Linear congruential generator (LCG) – makes numbers by a simple formula; fast but patterns leak through if you look closely
Xorshift – flips and shifts bits to make new numbers; extremely fast, decent for games, bad for security
Xoshiro / Xoroshiro – improved xorshift versions that hide patterns better while staying very fast
Mersenne Twister – produces very high-quality-looking randomness with a huge cycle, but attackers can reverse it
PCG (permuted congruential generator) – takes a simple generator and scrambles the output to remove obvious patterns
SplitMix – a fast, simple generator mainly used to start (seed) better generators
Middle-square – squares a number and takes the middle digits; clever, old, and breaks horribly
Lagged Fibonacci – creates numbers by combining earlier ones in the sequence; better than simple formulas
Additive / subtractive generators – mix previous values using addition or subtraction to get new ones
Blum Blum Shub – extremely secure mathematically, painfully slow, used mostly in theory
ChaCha20 RNG – uses a modern encryption algorithm to produce randomness; fast and secure
AES-CTR DRBG – uses the AES encryption algorithm to generate random-looking data safely
Hash-based RNG – repeatedly hashes data so outputs are unpredictable without knowing the input
Hardware RNG (HRNG) – reads physical noise from the computer hardware itself
Entropy pool mixers – collect randomness from many messy sources and blend them together safely
Quasi-random (low-discrepancy) sequences – not random at all, just evenly spaced to avoid clumping
True random number generator wrappers – software doors that let programs access physical randomness