Pre-Commitment Devices

A simple explanation

A pre-commitment device is a structure you build, in a clear moment, to make a predictable failure of a later moment costlier than the original goal. Ulysses asked the crew to lash him to the mast and refuse his orders when the sirens sang. He knew which version of himself would do the choosing later, and he did not trust it. So he moved the choice earlier, into a moment where the chooser was steady.

The modern forms are quieter but the shape is the same. You pre-pay the gym. You schedule the automatic transfer to savings. You install the internet blocker before the working day begins. You batch-cook on Sunday so Tuesday's tired self has no decision left to make. In each case present-self spends a small amount of design effort so future-self does not have to spend a large amount of willpower.

An everyday example

You know, by the eighth month of watching it happen, that you do not write the harder paragraphs of a project after 4 p.m. Your morning-self is patient with difficulty; your late-afternoon-self picks the email tab and never comes back. The honest reading is not that you lack discipline. It is that the two selves are reliably different people, and the afternoon one wins every contest of pure will.

So on Sunday evening you install Freedom on your laptop, set it to block social and email between 9 a.m. and 12 p.m. on writing days, and set the unlock fee to a real amount — not a symbolic dollar, an amount that makes you wince. Tuesday morning arrives. The blocker runs. The afternoon-self never gets the choice. The paragraph gets written.

What the device did was not generate willpower. It moved the decision into a moment where willpower was abundant, and removed the decision from a moment where it was not.

What is a pre-commitment device?

Any mechanism by which present-self deliberately constrains the option set of future-self toward a goal both selves, in principle, share. The constraint can be financial (a forfeit, a deposit, a pre-payment), social (a witness, a partner, a public stake), structural (a default, a friction, a removed option), or temporal (an irreversible scheduling). What unites them is the asymmetry: the cold-state chooser binds the hot-state chooser, anticipating that the hot state will see the world differently.

Thomas Schelling, who first named the family, treated it as a strategic problem inside the self. Two parts of the same person are bargaining across time, and they want different things. The cold part — the planner, the morning-self, the one currently reading this — can install a credible commitment that makes the hot part's preferred move costly. The credibility is the whole game. A pre-commitment without a real cost is a wish.

Do pre-commitment devices actually work?

They work when three conditions are met, and not when they are not.

The failure they target must be predictable. Pre-commitment is precision instrumentation; it is not a general fix for self-control. If you do not yet know where your hot-state self reliably defects, the device is shooting in the dark and will mostly bind your steady self for no reason.

The cost of breaking the commitment must be real. A forfeit you do not feel is a placebo. stickK pioneered the anti-charity twist — your forfeit goes to a cause you actively dislike — because money flowing to a neutral place is too easy to absorb. The sting is the mechanism.

The device must be proportionate. Lashing yourself to the mast against a one-time temptation is wise. Lashing yourself against every minor desire of every future self is self-tyranny, and the residue compounds even when the deposit is small.

Examples that show the range

A short tour, less as a menu than as a map:

• stickK.com — present-self pledges a behaviour, names a referee, attaches a forfeit (often to an anti-charity). Miss the goal, the money goes. The anti-charity move is what makes the forfeit bite.
• Beeminder — a graph with a road. Stay on the road, nothing happens. Drift off, you pledge the next amount up the ladder. The ladder is the device's teeth.
• Automatic savings transfer — the day after payday, before the brain has time to revise the budget, the money moves. The brain finds a slightly poorer account and adjusts. This is the most boring and one of the strongest pre-commitments most people ever install.
• Gym membership with a friend — the financial cost is mild; the social cost of standing them up is what does the work.
• Freedom / Cold Turkey / SelfControl — internet blockers that the cold-state self installs against the hot-state self. Strongest when the unlock cost is set high enough to be unpleasant.
• Batch-cooked meals — Sunday's-self makes Tuesday's-tired-self have no decision to make about dinner. A pre-commitment without a forfeit, working through the removal of the choice itself.
• Donor cards, advance directives, irrevocable trusts — the cold-state self binding a future hot-state self at much higher stakes. Same family, far end of the seriousness scale.

The behavioural loop

How a working pre-commitment runs, end to end:

1. Pattern recognition — present-self notices, honestly, that future-self predictably defects at a specific moment under a specific pull.
2. Cold-state design — in a clear hour, present-self builds the device: the forfeit, the witness, the block, the default. The design itself is a small deposit.
3. The hot moment arrives — the predictable pull comes on schedule. Future-self is now the chooser, and is exactly as unreliable as forecast.
4. The device runs — the option future-self wanted is either removed, made expensive, or made socially visible. The hot-state choice is not "do I resist?" but "do I pay?"
5. Residue and harvest — if the device fits, the deposit lands. The savings accumulate, the words get written, the workout happens. If the device was over-fitted — built against a desire that wasn't actually a failure mode — the residue is mild self-distrust and a small grudge against the design.

Emotional drivers

Three layers, often present at once:

• A specific tiredness, in the cold state, of fighting the same fight every week with future-self and losing. The device is an admission and a relief.
• A faint pride in the design itself — the satisfaction of having moved a real lever in the world, not just promised more discipline.
• A small bitterness if the device over-binds — the sense of being managed by an earlier self who did not know what this version was actually facing.

The third feeling is information. Frequent bitterness against your own pre-commitments is the system's signal that the design is too rigid or aimed at a fight that turned out not to be the real one.

What your nervous system does

Hot and cold are not metaphors here. The cold state is a more parasympathetic, prefrontal-led reading of the world in which costs and benefits sit in a single frame and patience is abundant. The hot state is a more sympathetic, limbic-led reading in which the proximal cue (the screen, the sugar, the inbox, the dread) takes up most of the field and the future stake fades.

The two states genuinely see different worlds. The cold-state self is not lying when it commits; the hot-state self is not weak when it defects. They are running on different neurochemical settings. A pre-commitment device works because it puts a structural object — a forfeit, a removed option, a witness — into the hot-state world that the hot-state self cannot easily make disappear, even with the field narrowed.

This is also why willpower-alone strategies fail at predictable moments. Asking the hot-state self to act on cold-state values is asking it to hold a frame its nervous system is not currently running.

The DojoWell interpretation

The Meaning System's job, when it is working, is to keep behaviour aligned with what the slow system has named as load-bearing. In the cold state this alignment is easy — the System and the chooser are reading the same map. In the hot state the map narrows, the Reward and Threat Systems spike, and the Meaning System's signal becomes a whisper against louder traffic.

The substitute on offer in the hot moment is "rely on future-self willpower." It wears the garb of virtue — strength, character, discipline — and it loses to predictable hot-state pull almost every time the pull is real. The deposit it promises (the chosen behaviour) does not land; the effort it costs (the contested moment) runs; the residue (another defection, another revision downward of self-trust) accumulates. Density collapses by exactly the mechanism the equation describes.

A well-fitted pre-commitment device is the Meaning System's cold-state intervention. Present-self, while the System's signal is clear, installs structure that does not require future-self to hold the frame alone. The deposit lands because the choice was made when the chooser was steady. The residue stays low because the hot-state self is not asked to do the impossible. The effort is paid up front, in the design, which is exactly where effort is cheapest.

This is also why the device must be matched to a predictable failure, not sprinkled across life as generalised self-management. Pre-commitment used precisely is a high-density instrument. Pre-commitment used as a substitute for knowing yourself — as a way of binding every future self because no future self is trusted — runs the equation in reverse: effort and residue accumulate without deposit, and the loop produces self-distrust rather than freedom. The signature delayed_harvest belongs to the first usage. The second collects effort without deposit and reads as low density.

The resolution, in the framework's voice, is small and specific. Identify the predictable hot-state failure point. Design the pre-commitment specifically against it. Attach a real cost — financial, social, or structural — to breaking it. Then leave the rest of future-self alone.

How do I set up a pre-commitment device that actually sticks?

You start from the failure, not the goal.

Most well-meaning pre-commitments fail because they are built downstream of a wish ("I want to write more") rather than upstream of a known defection point ("I do not write after 4 p.m. on days when X"). The wish-shaped device binds the wrong moments, and the cost falls on the steady self instead of the hot one.

Three moves, in order:

1. Find the predictable failure. Watch for one or two weeks. When, specifically, does future-self defect from what cold-state-self chose? What state is the defection in? What was the proximal cue? Name the moment with enough detail that you could write down the time of day and the trigger.
2. Design narrowly. Build the smallest device that binds that specific moment. A blocker for those hours. A transfer on that day. A friend who expects you at that slot. Wider devices generate residue without adding deposit.
3. Make the cost real. The forfeit that does not sting is not a forfeit. The witness who would not actually notice is not a witness. The default that you can re-open in one click is not a default. The cost is the mechanism.

Practical steps

1. Spend one week observing before you design. Pre-commitment built without honest pattern recognition is mostly self-blame in a sturdier costume.
2. Default-set wherever possible before adding forfeit-set. Automatic transfers, scheduled deliveries, pre-installed blockers — friction that runs without anyone choosing — is cheaper and quieter than forfeits and binds the hot self without bruising the cold one.
3. When a forfeit is needed, set it just high enough to wince. Too low and the device is theatre. Too high and you will spend the residue arguing with the design.
4. Use social pre-commitment where the social cost is real. A friend who will actually mind is a stronger device than a public post that will scroll away.
5. Sunset devices that have done their work. A pre-commitment that has trained the pattern can usually be retired. Carrying every device forever is how the instrument turns into self-distrust.
6. If you notice frequent bitterness toward your own devices, redesign rather than add more. The residue is the equation's report that the design has drifted from the failure it was meant to bind.

Reflection questions

• Where in your life does future-self predictably differ from present-self? Can you name the moment, the cue, the state?
• Which of your current pre-commitments are matched to a real failure, and which are general self-management?
• Where have you been asking willpower to do work that structure could have done more cheaply?
• Is there a pre-commitment you should have retired by now? What is keeping it in place?

Frequently Asked Questions