AreteDriver /estimate - Task Estimation Techniques
Structured approach to estimating development tasks.
Usage
/estimate "add user authentication" # Estimate a task
/estimate --breakdown # Detailed breakdown
/estimate --range # Provide ranges
/estimate --compare # Compare techniques
What This Skill Does
- Analyze Task - Understand scope, complexity
- Break Down - Decompose into subtasks
- Identify Risks - Unknowns, dependencies
- Apply Techniques - T-shirt, story points, hours
- Provide Ranges - Optimistic/realistic/pessimistic
Estimation Report Format
# Estimate: [Task Description]
## Task Analysis
### Scope
What's included in this task.
### Not Included
What's explicitly out of scope.
### Assumptions
- Assumption 1
- Assumption 2
### Dependencies
- Dependency 1
- Dependency 2
---
## Breakdown
| Subtask | Complexity | Estimate |
|---------|------------|----------|
| Design/planning | Low | S |
| Core implementation | Medium | M |
| Tests | Low | S |
| Documentation | Low | XS |
| Code review | Low | XS |
| **Total** | | **M-L** |
---
## Estimate
### T-Shirt Size
**M** (Medium)
| Size | Typical Range |
|------|---------------|
| XS | < 2 hours |
| S | 2-4 hours |
| M | 4-8 hours (1 day) |
| L | 1-3 days |
| XL | 3-5 days |
| XXL | 1-2 weeks |
### Three-Point Estimate
| Scenario | Estimate | Notes |
|----------|----------|-------|
| Optimistic | 4 hours | Everything goes smoothly |
| Realistic | 8 hours | Some minor issues |
| Pessimistic | 16 hours | Unexpected complications |
**Expected**: 8-10 hours (weighted average)
### Story Points
**5 points** (Fibonacci: 1, 2, 3, 5, 8, 13, 21)
Reference:
- 1 point = trivial change, < 1 hour
- 3 points = straightforward, half day
- 5 points = moderate complexity, 1 day
- 8 points = significant work, 2-3 days
- 13+ points = should be broken down
---
## Risk Factors
| Risk | Likelihood | Impact | Mitigation |
|------|------------|--------|------------|
| Unclear requirements | Medium | High | Clarify before starting |
| Integration issues | Low | Medium | Spike first |
| Learning curve | Low | Low | Time-box research |
### Confidence Level
**Medium** (70%)
Factors affecting confidence:
- [+] Well-understood domain
- [+] Similar work done before
- [-] External API integration
- [-] Some unknowns in requirements
---
## Recommendations
1. **Clarify before starting**: [specific questions]
2. **Spike if needed**: [area of uncertainty]
3. **Break down further**: [if XXL]
4. **Add buffer**: Include 20% for unknowns
Estimation Techniques
T-Shirt Sizing
Best for: Rough estimates, backlog grooming
| Size | Effort | Complexity | Risk |
|---|---|---|---|
| XS | Trivial | None | None |
| S | Small | Low | Low |
| M | Medium | Medium | Low-Med |
| L | Large | High | Medium |
| XL | Very Large | High | High |
| XXL | Epic | Very High | High |
Story Points (Fibonacci)
Best for: Sprint planning, relative sizing
1 - Trivial (< 1 hour)
2 - Simple (1-2 hours)
3 - Straightforward (half day)
5 - Moderate (1 day)
8 - Complex (2-3 days)
13 - Very complex (1 week)
21 - Epic (break it down!)
Three-Point Estimation
Best for: Schedule planning, risk assessment
Expected = (Optimistic + 4×Realistic + Pessimistic) / 6
Example:
- Optimistic: 2 days
- Realistic: 4 days
- Pessimistic: 10 days
- Expected: (2 + 16 + 10) / 6 = 4.7 days
PERT (with standard deviation)
Expected = (O + 4M + P) / 6
Std Dev = (P - O) / 6
95% confidence = Expected ± 2×StdDev
Common Estimation Mistakes
Underestimation
- Forgetting tests
- Forgetting code review time
- Forgetting documentation
- Ignoring integration time
- Optimism bias
Overestimation
- Adding too much buffer
- Accounting for same risk multiple times
- Not recognizing reusable components
General Mistakes
- Estimating without understanding scope
- Not breaking down large tasks
- Ignoring dependencies
- Treating estimates as commitments
- Not tracking actuals vs estimates
Estimation Checklist
Before estimating:
- [ ] Requirements are clear
- [ ] Scope is defined
- [ ] Dependencies identified
- [ ] Similar past work referenced
Include in estimate:
- [ ] Design/planning time
- [ ] Implementation
- [ ] Unit tests
- [ ] Integration tests
- [ ] Code review
- [ ] Documentation
- [ ] Bug fixes (buffer)
- [ ] Meetings/communication
After estimating:
- [ ] Confidence level stated
- [ ] Risks identified
- [ ] Assumptions documented
- [ ] Range provided (not single number)
Instructions for Claude
When /estimate is invoked:
- Understand the task - What's being estimated
- Clarify scope - What's in, what's out
- Break down - Decompose into subtasks
- Identify unknowns - Risks, dependencies
- Apply techniques - T-shirt, points, hours
- Provide range - Not a single number
- State confidence - How sure are we
- Document assumptions - What we're assuming
- Recommend actions - Clarify, spike, break down