Quick Answer: In information theory, entropy measures the amount of unpredictability or uncertainty in a system. For random number generation, higher entropy = harder to predict = more secure randomness. Computers collect entropy from unpredictable hardware events (key timings, mouse movements, network packets) to seed cryptographic random generators.
Two Meanings of Entropy
Entropy has meanings in both thermodynamics (a measure of physical disorder in a system) and information theory (a measure of uncertainty or unpredictability in information). For computer random number generation, information entropy is the relevant concept. Both definitions share the underlying idea: entropy measures how unpredictable or disordered something is.
Shannon Entropy: The Mathematical Definition
Claude Shannon defined information entropy as H = -∑ p(x) log₂ p(x) across all possible values x with probability p(x). Simplification: a single fair coin flip has exactly 1 bit of entropy — the maximum for a binary event. A biased coin with 90% heads has less than 1 bit of entropy — it is more predictable, so it carries less information.
How Computers Collect Entropy
- Keystroke timing: The precise inter-keystroke timing (measured in microseconds) is essentially random
- Mouse movement: Precise mouse position trajectory contains unpredictable micro-variations
- Network packet timing: Variable network latency contributes entropy
- CPU thermal noise: Hardware performance variations create measurable noise
- Dedicated hardware RNG chips: Modern CPUs include hardware entropy sources (Intel RDRAND, AMD equivalent)
Entropy in PickRandom.online
The Web Crypto API (window.crypto.getRandomValues()) uses the browser's OS-level entropy sources — the same hardware entropy feeds that power SSL key generation. When you generate a random number or flip a coin on PickRandom.online, the result is derived from genuine hardware entropy — making it computationally impossible to predict.