OpenAnalyst Commitment Contracts (Punishment System)
Core Psychology
Loss Aversion: People are 2-3x more motivated by losing something than gaining something.
Research shows: Users who put financial stakes on commitments are 3x more likely to succeed.
How It Works
1. Contract Creation (During Challenge Onboarding)
code
Coach: "Now let's add some real stakes. What happens if you miss a day?"
User: "I'll send $100 to my girlfriend"
Coach: "Perfect. Let me confirm:
📜 COMMITMENT CONTRACT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Challenge: Learn Python in 30 Days
Commitment: Daily 1-hour session
PUNISHMENT TERMS:
• Trigger: Missing a scheduled session
• Consequence: Send $100 to girlfriend
• Recipient: [Girlfriend's name/contact]
• Grace Period: 24 hours to make up
REFEREE:
• Name: [Girlfriend's name]
• Role: Confirms punishment was received
Do you agree to these terms?"
User: "Yes"
Coach: "Contract signed. I've notified your referee.
She'll receive updates on your progress and confirm
if punishment is needed.
Day 1 starts now. Make us proud. 💪"
2. Daily Tracking
Claude (the brain) automatically tracks:
- •Did user check in today?
- •Did user complete their commitment?
- •Is this within grace period?
- •Should punishment trigger?
3. Punishment Trigger
code
[User missed Day 5, no check-in, grace period expired]
Coach: "Alex, we need to talk.
Yesterday you missed your Python session.
I waited 24 hours for a makeup or explanation.
Nothing.
According to your commitment contract:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
❌ Day 5: MISSED
💸 Punishment: Send $100 to [Girlfriend]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
I've already notified [Girlfriend] that the
punishment is due. She's been sent the contract
you signed.
Your move:
A) Send the $100 now and confirm
B) [Girlfriend] confirms she received it
C) Appeal (requires valid emergency proof)
Your streak: RESET to 0
Next session: Today 8pm
This isn't punishment for punishment's sake.
It's the cost of breaking your word.
Ready to restart?"
Punishment Types
1. Financial Punishments
markdown
- Send money to friend/family - Donate to charity - Donate to ANTI-charity (cause you hate) - Pay into savings (forced saving)
2. Social Punishments
markdown
- Referee gets notified of failure - Public post on social media - Tell X people about the miss - Embarrassing task
3. Forfeit Punishments
markdown
- Lose a privilege (no Netflix for a week) - Give away something (video game time) - Skip something fun (Friday drinks)
4. Escalating Stakes
markdown
First miss: $20 Second miss: $50 Third miss: $100 Fourth miss: $200 + public confession
Contract Storage
~/.openanalyst/contracts/active-contract.json
json
{
"contract_id": "ctr_abc123",
"challenge_id": "learn-python-30",
"created_at": "2025-12-26T10:00:00Z",
"signed_at": "2025-12-26T10:05:00Z",
"commitment": {
"description": "Complete 1 hour of Python daily",
"frequency": "daily",
"start_date": "2025-12-26",
"end_date": "2026-01-25"
},
"punishment": {
"type": "financial",
"amount": 100,
"currency": "USD",
"recipient": {
"name": "Sarah",
"relationship": "girlfriend",
"contact": "sarah@email.com"
},
"method": "venmo"
},
"referee": {
"name": "Sarah",
"email": "sarah@email.com",
"notified": true
},
"escalation": {
"enabled": true,
"levels": [
{"miss_count": 1, "amount": 100},
{"miss_count": 2, "amount": 150},
{"miss_count": 3, "amount": 200}
]
},
"grace_period_hours": 24,
"history": []
}
~/.openanalyst/contracts/punishment-history.md
markdown
# Punishment History ## Active Contract: Learn Python in 30 Days ### Miss #1 - December 30, 2025 - **Trigger:** Missed Day 5, no makeup within 24 hours - **Punishment:** $100 to Sarah - **Status:** ✅ Paid and confirmed by referee - **User Note:** "I accept. Won't happen again." ### Appeals None filed. ## Statistics - Total Misses: 1 - Total Paid: $100 - Punishment Success Rate: 100%
Referee System
Referee Role
- •Gets notified when contract is signed
- •Receives progress updates (weekly)
- •Gets alerted when punishment triggers
- •Confirms punishment was received
- •Can verify legitimate emergencies
Referee Notifications
On Contract Sign:
code
Subject: You're Alex's Accountability Referee Hi Sarah, Alex has committed to learning Python for 30 days and chosen you as their referee. Here's what that means: CONTRACT DETAILS: - Challenge: Learn Python in 30 Days - Daily commitment: 1 hour of study - If missed: Alex sends you $100 YOUR ROLE: - You'll get weekly progress updates - If Alex misses, you'll be notified - Confirm when punishment is received Alex is trusting you to hold them accountable. No letting them off easy! 😄 View contract: [link]
On Punishment Trigger:
code
Subject: 🚨 Alex missed their commitment Hi Sarah, Alex missed their Python session yesterday. Per their commitment contract, they owe you $100. DETAILS: - Missed: December 30, 2025 - Commitment: 1 hour Python study - Punishment: $100 to you Please confirm when you receive the payment. [Confirm Payment Received] [Alex Appealed - Review]
Anti-Charity Option
For extra motivation, users can donate to causes they oppose:
code
Coach: "Want to make the stakes even higher?
Instead of sending money to someone you like,
you can set it to donate to an anti-charity -
a cause or organization you really don't support.
For example:
- Political party you oppose
- Rival sports team's charity
- Organization with values you disagree with
Studies show this is VERY effective motivation.
Interested?"
Integration Points
- •Claude Brain: Monitors all check-ins, triggers punishments
- •Streak Skill: Resets streak on missed day
- •Daily Check-in: Warns when approaching deadline
- •Schedule Replanner: Legitimate replans avoid punishment
- •Motivation Generator: Extra motivation when stakes are high
Appeal System
Users can appeal with valid emergencies:
code
User: "I need to appeal yesterday's miss"
Coach: "Tell me what happened."
User: "I was in the hospital with food poisoning"
Coach: "That's a valid emergency. Can you provide any proof?
- Hospital discharge papers
- Doctor's note
- Photo evidence
If verified, I'll void this punishment and pause
your challenge for recovery."
Safety Features
- •24-hour grace period - Always a chance to make up
- •Appeal system - Legitimate emergencies respected
- •Max punishment cap - Won't bankrupt anyone
- •Referee verification - Human check on the system
- •Pause option - Extended breaks don't trigger
Sources
This system is inspired by research and apps that have proven this approach works:
- •StickK - Yale economists' commitment contract platform
- •Beeminder - Data-driven accountability with money stakes
- •ActionBuddy - Human accountability partnerships