About

Generating uniformly distributed odd integers across an arbitrary range is not as trivial as it appears. A naive approach - generate a random integer, check if it is odd, retry if not - introduces rejection bias and unbounded loop risk at scale. This tool computes the exact cardinality of odd integers in the closed interval [a, b] and maps a uniform random index directly onto that subset, guaranteeing O(1) generation per number with zero rejections. When duplicate-free output is required, the generator builds the full odd-integer pool and applies a Fisher-Yates shuffle, which runs in O(n) time and produces an unbiased permutation. Requesting more unique values than the pool contains is detected before generation and flagged as an error - not silently truncated.

Applications range from cryptographic nonce seeding (many schemes require odd moduli) to Monte Carlo sampling on odd lattices, classroom exercises in parity, and lottery-style draws. The tool assumes a standard uniform PRNG (Math.random); it is not suitable for cryptographic key generation. Pro tip: if your range bounds are even, they are automatically narrowed inward to the nearest odd values - verify the adjusted bounds in the output summary.

Formulas

The count of odd integers in a closed interval [a, b] where a ≤ b is computed as:

C = b − a2 + 1

where a is adjusted to the nearest odd ceiling and b to the nearest odd floor if they are even. More precisely, using floor division:

C = floor(b ÷ 2) − floor((a − 1) ÷ 2)

To generate a uniformly random odd integer without rejection, the generator maps a random index k ∈ {0, 1, …, C − 1} to an odd value:

n = a_odd + 2k

where a_odd is the smallest odd integer ≥ a. For duplicate-free generation, a Fisher-Yates shuffle on the pool array guarantees an unbiased permutation in O(C) time.

Where: C = total count of odd integers in range, a = minimum bound, b = maximum bound, k = uniform random index, n = generated odd number, a_odd = adjusted odd minimum.

Reference Data

Range	Odd Count Formula	Count	Example Odds	Notes
1 - 10	⌊10÷2⌋ − ⌊0÷2⌋	5	1, 3, 5, 7, 9	Symmetric around 5
1 - 100	⌊100÷2⌋ − ⌊0÷2⌋	50	1, 3, …, 99	Half of all integers
1 - 1000		500	1, 3, …, 999	Dense pool
0 - 1		1	1	Only one odd value
−10 - 10		10	−9, −7, …, 9	Negative odds included
−99 - −1		50	−99, −97, …, −1	Fully negative range
2 - 2		0	-	No odd integers exist
7 - 7		1	7	Single-value range
1 - 10000		5000	1, 3, …, 9999	Large pool, fast generation
−1000 - 1000		1000	−999, …, 999	Centered symmetric range
100 - 200	⌊200÷2⌋ − ⌊99÷2⌋	50	101, 103, …, 199	Even bounds adjusted inward
0 - 0		0	-	Zero is even
1 - 2		1	1	Minimal valid range
999 - 1001		2	999, 1001	Tight odd-only range
−5 - 5		5	−5, −3, −1, 1, 3, 5	Includes zero (even, excluded)

Frequently Asked Questions

Rather than generating arbitrary integers and filtering out evens (which wastes entropy and biases timing), the tool calculates the exact count C of odd integers in your range, generates a uniform random index k in [0, C − 1], and maps it directly to a_odd + 2k. Every odd integer in the range has exactly 1C probability of selection.

The tool detects that the adjusted odd ceiling exceeds the adjusted odd floor, yielding a pool size of 0. It displays an error notification explaining that no odd integers exist within the specified bounds. This prevents silent empty output.

No. If duplicates are disabled, the tool computes the pool size C and compares it to your requested count. If count > C, generation is blocked with a message stating the maximum available. For example, the range [1, 10] contains only 5 odd integers, so requesting 6 unique values is impossible.

Yes. The parity check uses the mathematical definition: an integer n is odd if n mod 2 ≠ 0. In JavaScript, the % operator preserves sign, so −3 % 2 = −1 (truthy). Negative bounds are fully supported. The range [−9, −1] yields {−9, −7, −5, −3, −1}.

The tool uses Math.random, which provides a pseudo-random number generator adequate for simulations, sampling, educational exercises, and casual draws. It is NOT cryptographically secure. If you need unpredictable randomness for security-sensitive applications (nonce generation, token creation), you should post-process the output or use a dedicated CSPRNG. This tool is designed for general-purpose odd number generation.

The Fisher-Yates (Knuth) shuffle iterates from the last element to the first, swapping each element with a randomly chosen element from the remaining unshuffled portion. This produces each of the C! permutations with equal probability 1C!. The tool then slices the first n elements, equivalent to drawing without replacement.