Engineering Fluency OS

Why This Matters

Software does not live in a world of certainty. Will a user click this button? Is this email spam? Will the server respond in under 100ms? Every one of these questions is a probability question. Probability is the math of uncertainty, and it is the foundation of machine learning, A/B testing, security, and data-driven decision making.

Conditional probability asks: given that something has already happened, how does that change the odds? And Bayes' theorem lets you flip conditional probabilities around -- going from "what is the probability of a positive test given you are sick" to "what is the probability you are sick given a positive test." This reasoning pattern powers spam filters, medical diagnostics, and Bayesian machine learning.

Define Terms

Visual Model

Event

P(A)Outcome A

P(not A)Outcome not-A

P(B|A)

P(B|not A)

P(A|B)Conditional

Bayes Theorem

The full process at a glance. Click Start tour to walk through each step.

Probability quantifies uncertainty. Bayes' theorem lets you update beliefs with new evidence.

Code Example

Code

// Basic probability: favorable / total
function probability(favorable, total) {
  return favorable / total;
}

// Rolling a 6-sided die
const pEven = probability(3, 6); // {2,4,6} out of {1..6}
console.log("P(even):", pEven); // 0.5

// Independent events: P(A and B) = P(A) * P(B)
const pTwoHeads = 0.5 * 0.5;
console.log("P(2 heads):", pTwoHeads); // 0.25

// Simulate probability with random trials
function simulate(trials, eventFn) {
  let successes = 0;
  for (let i = 0; i < trials; i++) {
    if (eventFn()) successes++;
  }
  return successes / trials;
}

const simEven = simulate(10000, () => {
  const die = Math.floor(Math.random() * 6) + 1;
  return die % 2 === 0;
});
console.log("Simulated P(even):", simEven.toFixed(3)); // ~0.500

// Bayes Theorem
// P(sick | positive) = P(positive | sick) * P(sick) / P(positive)
function bayes(pBA, pA, pB) {
  return (pBA * pA) / pB;
}

const pSick = 0.01;              // 1% of people are sick
const pPosGivenSick = 0.99;      // 99% true positive rate
const pPosGivenHealthy = 0.05;   // 5% false positive rate
const pPos = pPosGivenSick * pSick + pPosGivenHealthy * (1 - pSick);
const pSickGivenPos = bayes(pPosGivenSick, pSick, pPos);
console.log("P(sick|positive):", pSickGivenPos.toFixed(3)); // ~0.167

Interactive Experiment

Try these exercises:

Roll a die 1000 times in code. How close is the simulated probability of rolling a 6 to the theoretical 1/6?
Flip two coins 10000 times. How often are both heads? Does the simulation match 0.25?
If P(rain) = 0.3 and P(umbrella | rain) = 0.9, what is P(rain AND umbrella)?
A disease affects 1 in 1000 people. A test is 95% accurate. If you test positive, what is the actual probability you have the disease? (Hint: use Bayes.)
Why is the answer to the previous question surprisingly low?

Quick Quiz

Coding Challenge

Monte Carlo Simulation

Write a function called `estimatePi` that uses a Monte Carlo simulation to estimate the value of Pi. Generate `n` random points with x and y in the range [0, 1). Count how many fall inside a quarter circle (x^2 + y^2 < 1). Return 4 * (inside / n). With n=100000 you should get a value close to 3.14.

Loading editor...

Real-World Usage

Probability is the backbone of data-driven software:

A/B testing: "Is variant B actually better, or was it luck?" Statistical significance tests use probability to answer this.
Spam filters: Naive Bayes classifiers compute P(spam | words in email) using Bayes' theorem to classify messages.
Machine learning: Training a model means finding parameters that maximize the probability of the observed data.
Reliability engineering: "What is the probability all 5 servers fail at once?" drives redundancy decisions.
Cryptography: Security depends on the probability of guessing a key being astronomically low.

Probability