How do complex Duplicate Rules with fuzzy matching work?

Complex Duplicate Rules combine multiple Matching Rules with different algorithms:

1. Exact Match: Records must have identical values (case-insensitive). Good for email addresses, phone numbers.
2. Fuzzy Match: Uses phonetic algorithms (First Name fuzzy, Last Name fuzzy) to catch "John Smith" vs "Jon Smyth". Good for name-based matching.
3. Acronym: Matches abbreviated vs full company names.
4. Initials: Matches "J. Smith" to "John Smith". Matching rules use a match key (a transformed value) — two records are potential duplicates if their match keys are equal or meet the fuzzy threshold. Duplicate rules can apply different actions for leads vs contacts, block on insert but allow on edit, or only report duplicates without blocking.