METHOD NOTES

Decide what
a draw means.

Random selection starts with a defined population. A clean list and an explicit method are more useful than a mysterious spinning animation.

Without replacement

Each eligible input row may appear at most once in the result. A sample of 5 therefore needs at least 5 eligible rows. The tool uses a partial Fisher–Yates shuffle: it swaps a randomly chosen remaining row into each output position, then returns the requested prefix.

If duplicate labels are kept as separate entries, a label can appear twice even though no row was selected twice. If you mean distinct labels, choose “Deduplicate exact labels.”

With replacement

Every draw starts from the full eligible population. A row selected on the first draw remains eligible for the second. A sample may exceed the population size, subject to the tool’s 10,000-result limit. Repetitions are expected and do not indicate a malfunction.

Shuffling the full list

A shuffle returns every eligible row in a randomized order. It always runs without replacement, regardless of the sampling-method setting. The duplicate policy still applies, so deduplicating first changes how many entries appear.

What counts as a duplicate?

Leading and trailing spaces are removed, and blank lines are ignored. Exact remaining labels match case-sensitively. “Maple” and “MAPLE” are different labels. Entries are separated by lines, not commas; a comma within an entry remains part of its label.

Document the selection

The CSV contains position, entry, method, duplicate policy, and eligible population size. Keep your original source list alongside it. The export is not an independent audit record, and it does not prove that a source list was complete or unbiased.

Limits worth keeping in view

The tool uses the browser’s pseudorandom generator without a user-controlled seed. Results cannot be replayed from a seed. It is suitable for everyday list organization and learning, not regulated draws or security-sensitive selection. It cannot make an unrepresentative input population representative.

Return to the sampler

Open the full wheel studio