An RNG, or random number generator, is a system that produces numbers with no predictable pattern. Computers use RNGs constantly: to shuffle a playlist, deal cards in a poker app, generate a unique world in Minecraft, or encrypt your bank transactions. The concept is simple, but the way machines actually produce “randomness” varies widely, and understanding the difference matters more than you might expect.
How Computers Fake Randomness
Most RNGs you encounter are pseudo-random number generators (PRNGs). They don’t produce truly random numbers. Instead, they run a mathematical algorithm that starts from a single number called a “seed” and churns out a long sequence of values that look random. The sequence passes casual inspection, but it’s entirely determined by that starting seed. If you know the seed and the algorithm, you can reproduce the exact same sequence every time.
That predictability is sometimes a feature, not a bug. Scientists running simulations need to reproduce their results, so they deliberately reuse the same seed. Game developers use fixed seeds to let players share identical procedurally generated worlds. When you enter a “world seed” in Minecraft, you’re telling the PRNG exactly where to start, which is why your friend can type the same seed and get the same landscape.
To avoid generating the same sequence on every launch, most software initializes the seed from something that changes constantly, like the current time down to the millisecond. This makes the output feel unpredictable to the user even though the underlying process is fully deterministic.
The Most Common Algorithm
The workhorse behind randomness in most programming languages is the Mersenne Twister, an algorithm introduced in 1997. It’s built into Python, Ruby, MATLAB, and many game engines. Its defining feature is an extraordinarily long cycle: it can produce 2^19937 minus 1 numbers before the sequence repeats. That number is so large it dwarfs the estimated number of atoms in the observable universe. For almost every practical purpose, the sequence never repeats during a program’s lifetime.
The Mersenne Twister also distributes its output evenly across a range, meaning no particular number is favored over another. That uniformity is critical for simulations, games, and statistical work. It is not, however, suitable for security applications, because someone who observes enough of its output can reverse-engineer the internal state and predict future numbers.
True Random Number Generators
When unpredictability genuinely matters, as in cryptography and online gambling, systems rely on true random number generators (TRNGs). These pull randomness not from an algorithm but from physical processes that are inherently unpredictable: tiny fluctuations in electrical thermal noise, timing jitter in oscillating circuits, atmospheric electromagnetic noise, or even quantum phenomena like radioactive decay.
Because TRNGs harvest randomness from physics rather than math, they have no seed. There is no starting value that could let someone reconstruct the output. Even if you captured one number, you could not predict the next. That non-deterministic quality is what makes them “true” random generators.
The tradeoff is speed. Sampling physical noise is slower than running an algorithm, so many real-world systems use a hybrid approach: a TRNG feeds a small stream of genuine randomness into a PRNG, periodically refreshing its internal state. Modern Intel and AMD processors include on-chip hardware random number generators for exactly this purpose.
How Quality Gets Tested
Not all RNGs are created equal. A flawed generator might favor certain numbers, fall into short repeating cycles, or produce subtle patterns that compromise security or skew simulation results. To catch these problems, the National Institute of Standards and Technology (NIST) publishes a standardized battery of statistical tests known as SP 800-22. The suite checks output for the kinds of patterns that shouldn’t appear in a truly random sequence: runs of identical digits, frequency imbalances, spectral regularities, and more. Any RNG used in cryptographic applications is expected to pass this suite before deployment.
RNG in Video Games
If you first heard the term “RNG” in a gaming context, you’re not alone. Gamers use it as shorthand for any element of chance built into a game’s mechanics. When a boss in Diablo drops a legendary weapon, RNG determined the item’s rarity, its stats, and possibly its visual appearance. When a roguelike game like Hades or The Binding of Isaac rearranges its level layouts, enemy spawns, and power-up placements each run, that’s RNG creating a fresh experience every time.
Procedural generation takes RNG further. Games like Minecraft and No Man’s Sky use random number generators to build entire worlds from scratch, placing mountains, rivers, biomes, and cave systems according to randomized parameters. The result is a near-infinite variety of environments from a relatively small set of rules.
RNG is also at the center of ongoing debates about fairness. Loot boxes and gacha systems use random number generation to determine what players receive for real money, and the opacity of those systems has drawn criticism. Players increasingly push for transparency around drop rates, especially when monetization is involved, to distinguish exciting randomness from exploitative design.
RNG in Science and Security
Monte Carlo simulations are one of the most important scientific applications of RNGs. Researchers model complex systems, from radiation transport in nuclear physics to protein folding in biology, by running thousands or millions of randomized trials and analyzing the statistical results. The quality of the RNG directly affects the reliability of those results, which is why scientific computing libraries are careful about which algorithms they ship.
In cryptography, RNG is foundational. Encryption keys, session tokens, one-time passwords, and secure communications all depend on numbers that an attacker cannot guess. A weak or predictable RNG in a security system isn’t a minor flaw; it’s a catastrophic vulnerability. This is why cryptographic systems insist on true randomness or, at minimum, cryptographically secure PRNGs that are specifically designed to resist reverse-engineering, unlike general-purpose algorithms such as the Mersenne Twister.
Why “Random” Is Harder Than It Sounds
Humans are surprisingly bad at recognizing real randomness. A truly random sequence of coin flips will contain long streaks of heads or tails that feel “non-random” to us. Our brains look for patterns even where none exist. This mismatch is partly why RNG in games feels unfair sometimes: a player who loses five 50/50 rolls in a row suspects the system is rigged, but that outcome is perfectly normal in a random distribution.
Some games address this perception gap with “pity timers” or weighted systems that guarantee a rare drop after a certain number of failures. Technically, these systems are less random, but they feel more fair. It’s a reminder that the goal of an RNG isn’t always pure statistical randomness. It’s to serve the purpose it was built for, whether that’s unbreakable encryption, reproducible science, or a game that keeps you coming back.