What Is the Practical Difference between “Opt-in RBF” and “Full RBF”?

The practical difference lies in the replacement policy of network nodes. Opt-in RBF (BIP125) requires the original transaction to explicitly signal its replaceability by setting the 'nSequence' field.

If the flag is absent, nodes generally won't relay a replacement. Full RBF is a node policy where any unconfirmed transaction, regardless of its RBF flag, can be replaced by a new, higher-fee transaction.

Full RBF maximizes the fee market but completely invalidates 0-conf security.

What Are the Fee-Bumping Rules (E.g. Minimum Fee Increase) for a BIP125 RBF Replacement?
How Does a Node’s Mempool Size Limit Influence Its RBF Relay Policy?
What Is the Significance of the ‘Nsequence’ Field in Enabling RBF?
How Do Replace-by-Fee (RBF) Mechanisms Help Users on a Congested Network?
What Is the Replace-by-Fee (RBF) Protocol and How Is It Activated?
How Does the ‘Nsequence’ Field Differ in Usage for Opt-in RBF versus Relative Time-Locks?
What Is a ‘Replace-by-Fee’ (RBF) Transaction?
What Are the Risks of Using RBF for a Sender or Receiver?

Glossar