = p(A) + p(B)$) which requires events to be mutually exclusive, and the AND rule ($p(A \text{ and } B) = p(A) \times p(B)$). Finally, the concept of exhaustive events ($p(A) + p(\text{not }A) = 1$) provides a powerful tool for simplifying complex probability problems, as demonstrated in the machine breakdown and worker dismissal examples.
π§ Quick Revision Questions
- A committee of 4 is to be formed from 8 people. How many different committees are possible? How does this differ from a permutation problem?
- What is the condition for using the OR rule $p(A \text{ or } B) = p(A) + p(B)$? What happens if this condition is not met?
- If the probability of rain on a given day is 0.3, what is the probability it does not rain? Which rule of probability does this illustrate?
- A bag contains 3 red and 5 blue marbles. What is the probability of drawing a red marble and then, without replacement, a blue marble? Which probability rule is used here?
- Explain the difference between a priori probability (like drawing a card from a deck) and empirical probability (like finding the chance a machine will fail based on past data).
π Lecture 36 β Elementary Probability Part 2
π Overview: This lecture continues the study of elementary probability concepts introduced in Lecture 35. It reviews fundamental probability rules including the OR Rule for mutually exclusive events and the AND Rule for simultaneous events, and explores how to handle sets of mutually exclusive and exhaustive events through worked examples.
ποΈ Topics Covered
The lecture begins with a review of basic probability concepts and the PERMUT function. It then revisits the OR Rule with an example involving dice outcomes, followed by a review of the AND Rule with a workforce management example. The final section covers sets of mutually exclusive (exhaustive) events, using a production line example with three machines to demonstrate how to calculate probabilities of complementary events.
π Lecture Summary
PROBABILITY CONCEPTS REVIEW
Probability is about making assessments of chances. The simplest example was the probability of Rashid getting the lottery when he was one of 600 people. The probability of the event was 1/600.
PERMUT EXAMPLE
The function PERMUT can be used for calculations of permutations. This function was introduced in Lecture 35.
OR RULE REVIEW
When two events are mutually exclusive, the probability of either one of those occurring is the sum of individual probabilities. This is the OR Rule. This is a very extensively used rule. A and B must be mutually exclusive.
π Definition β OR Rule: The probability that event A or event B occurs (when A and B are mutually exclusive) equals the sum of their individual probabilities.
π Formula: p(A or B) = p(A) + p(B)
π Example: If a dice is thrown, what is the chance of getting an odd number or a number divisible by three?
- P(odd) = 3/6
- P(divisible by 3) = 2/6
- P(odd or divisible by 3) = 3/6 + 2/6 = 5/6
- Note: The number 6 is not mutually exclusive (needs checking), hence correct answer = 5/6.
AND RULE REVIEW
The AND Rule requires that the events occur simultaneously.
π Example: 60% of the workforce are men. 25% of females are in management grade. 30% of males are in management grade. What is the probability that a worker selected is a man from management grade?
- Total workforce = 100
- p(man) = 0.6
- p(management) = 0.3
- p(man & management grade) = p(man) Γ p(management)
- = 0.6 Γ 0.3 = 0.18 or 18%
SET OF MUTUALLY EXCLUSIVE EVENTS REVIEW
Mutually exclusive events between them cover all possibilities. These are also called exhaustive events. Probabilities of all these events together add up to 1. Exhaustive events are events that happen or do not happen.
π Formula: p(it does not rain) = 1 - p(it rains)
- Example: p(it rains) = 0.9, then p(it does not rain) = 1 - 0.9 = 0.1
π Example β Three Machines Problem: A production line uses 3 machines.
- Chance that 1st machine breaks down in any week = 1/10
- Chance for 2nd machine = 1/20
- Chance of 3rd machine = 1/40
- Question: What is the chance that at least one machine breaks down in any week?
Solution Approach:
- The probability that one or two or three machines are not working (at least one not working) AND that all three are working are exhaustive events that add up to 1.
- P(at least one not working) + p(all three working) = 1
Step 1 β Calculate probability of each machine working:
- p(1st working) = 1 - p(1st not working) = 1 - 1/10 = 9/10
- p(2nd working) = 1 - 1/20 = 19/20
- p(3rd working) = 1 - 1/40 = 39/40
Step 2 β Calculate probability all three are working (AND Rule):
- p(all three working) = p(1st working) Γ p(2nd working) Γ p(3rd working)
- = 9/10 Γ 19/20 Γ 39/40 = 6669/8000
Step 3 β Calculate probability at least one is not working:
- P(at least one not working) = 1 - p(all three working)
- = 1 - 6669/8000 = 1331/8000
π‘ Why this matters: This example demonstrates how to handle complex probability problems by first finding the probability of the complementary event (all machines working) and then subtracting from 1.
β Key Takeaways
The OR Rule is used for mutually exclusive events where the probability of either event occurring is the sum of individual probabilities, as demonstrated with the dice example. The AND Rule applies when events must occur simultaneously, requiring multiplication of their probabilities. The most critical concept for this lecture is the method of solving "at least one" problems by first finding the probability of the complementary event (all events working/occurring) and subtracting from 1, as shown in the three machines example. Understanding that exhaustive events (like a machine working or not working) always have probabilities that sum to 1 is fundamental to probability calculations. Finally, the PERMUT function offers a computational tool for permutation calculations.
π§ Quick Revision Questions
- What is the OR Rule formula, and what condition must be met for its application?
- In the dice example (odd number OR divisible by three), what was the probability and why was 5/6 the correct answer?
- In the workforce example, what was the probability that a randomly selected worker was a man from management grade, and how was it computed?
- In the three machines problem, what was the probability that all three machines are working during a week?
- What was the probability that at least one machine breaks down in any week for the three machines, and what is the key insight for solving such problems?
π Lecture 37 β Patterns of Probability: Binomial, Poisson and Normal Distributions Part 1
π Overview: This lecture reviews elementary probability concepts from previous sessions and introduces the foundation for understanding probability patterns, specifically the Binomial distribution. It covers multiple worked examples including conditional probability, expected value calculations, and sets up the conditions for binomial probability situations.
ποΈ Topics Covered
The lecture begins with a review of Module 7 structure and continues with detailed worked examples: Example 1 examines a worker's probability of being dismissed for lateness with multiple sequential events; Example 2 covers competing firms bidding for contracts with mutually exclusive events; Example 3 demonstrates conditional probability using workforce demographics; Example 4 calculates expected value from pie sales data; and finally, the lecture introduces binomial probability conditions and cumulative binomial probability tables using a biscuit factory production problem.
π Lecture Summary
Review Lecture 36
The lecture reviews elementary probability concepts covered in Handouts 35 and 36. Material concerning the employee who is warned for being late and dismissed if late twice is discussed in detail.
EXAMPLE 1
A firm has rules: when a worker comes late, there is a ΒΌ chance of being caught. First time: warning. Second time: dismissal.
What is the probability that a worker late three times is not dismissed?
Solution: Options are developed through sequential stages:
- First time: Caught (1C) or Not Caught (1NC)
- Second time: Caught (2C) or Not Caught (2NC)
- Third time: Caught (3C) or Not Caught (3NC)
Combinations up to 2nd stage: 1C&2C, 1C&2NC, 1NC&2C, 1NC&2NC
Combinations up to 3rd stage: 1C&2C&3C, 1C&2NC&3C, 1C&2NC&3NC, 1NC&2C&3C, 1NC&2C&3NC, 1NC&2NC&3C, 1NC&2NC&3NC
The probability of being caught = ΒΌ, probability of not being caught = 1 - ΒΌ = ΒΎ
Probabilities:
- 1C & 2C = 1/4 Γ 1/4 = 1/16 = 4/64 (Dismissed 1)
- 1C & 2NC & 3C = 1/4 Γ 3/4 Γ 1/4 = 3/64 (Dismissed 2)
- 1C & 2NC & 3NC = 1/4 Γ 3/4 Γ 3/4 = 9/64 (Not dismissed 1)
- 1NC & 2C & 3C = 3/4 Γ 1/4 Γ 1/4 = 3/64 (Dismissed 3)
- 1NC & 2C & 3NC = 3/4 Γ 1/4 Γ 3/4 = 9/64 (Not dismissed 2)
- 1NC & 2NC & 3C = 3/4 Γ 3/4 Γ 1/4 = 9/64 (Not dismissed 3)
- 1NC & 2NC & 3NC = 3/4 Γ 3/4 Γ 3/4 = 27/64 (Not dismissed 4)
p(caught): These are mutually exclusive events. All dismissal situations considered: p(caught) = p(dismissed 1) + p(dismissed 2) + p(dismissed 3) = 4/64 + 3/64 + 3/64 = 10/64
p(not caught): As an exhaustive event: p(not caught) = 1 - p(caught) = 1 - 10/64 = 54/64
EXAMPLE 2
Two firms compete for contracts. A has probability ΒΎ of obtaining one contract, B has probability ΒΌ.
What is the probability that when they bid for two contracts, firm A will obtain either the first or second contract?
Solution: Wrong approach: P(A gets first or A gets second) = ΒΎ + ΒΎ = 6/4 β Probability cannot exceed 1! This ignored the mutually exclusive restriction.
Correct approach: We want probability that A gains the first or second or both. We are not interested in B getting both contracts. p(B gets first) Γ p(B gets both) = ΒΌ Γ ΒΌ = 1/16 p(A gets one or both) = 1 - 1/16 = 15/16
Alternative Method: Split "A gets first or second or both" into 3 mutually exclusive parts:
- A gets first but not second = ΒΎ Γ ΒΌ = 3/16
- A does not get first but gets second = ΒΌ Γ ΒΎ = 3/16
- A gets both = ΒΎ Γ ΒΎ = 9/16
P(A gets first or second or both) = 3/16 + 3/16 + 9/16 = 15/16 β
EXAMPLE 3
In a factory, 40% workforce is female. 25% of females belong to management cadre. 30% of males are from management cadre.
If a management grade worker is selected, what is the probability that it is a female?
Solution: First, draw up a table:
| Male | Female | Total | |
|---|---|---|---|
| Management | ? | ? | ? |
| Non-Management | ? | ? | ? |
| Total | ? | 40 | 100 |
Calculations:
- Total male = 100 - 40 = 60
- Management female = 0.25 Γ 40 = 10
- Non-Management female = 40 - 10 = 30
- Management male = 0.3 Γ 60 = 18
- Non-Management male = 60 - 18 = 42
- Management total = 18 + 10 = 28
- Non-Management total = 42 + 30 = 72
Final table:
| Male | Female | Total | |
|---|---|---|---|
| Management | 18 | 10 | 28 |
| Non-Management | 42 | 30 | 72 |
| Total | 60 | 40 | 100 |
p(management grade worker is female) = 10/28
π‘ Why this matters: This is a classic conditional probability problem β the condition is "given management grade," so the denominator is the management total (28), not the overall total (100).
EXAMPLE 4
A pie vendor collected data over sale of pies:
| No. Pies sold | Income per pie (X) Rs. | % Days (f) | fX Rs. |
|---|---|---|---|
| 40 | 35 Γ 40 = 1400 | 20 | 28000 |
| 50 | 1750 | 20 | 35000 |
| 60 | 2100 | 30 | 63000 |
| 70 | 2450 | 20 | 49000 |
| 80 | 2800 | 10 | 28000 |
| Total | 100 | 203000 |
Mean/day = 203000/100 = 2030 Rs.
Selling price per pie was Rs. 35. The mean sale per day is calculated by multiplying income from each quantity by % days (which serves as probability), summing, and dividing by number of days.
EXPECTED VALUE
π Formula: EMV = Ξ£ (probability of outcome Γ financial result of outcome)
Example: In an insurance company:
- 80% of policies have no claim = 0 Rs.
- 15% of policies have claim = 5000 Rs.
- 5% of policies have claim = 50000 Rs.
What is the expected value of claim per policy?
EMV = 0.8 Γ 0 + 0.15 Γ 5000 + 0.05 Γ 50000 = 0 + 750 + 2500 = 3250 Rs.
π‘ Why this matters: This expected monetary value represents the average claim amount per policy that the insurance company must anticipate β critical for setting premium prices.
TYPICAL PRODUCTION PROBLEM
In a biscuit factory, the packing machine breaks 1 biscuit out of twenty (p = 1/20 = 0.05).
What proportion of boxes will contain more than 3 broken biscuits?
This is a typical Binomial probability situation! The individual biscuit is either broken or not β two possible outcomes.
Conditions for Binomial Situation
- Either/or situation (two possible outcomes)
- Number of trials (n) is known and fixed
- Probability of success on each trial (p) is known and fixed
CUMULATIVE BINOMIAL PROBABILITIES
The Cumulative Probability table gives the probability of r or more successes in n trials, with probability p of success in one trial.
In the table:
- Total number of trials n = 1 to 10
- Number of successes r = 1 to 10
- Probability p = 0.05, 0.1, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5
β Key Takeaways
- For sequential probability problems with dismissal rules, systematically list all possible combinations of outcomes (caught/not caught) across each time period, calculate each path's probability by multiplying along branches, and sum probabilities for the condition of interest β remembering that once dismissal occurs, the process stops.
- When calculating "A or B" probabilities, ensure events are mutually exclusive; otherwise, simply adding probabilities can produce values exceeding 1, which is impossible β the complement method (1 minus probability of the undesired event) often provides a cleaner solution.
- In conditional probability problems (e.g., "given management grade"), the denominator is always the condition subgroup total, not the overall population total β create a contingency table to organize the given and calculated values systematically.
- Expected Monetary Value (EMV) is calculated by multiplying each outcome's probability by its financial result and summing all products β this represents the long-run average outcome per trial or policy.
- A Binomial probability situation requires exactly three conditions: a binary either/or outcome, a fixed number of trials (n), and a constant probability of success (p) for each trial β cumulative binomial probability tables provide quick answers for "r or more successes" problems.
π§ Quick Revision Questions
- In Example 1, why is the combination 1C&2C terminated at the second stage, and how does this affect the probability calculation for the third late occurrence?
- In Example 2, why does simply adding ΒΎ + ΒΎ = 6/4 give an incorrect answer, and what is the correct complement-based approach?
- In Example 3, when calculating the probability that a management grade worker is female, why is the denominator 28 rather than 100?
- Using the EMV formula, if an insurance company has 70% no claims, 20% claims of Rs. 3000, and 10% claims of Rs. 20000, what is the expected value of claim per policy?
- What are the three essential conditions that must be satisfied for a situation to qualify as a Binomial probability problem?
π Lecture 38 β Patterns of Probability: Binomial, Poisson and Normal Distributions PART 2
π Overview: This lecture continues the study of probability distributions, focusing on cumulative binomial probabilities and advanced binomial calculations. It also introduces the negative binomial distribution and Excel functions like BINOMDIST, NEGBINOMDIST, and CRITBINOM, which are essential for solving real-world probability problems in quality control, medicine, and communications.
ποΈ Topics Covered
The lecture covers cumulative binomial probabilities and their calculation using tables and formulas. It explains the BINOMDIST function in Excel with syntax and examples for problems like medical acceptance rates, weather probability, and transmission errors. The negative binomial distribution is introduced with properties and formula, along with the NEGBINOMDIST function for scenarios like finding qualified candidates. Finally, the CRITBINOM function is explained for quality assurance applications.
π Lecture Summary
Cumulative Binomial Probabilities
Cumulative binomial probability refers to the probability that the binomial random variable falls within a specified range, such as being greater than or equal to a stated lower limit and less than or equal to a stated upper limit. For example, the probability of obtaining 45 or fewer heads in 100 coin tosses is the sum of all individual binomial probabilities from 0 to 45 successes. This is not a single probability but the sum of multiple binomial probabilities. π‘ Why this matters: Cumulative probabilities allow us to answer "at most" or "at least" questions, which are common in real-world decision making.
π Definition β Cumulative Binomial Probability: The probability that the binomial random variable falls within a specified range, calculated as the sum of individual binomial probabilities for all values in that range.
π Formula: ( P(x \leq c) = \sum_{x=0}^{c} \binom{n}{x} P^x (1-P)^{n-x} ) β The sum of binomial probabilities from 0 up to c successes.
π Example: The probability that a student is accepted to a prestigious college is 0.3. If 5 students from the same school apply, what is the probability that at most 2 are accepted?
- n = 5, p = 0.3, we need P(x β€ 2)
- Compute: P(x=0) + P(x=1) + P(x=2) using the binomial formula
- Result: b(x β€ 2; 5, 0.3) = 0.837
BINOMDIST
BINOMDIST returns the individual term binomial distribution probability. Use BINOMDIST in problems with a fixed number of tests or trials, when the outcomes of any trial are only success or failure, when trials are independent, and when the probability of success is constant throughout the experiment. For example, BINOMDIST can calculate the probability that two of the next three babies born are male.
π Definition β BINOMDIST: An Excel function that returns the binomial distribution probability, either as the probability mass function (exact number of successes) or cumulative distribution function (at most number_s successes).
π Syntax: BINOMDIST(number_s, trials, probability_s, cumulative)
- number_s: number of successes in trials
- trials: number of independent trials
- probability_s: probability of success on each trial
- cumulative: TRUE returns cumulative distribution function (at most number_s successes); FALSE returns probability mass function (exact number_s successes)
π Example 1: The probability of wet days is 60%. What is the probability of 5 or more wet days in 7 days?
- p(wet) = 0.6 is beyond table values (max 0.5)
- Convert to p(dry) = 1 - 0.6 = 0.4
- P(5 or more wet days) = P(2 or less dry days)
- Convert using complement: P(2 or less dry days) = 1 - P(3 or more dry days)
- Using BINOMDIST with n=7, r=3, p=0.4, cumulative=TRUE
- Answer: 0.4199
π Example 2: In a transmission where 8 bit message is transmitted electronically, there is 10% probability of one bit being transmitted erroneously. What is the chance that entire message is transmitted correctly?
- We need P(0 errors) in 8 trials, p(error)=0.1
- Using formula: P(x=0) = C(8,0) Γ 0.1β° Γ 0.9βΈ = 1 Γ 1 Γ 0.430
- Answer: 0.430
π Example 3: A surgery is successful for 75% patients. What is the probability of its success in at least 7 cases out of randomly selected 9 patients?
- p(success)=0.75 is outside the table (max 0.5)
- Invert: p(failure)=1-0.75=0.25
- "Success at least 7" means "Failure 2 or less"
- P(xβ₯7; n=9; p=0.75) = 1 - P(xβ€6; n=9; p=0.75) = 1 - 0.3995 = 0.6005 = 60%
Negative Binomial Distribution
A negative binomial experiment is a statistical experiment with the following properties: The experiment consists of x repeated trials; each trial has two outcomes (success/failure); the probability of success P is constant; trials are independent; and the experiment continues until r successes are observed, where r is specified in advance.
π Definition β Negative Binomial Distribution: A probability distribution that models the number of trials needed to achieve a fixed number of successes, where trials are independent and the probability of success is constant.
π Formula: ( b^*(x; r, P) = \binom{x-1}{r-1} \times P^r \times (1-P)^{x-r} )
- x = total number of trials
- r = number of successes (fixed)
- P = probability of success on each trial
π Example: Bob is a high school basketball player with 70% free throw shooting. What is the probability that Bob makes his first free throw on his fifth shot?
- This is a geometric distribution (special case of negative binomial with r=1)
- P = 0.70, x = 5, r = 1
- b*(5; 1, 0.7) = C(4,0) Γ 0.7ΒΉ Γ 0.3β΄
- b*(5; 1, 0.7) = 0.00567
NEGBINOMDIST
NEGBINOMDIST returns the negative binomial distribution. It returns the probability that there will be number_f failures before the number_s-th success, when the constant probability of a success is probability_s. This function is similar to the binomial distribution, except that the number of successes is fixed, and the number of trials is variable.
π Definition β NEGBINOMDIST: An Excel function that returns the probability of a given number of failures before a specified number of successes occurs.
π Syntax: NEGBINOMDIST(number_f, number_s, probability_s)
- number_f: the number of failures
- number_s: the threshold number of successes
- probability_s: the probability of a success
π Formula: ( \binom{x+r-1}{r-1} \times p^r \times (1-p)^x ) where x is number_f, r is number_s, and p is probability_s.
π Example: You need to find 10 people with excellent reflexes, and you know the probability that a candidate has these qualifications is 0.3. NEGBINOMDIST calculates the probability that you will interview a certain number of unqualified candidates before finding all 10 qualified candidates.
CRITBINOM
CRITBINOM returns the smallest value for which the cumulative binomial distribution is greater than or equal to a criterion value. Use this function for quality assurance applications, such as determining the greatest number of defective parts that are allowed to come off an assembly line run without rejecting the entire lot.
π Definition β CRITBINOM: An Excel function that returns the smallest value for which the cumulative binomial distribution meets or exceeds a specified criterion (alpha).
π Syntax: CRITBINOM(trials, probability_s, alpha)
- trials: number of Bernoulli trials
- probability_s: probability of success on each trial
- alpha: criterion value
π Example: With 6 Bernoulli trials, probability of success 0.5, and criterion value 0.75:
- CRITBINOM(6, 0.5, 0.75) returns 4
- This means the smallest number of successes for which the cumulative probability is at least 0.75 is 4
β Key Takeaways
Students must remember that cumulative binomial probabilities sum individual probabilities for ranges like "at most" or "at least" a certain number of successes, and these can be calculated using formulas or tables. The BINOMDIST function is essential for Excel-based calculations, where setting cumulative to TRUE gives the cumulative probability and FALSE gives the exact binomial probability. When probability values exceed table limits (0.5), the problem must be converted by switching success to failure (using p = 1 - original p). The negative binomial distribution differs from the binomial in that it fixes the number of successes and varies the number of trials. Finally, CRITBINOM is specifically designed for quality control to find the critical threshold number of successes.
π§ Quick Revision Questions
- What is the difference between the binomial distribution and the negative binomial distribution?
- How do you calculate cumulative binomial probability when the probability of success is greater than 0.5?
- In the BINOMDIST function, what does setting the cumulative parameter to TRUE return?
- What is the formula for the negative binomial distribution, and what does each variable represent?
- In what real-world application would you use the CRITBINOM function?