Hypothesis-Elimination (HE)

Find the root CAUSE of issues through systematic hypothesis testing and elimination.

When to Use

• •Debugging complex issues
• •Root cause analysis
• •Diagnostic problems
• •Evidence can be gathered to test hypotheses
• •Multiple possible causes exist

Process

Phase 1: Hypothesis Generation

Generate 5-8 possible causes:

Symptom: [observed problem]

H1: [possible cause 1]
H2: [possible cause 2]
H3: [possible cause 3]
H4: [possible cause 4]
H5: [possible cause 5]

Phase 2: Test Design

For each hypothesis, design a test:

Phase 3: Evidence Gathering

Execute tests in order of:

1. •Easiest to perform
2. •Most discriminating (eliminates most hypotheses)
3. •Lowest risk

Phase 4: Elimination

After each test, update hypothesis status:

• •ELIMINATED: Evidence contradicts
• •SUPPORTED: Evidence consistent
• •INCONCLUSIVE: Need more data

Phase 5: Convergence

Continue until:

• •One hypothesis remains (root cause found)
• •Multiple remain (need more tests)
• •All eliminated (missing hypothesis - regenerate)

Output Template

## HE Analysis: [Problem]

### Symptom
[Detailed description of observed issue]

### Hypotheses
| ID | Hypothesis | Prior Confidence |
|----|------------|------------------|
| H1 | [cause] | X% |
| H2 | [cause] | X% |
| H3 | [cause] | X% |

### Tests Performed

#### Test 1: [description]
- **Result**: [observation]
- **Eliminates**: H2, H4
- **Supports**: H1, H3

#### Test 2: [description]
- **Result**: [observation]
- **Eliminates**: H3
- **Supports**: H1

### Elimination Matrix
| Hypothesis | Test 1 | Test 2 | Test 3 | Status |
|------------|--------|--------|--------|--------|
| H1 | ✓ | ✓ | ✓ | **ROOT CAUSE** |
| H2 | ✗ | - | - | Eliminated |
| H3 | ✓ | ✗ | - | Eliminated |
| H4 | ✗ | - | - | Eliminated |

### Root Cause
**H1: [description]**

### Evidence Summary
- [evidence 1 supporting H1]
- [evidence 2 supporting H1]

### Recommended Fix
[action to address root cause]

Example

Symptom: API requests failing intermittently

Hypotheses:

• •H1: Database connection pool exhaustion
• •H2: Memory leak causing OOM
• •H3: Network timeouts to external service
• •H4: Rate limiting by upstream API
• •H5: DNS resolution failures

Tests:

1. •Check connection pool metrics → H1 supported (pool at 100%)
2. •Check memory graphs → H2 eliminated (memory stable)
3. •Check external service logs → H3 eliminated (all succeed)
4. •Check rate limit headers → H4 eliminated (under limit)

Root Cause: H1 - Database connection pool exhaustion Fix: Increase pool size, add connection timeout