Probability and Distributions

TL;DR

Probability helps you quantify how likely events are, from coin flips to complex data. Distributions describe the patterns these probabilities follow, showing which outcomes are common and which are rare. Understanding these lets you make informed predictions and analyze data more effectively.

1. The Mental Model

Think of probability as a way to measure uncertainty, giving numbers to "maybe." Distributions are like charts or graphs that show you all the possible "maybes" and how often each one is expected to happen.

2. The Core Material

Probability is all about chance. You're trying to figure out the likelihood of something happening. It's always a number between 0 (impossible) and 1 (certain). You often express it as a percentage too.

The basic formula for probability of an event (let's call it A) is:

P(A) = (Number of favorable outcomes) / (Total number of possible outcomes)

Types of Events

Photo by SINAL Multimédia on Pexels

• Independent Events: One event doesn't affect the other. Like flipping a coin twice – the first flip doesn't change the odds of the second.
  • P(A and B) = P(A) * P(B)
• Dependent Events: The outcome of one event influences the next. Drawing two cards from a deck without replacement is a good example.
  • P(A and B) = P(A) * P(B|A) (where P(B|A) is the probability of B happening given that A has already happened).
• Mutually Exclusive Events: Events that can't happen at the same time. You can't roll a 2 and a 5 on a single die roll.
  • P(A or B) = P(A) + P(B)

Random Variables

Photo by Markus Spiske on Pexels

A random variable is just a number that represents the outcome of a random event.
* Discrete Random Variable: Can only take specific, separate values (like integers). Examples: number of heads in 3 coin flips (0, 1, 2, 3), number of cars passing a point in an hour.
* Continuous Random Variable: Can take any value within a range. Examples: height, weight, time taken to complete a task.

Probability Distributions

Photo by DS stories on Pexels

A probability distribution tells you all the possible values a random variable can take and how likely each value is.

graph TD
    A["Probability Distributions"] --> B["Discrete Distributions"]
    A --> C["Continuous Distributions"]

    B --> B1["Binomial Distribution"]
    B1 --"Fixed number of trials"--> B1a["Each trial: two outcomes (success/failure)"]
    B1 --"Probability of success (p) is constant"--> B1b["Trials are independent"]

    B --> B2["Poisson Distribution"]
    B2 --"Events occur at a constant average rate"--> B2a["Events are independent"]
    B2 --"Occurrences in an interval"--> B2b["Number of events is countable"]

    C --> C1["Normal Distribution (Gaussian)"]
    C1 --"Bell-shaped curve"--> C1a["Symmetric around the mean"]
    C1 --"Mean, median, mode are equal"--> C1b["Defined by mean (μ) and standard deviation (σ)"]

    C --> C2["Uniform Distribution"]
    C2 --"All outcomes equally likely within an interval"--> C2a["Constant probability density"]
    C2 --"Rectangular shape"--> C2b["Min and Max values define it"]

Common Discrete Distributions:

• Binomial Distribution: Used for situations where you have a fixed number of independent trials, each with only two possible outcomes (success/failure), and the probability of success is constant.
  • Example: What's the probability of getting exactly 3 heads in 5 coin flips?
• Poisson Distribution: Used for counting the number of events that occur within a fixed interval of time or space, when these events happen with a known constant mean rate and independently of the time since the last event.
  • Example: Number of phone calls a call center receives in an hour.

Common Continuous Distributions:

• Normal Distribution (Gaussian Distribution): This is perhaps the most important distribution. It's bell-shaped and symmetric, with most values clustering around the mean. Many natural phenomena (like heights, IQ scores) follow this. It's defined by its mean ($\mu$) and standard deviation ($\sigma$).
  • A key concept here is the Standard Normal Distribution (Z-distribution), which has a mean of 0 and a standard deviation of 1. You can convert any normal distribution to a standard normal distribution using the formula: Z = (X - $\mu$) / $\sigma$. This allows you to use standard Z-tables to find probabilities.
• Uniform Distribution: All values within a given range are equally likely.
  • Example: Rolling a fair die (if treated continuously), or a random number generator that produces numbers between 0 and 1.

3. Worked Example

Let's say you're a quality control inspector. You know that 10% of the products coming off an assembly line are defective. You randomly select 8 products for inspection. What's the probability that exactly 2 of them are defective?

This is a classic Binomial Distribution problem because:
1. Fixed number of trials (n): You inspect 8 products.
2. Two possible outcomes: Defective (success) or Not Defective (failure).
3. Constant probability of success (p): Probability of a product being defective is 0.10.
4. Independent trials: One product being defective doesn't affect another.

The formula for the probability of k successes in n trials in a Binomial distribution is:

P(X=k) = C(n, k) * p^k * (1-p)^(n-k)

Where:
* C(n, k) = n! / (k! * (n-k)!) is the number of combinations.
* n = number of trials = 8
* k = number of successes (defective products) = 2
* p = probability of success (defective) = 0.10
* 1-p = probability of failure (not defective) = 0.90

Let's calculate:
1. C(8, 2): 8! / (2! * (8-2)!) = 8! / (2! * 6!) = (8 * 7) / (2 * 1) = 28
2. p^k: (0.10)^2 = 0.01
3. (1-p)^(n-k): (0.90)^(8-2) = (0.90)^6 $\approx$ 0.531441

Now, multiply these together:
P(X=2) = 28 * 0.01 * 0.531441 $\approx$ 0.1488

So, there's approximately a 14.88% chance that exactly 2 out of the 8 products you inspect will be defective.

4. Key Takeaways

• Probability quantifies the likelihood of events, always between 0 and 1.
• Random variables are numerical outcomes of random phenomena; they can be discrete or continuous.
• Probability distributions map out all possible outcomes of a random variable and their associated probabilities.
• Binomial distribution is for a fixed number of trials with two outcomes (success/failure).
• Poisson distribution models events occurring at a constant average rate over an interval.
• Normal distribution is bell-shaped, symmetric, and characterized by its mean and standard deviation.
• The Z-score helps standardize any normal distribution, allowing for easier probability calculations.

Common Mistakes to Avoid:
- Confusing independent and dependent events when calculating joint probabilities.
- Applying discrete distribution formulas to continuous problems, or vice-versa.
- Misinterpreting the mean ($\mu$) and standard deviation ($\sigma$) in a normal distribution.
- Forgetting that the sum of all probabilities in any distribution must equal 1.

5. Now Try It

You're tracking the number of emails you receive on a typical workday. Based on past data, you've found that you receive an average of 10 emails per hour. Assuming email arrivals follow a Poisson distribution, calculate the probability that you receive exactly 7 emails in the next hour.

What success looks like: You should be able to identify the correct distribution, set up the parameters ($\lambda$ and k), and use the Poisson probability mass function (P(X=k) = ($\lambda^k$ * e^($-\lambda$)) / k!) to get the final probability, which should be around 9%.