Node-locking

Node-locking means fixing one player's strategy at a chosen point in the game tree to a specific, usually imperfect, plan, then re-solving the rest of the tree around that fixed play. Instead of letting both sides search for equilibrium, one side is told exactly what it does at that node, and the solver only optimises the other side's response.

This turns a GTO tool into an exploit-finding tool. A pure solve tells you the balanced play against a perfect opponent, which is not what you are facing at most tables. Lock a node to the tendency you have actually observed, such as a player who never bluffs a certain river, and the re-solve returns the maximally profitable response to that specific, flawed strategy rather than the cautious, balanced one.

Worked example. Lock an opponent's river strategy to betting only strong hands, then re-solve your side. The result stops calling with bluff-catchers that only beat bluffs, since the locked strategy has none, and shifts that value elsewhere.

Node-locking is how solver study connects theory to a real opponent: the equilibrium sets the baseline, and locking a node measures the exact cost of a specific deviation from it.