How UUID collision probability works
Collision risk follows the birthday problem: the chance rises with the square of the number of generated IDs and falls with the size of the ID space. A UUID v4 has 122 random bits, so its available random space is huge even at very large generation volumes.
Why UUID v7 and ULID use a millisecond count
Time-ordered IDs split the identifier into a timestamp and a random section. Collision risk is concentrated inside a single timestamp bucket, so the relevant input is the peak number of IDs generated in the same millisecond.
When custom IDs need extra care
Short custom IDs can be useful for public URLs and human-facing codes, but each removed character reduces the ID space. The calculator makes that tradeoff visible before the format becomes part of an API or database schema.