EXPLOITATIVE PLAY · FIELD NOTE
Node Locking in Plain English: Test the Mistake, Not a Fantasy
Use node locking to model opponent mistakes, calculate best responses, avoid unrealistic assumptions, and convert solver findings into practice.
Node locking sounds technical, but the idea is simple. You tell the solver, “Pretend my opponent makes this specific mistake,” and it shows you the best response. The dangerous part is feeding it a fantasy instead of a measured mistake.
The short answer
Solve the normal strategy first. Change one opponent behavior that you have evidence for, then compare the result. Make small locks, test nearby assumptions, and turn the stable changes into rules you can actually use.
Three useful words
A node is one decision point, such as the big blind facing a one-third-pot flop bet. A node lock forces the opponent to use the behavior you choose at that point—for example, folding 60% instead of 45%. A best response is the most profitable way to play against that forced mistake.
Two quick examples
- Reasonable: Your database shows the big blind folds 60% on
K♦ 7♣ 2♥where the baseline folds 45%. Lock that response near 60% and see whether hands such asQ♠ J♠become profitable bets. - Not reasonable: You delete every check-raise and strong call on that flop because the pool “seems passive.” The solver may tell you to bet
9♣ 8♣, but it is answering an opponent who does not exist.
Practice this spot first
Assume the big blind overfolds this river by 15 percentage points. Decide whether queen-jack becomes a bluff, then ask whether the answer survives a smaller 5-point overfold.
Solve the baseline first
Record the solver's normal actions and EVs before changing anything. That gives you something to compare with the exploit. Without the normal answer, a dramatic new strategy does not tell you what changed or why.
Lock one measurable mistake
Change one behavior supported by data or a clear read: too much folding to one size, too few raises, or too many calls. Leave the rest of the strategy alone. Forcing every action to an extreme creates an opponent who does not exist.
Look at all the hands that change
The best response may bluff more hands, value bet more thinly, change size, or add strong hands to a check. Look at which kinds of hands change first. If complete misses with useful blockers become bluffs before hands that can win at showdown, you have found a rule you can reuse.
Stress-test the assumption
Test a smaller and larger mistake. If the exploit appears only when you force the opponent to make an extreme error, it may not work against a real player. Check similar boards and sizes too, so you know the lesson is not unique to this one decision.
Convert output into drills
Practice three versions: the normal opponent, an opponent making a small mistake, and one making a large mistake. Guess which action changes before seeing the answer. This turns node locking into a learning tool instead of a way to make a colorful chart.
Rule of thumb: Force the solver to copy a mistake you have evidence for, not one you imagined. The answer is useful only when the model resembles a real opponent.
Document every lock
Save the original frequency, the frequency you forced, your evidence and sample size, and the change in EV. Also record the board, positions, stacks, and bet size. You will be able to repeat the test and avoid turning one surprising result into a rule for unrelated situations.
Keep studying
Place each experiment inside the broader GTO-versus-exploitative framework, ground the lock in observable loose-passive behavior, and document the result with a repeatable poker-hand review process.