Random Number Generator

How It Works

Random number generation for everyday use does not need to be cryptographically secure, but this generator uses window.crypto.getRandomValues() anyway, which eliminates any statistical bias present in Math.random().

Generating a single uniform random integer in [min, max]:
1. Calculate range = max − min + 1 (the number of possible values, inclusive of both endpoints).
2. Generate a random 32-bit unsigned integer R using crypto.getRandomValues(new Uint32Array(1)).
3. Result = min + (R % range).

This gives a uniform distribution over all integers from min to max inclusive.

Generating multiple numbers without duplicates (sampling without replacement):
This uses a Fisher-Yates shuffle on an array of all valid integers, then takes the first N values.
1. Create array [min, min+1, ..., max].
2. Shuffle using Fisher-Yates: for i from (range−1) down to 1, pick random j in [0, i], swap arr[i] and arr[j].
3. Return first N elements.

Fisher-Yates is O(range) time, which is fine for typical ranges but could be slow for very large ranges (e.g., min=1, max=1,000,000, count=5). For large ranges with small counts, a hash-based approach avoids building the full array — but the Fisher-Yates method is used here for simplicity and transparency.

Worked example: Pick 6 numbers from 1 to 49 (lottery).
Build array [1, 2, ..., 49]. Shuffle. Return first 6.
Result might be: 7, 23, 14, 41, 2, 35.
These are guaranteed unique and uniformly distributed.

Generating duplicates: When duplicates are allowed, each number is drawn independently. This means the same number can appear multiple times, with each draw being independent.

Applications: Lottery picks, card dealing, random sampling for surveys, randomizing quiz questions, dice roll simulations, assigning random teams, generating test data.