Prevention Architecture

#system-architecture #core-principle

What It Is

Prevention architecture means removing the decision entirely instead of resisting it repeatedly.

The core insight: Prevention costs 0 willpower units. Resistance costs 2-3 units per instance.

The Math

With temptation accessible:

  • Donut sitting on desk
  • See it 20 times per day
  • Resist each time: 2 units × 20 = 40 willpower units/day
  • Eventually you give in (run out of units)

With temptation removed:

  • Don't buy donuts
  • Never see it
  • Never decide
  • Cost: 0 willpower units/day

Same outcome (not eating donut), massively different cost.

This isn't moral superiority. This is architecture.

Why Resistance Fails

Every act of resistance depletes your capacity for subsequent resistance.

Your prefrontal cortex runs on glucose and neurotransmitters. After multiple inhibition tasks, it's literally running low on metabolic resources.

You're not morally weak for giving in at 8 PM. Your decision-making hardware is running on fumes.

The self-control research is clear: it operates like a muscle that fatigues with use. Each resistance depletes the resource.

High Self-Control People Avoid Situations Requiring Self-Control

They don't keep junk food in the house.

They don't scroll social media because they never installed the apps.

They don't fight temptation heroically. They engineer environments where temptation doesn't exist.

They're not stronger. They're smarter about architecture.

Examples of Prevention Architecture

Phone Management

Resistance approach:

  • Phone on desk during work
  • Try to ignore it
  • Resist checking when you see notifications
  • Cost: 2-3 units every time you notice it
  • Failure rate: high (eventually you check)

Prevention approach:

  • Phone in locked drawer in different room
  • No notifications possible
  • Can't check even if you wanted to (4-unit cost to retrieve)
  • Cost: 0 units during work
  • Failure rate: near zero

Food Environment

Resistance approach:

  • Buy snacks "for guests"
  • Keep them in pantry
  • Try to resist eating them
  • Cost: 2-3 units per exposure
  • Failure rate: high

Prevention approach:

  • Don't buy snacks
  • Nothing in pantry to resist
  • Cost: 0 units
  • Failure rate: zero (can't eat what isn't there)

Digital Distractions

Resistance approach:

  • Netflix tab open "just in case"
  • Try to resist clicking
  • Fight the urge repeatedly
  • Cost: 2-3 units per temptation
  • Failure rate: moderate

Prevention approach:

  • Website blocker active during work hours
  • Netflix literally inaccessible
  • Cost: 0 units
  • Failure rate: zero (can't access what's blocked)

State Machine Perspective

Prevention means removing state transitions from the decision tree.

Old state machine:

  • work_state + phone visible → distraction_state (high probability)
  • Must actively resist transition (expensive)

Engineered state machine:

  • work_state + phone absent → transition to distraction_state doesn't exist
  • Can't transition to a state if the trigger is removed

The distraction_state option is literally not in the decision tree anymore. You're not resisting. The choice doesn't exist.

The Boltzmann Distribution

From statistical mechanics: systems naturally fall into low-energy configurations.

The behavior that executes is whichever has the lowest activation energy at decision time.

Phone on desk:

  • Checking: 0.1 units
  • Resisting: 2 units
  • Boltzmann says: you'll check it

Phone in drawer:

  • Checking: 4 units (stand, walk, unlock, retrieve)
  • Continuing work: 0.5 units (already in work state)
  • Boltzmann says: you'll keep working

Same person. Different energy landscape. Different behavior. Not discipline. Thermodynamics.

Application Domains

Work Environment

Remove:

  • Phone from workspace
  • Distracting websites (block at router/app level)
  • Unnecessary browser tabs
  • Visual clutter that captures attention

Result: Focus becomes default because competing attention sources don't exist.

Food System

Remove:

  • Junk food from house
  • Delivery apps from phone
  • Decision points about meals (predetermined)

Result: Eating discipline becomes automatic because temptation isn't accessible.

Social Media

Remove:

  • Apps from phone entirely (not just logged out - deleted)
  • Browser extensions that block sites
  • Accounts deactivated if necessary

Result: Can't check what isn't accessible.

Time-Based Prevention

You can also use time-boxing to prevent decisions:

No outside breakfast rule:

  • Eliminate 90-minute deliberation about dim sum
  • No decision required (rule handles it)
  • Zero willpower spent on morning food decisions

Eating window closes at 2 PM:

  • After 2 PM, eating isn't an option
  • No decisions about dinner/snacks
  • Zero willpower spent resisting evening eating

Why This Works When Trying Harder Doesn't

Trying harder = using more willpower to resist = depleting a finite resource = inevitable failure

Prevention architecture = changing environment so resistance is unnecessary = sustainable indefinitely

You're not using moral effort to override impulses. You're removing the impulses from your environment.

The Planning Phase

Prevention requires upfront work:

  1. Identify temptation sources
  2. Remove them from environment
  3. Install blocking mechanisms
  4. Set up systems that maintain prevention

This feels like effort. But it's one-time effort that eliminates ongoing daily effort.

  • 1 hour to set up website blockers
  • Saves 2-3 units × 20 instances × 365 days = 14,600+ willpower units/year

The ROI is absurd.

Limitations

Prevention architecture works when:

  • Temptation is discretely removable (can be eliminated from environment)
  • You control your environment (can modify it)
  • The behavior is not socially required (can avoid without consequences)

Prevention fails when:

  • Temptation is everywhere and unavoidable
  • You don't control the environment
  • Social context requires exposure

In those cases, you need other strategies (habit building, reframing, etc.). But for 80% of cases, prevention is available and optimal.

Intervention Design Questions

Intervention design questions systematically generate prevention architecture solutions. Instead of asking "How do I have more willpower?" (which treats willpower as an input variable you can increase), ask "What forcing function prevents this?" This presupposes that mechanical solutions exist and searches for implementations rather than relying on resistance.

"What's the earliest intervention point?" triggers backward temporal traverse of the causal chain to find the first modifiable node. For late-night eating, the chain might be: eating occurs ← hunger signal ← boredom ← lounge state ← no gym ← woke late ← sleep time was 1am. The earliest intervention point is sleep time—fixing that cascades forward and prevents the entire chain. This is computationally superior to resisting at the final moment when willpower is already depleted.

"What category of forcing function applies?" classifies the failure mode and retrieves appropriate prevention patterns from the taxonomy. Accessibility-based prevention (remove from environment), temporal prevention (time-boxing), commitment devices (external accountability), or environmental redesign (change default options). The question structure guides systematic exploration of prevention mechanisms rather than ad-hoc guessing.

"How do I make this automatic?" identifies decision points in the current behavior and installs default scripts or prevention architecture to reduce willpower cost to zero. Manual resistance at each decision costs 2-3 units repeatedly. Automating through prevention costs 0 units after one-time setup. The question forces focus on architecture rather than effort.

  • Willpower - The resource you conserve through prevention
  • Self-Control - Prevention as primary strategy
  • Discipline - What it looks like when prevention is working
  • State Machines - Removing transitions from decision tree
  • Activation Energy - Prevention makes undesired behaviors high-cost
  • Focus - Achieved through removing distraction sources
  • Question Theory - Intervention design questions generate forcing functions
  • Epistemic Contamination - Same prevention principle for information exposure—quarantine sources rather than resisting influence

Key Principle

Remove decisions, don't resist them - Engineer environments so desired behavior is the only option, rather than continuously fighting temptation with finite willpower.

The best way to resist temptation is to never encounter it. Prevention costs nothing. Resistance costs everything.