BTCRecover wildcards and typos: control the candidate count

A wildcard describes an unknown character

A wildcard is a rule inside a token. In the maintained tokenlist syntax, %d expands to one decimal digit and %2d to exactly two. A restricted set such as %[27] considers only the listed characters. Use the full syntax reference for ranges, case options and special characters.

Three measured examples

+ ^River
+ %d$

The one-digit fixture produces 10 candidates: River0 through River9. Replacing %d with %2d produces 100, including River00. Replacing it with %[27] produces just River2 and River7. Narrowing a set is useful only when supported by your memory; a faster search that excludes the answer will not help.

Variable length adds several search spaces

For decimal strings of one to three digits, the unconstrained count is 10 + 100 + 1,000 = 1,110 possibilities. It is not 30. This calculation assumes all lengths and leading zeros are allowed. BTCRecover’s %1,3d represents that length range. We verified fixed one- and two-digit fixtures; the 1,110 figure here is a mathematical explanation, not an additional reported software test.

A literal space inside a token can be expressed with %s. Our space exercise confirms that + River%sStone generates the one candidate River Stone. A regular space separating tokenlist alternatives has a different meaning.

Typos transform candidates after their construction

If you remember two nearly complete passwords, a passwordlist plus a focused typo rule may be clearer than a broad tokenlist. In our case-change exercise, two River passwords with --typos 1 --typos-case generate twelve candidates: each original and five variants that flip one letter’s case.

That count depends on these particular strings. Several typo types, repeated characters and higher typo limits can produce overlaps or much larger expansions. Read the typo guide rather than applying every option at once.

Keep hypotheses distinct

Try a narrowly supported hypothesis first, then expand one uncertainty. For example, test known year choices before all four-digit strings. Record what a completed run actually covered. Avoid presenting a no-match result as proof that a password was never used; it proves only that this configured search did not find a match.