UCP Calculation for Planned Work

Calculate Use Case Points for software development projects using Onshore's v4.0 complete work coverage methodology.

Quick Start

Minimal invocation:

"Calculate UCP for the Agent Studio Pro project. Input: git repository. Output: comprehensive UCP report."

With Azure DevOps:

"Calculate UCP from Azure DevOps User Stories for the Lighthouse project. Use references/azure-devops-hierarchy.md for work item guidance."

Using scripts:

python scripts/count_transactions.py "1. User enters credentials
2. System validates
3. System generates JWT
4. System returns token"
# Output: 3.6 transactions � Simple � 5 UUCW

python scripts/calculate_ucp.py --uaw 24 --uucw 562.4 --tcf 1.40 --ecf 0.86 --verbose
# Output: Realistic UCP: 494.22

Core Workflow

Step 1: Determine Input Source

Option A: Azure DevOps Work Items

1. •Check work item types available (Epic/Feature/User Story/Task)
2. •Prefer User Stories (95-98% accuracy)
3. •If only Features: apply 1.8x calibration factor
4. •Read references/azure-devops-hierarchy.md for guidance

Option B: Git Repository

1. •Analyze codebase structure
2. •Identify use cases from code (API endpoints, services, features)
3. •Count transactions per use case
4. •Read references/git-repository-analysis.md for methodology

Step 2: Identify Actors

Count external entities that interact with the system:

• •Simple (Weight 1): API/SDK (standard protocols)
• •Average (Weight 2): Interactive UI, automated systems
• •Complex (Weight 3): Custom protocols, admin dashboards

Calculate UAW (Unadjusted Actor Weight):

UAW = (Simple � 1) + (Average � 2) + (Complex � 3)

Step 3: Count and Classify Use Cases

v4.0 Complete Coverage - Count ALL work across 4 categories:

Category 1: New Functionality (1.0x multiplier)

• •User stories, features delivering new capabilities
• •Count transactions from acceptance criteria
• •Use scripts/count_transactions.py for automation

Category 2: Enhancements (0.6x multiplier)

• •Tasks, refactoring, technical improvements
• •Implementation without full scope

Category 3: Maintenance (0.4x multiplier)

• •Bugs, defects, corrections
• •Apply urgency multipliers: Normal 1.0x, High 1.2x, Critical 1.5x

Category 4: Quality Assurance (0.3x multiplier)

• •Test cases, test suites, QA work

Transaction Counting Rules:

• •Framework/SDK operations: 0.5 weight (FastAPI routes, ORM queries, API calls)
• •Custom business logic: 1.0 weight (workflows, algorithms, state management)

Classification:

• •Simple: d3 transactions � 5 UUCW
• •Average: 4-7 transactions � 10 UUCW
• •Complex: >7 transactions � 15 UUCW

Calculate UUCW per category:

Final UUCW = Base UUCW � Category Multiplier � (0.7 + 0.3 � SAF)

SAF (Specificity Adjustment Factor): Assess work item quality (0-1 scale)

• •Read references/saf-methodology.md for detailed rubric
• •5 dimensions: Clarity, Acceptance Criteria, Technical Detail, Scope, Dependencies

Step 4: Calculate UUCP

UUCP = UAW + Total UUCW (sum of all 4 categories)

Step 5: Assess Technical Complexity (TCF)

Evaluate 13-20 technical factors (ratings 0-5):

• •Standard factors (T1-T13): Distributed system, performance, security, etc.
• •AI/ML factors (T14-T20): Model complexity, data volume, explainability, etc.

Read references/tcf-factors.md for complete factor definitions and rating criteria.

TFactor = �(Rating � Weight)
TCF = 0.6 + (0.01 � TFactor)
TCF (capped) = min(TCF, 1.4)

Use scripts/calculate_ucp.py to automate:

python scripts/calculate_ucp.py --uaw 24 --uucw 562.4 --tcf 1.40 --ecf 0.86

Step 6: Assess Environmental Complexity (ECF)

Evaluate 8 environmental factors (ratings 0-5):

• •E1-E6: Favorable when HIGH (experience, motivation, stable requirements)
• •E7-E8: Unfavorable when HIGH (part-time staff, difficult language)

EFactor = �(Rating � Weight)
ECF = 1.4 + (-0.03 � EFactor)

Step 7: Calculate Final UCP

Adjusted UCP = UUCP � TCF � ECF

Realistic UCP = Adjusted UCP � Framework Leverage Factor
(Framework Leverage: 0.6-1.0, typically 0.7 for framework-heavy projects)

Step 8: Generate Report

Use assets/templates/ucp-report-template.md as structure.

Include:

1. •Executive Summary (all metrics)
2. •Actor Summary Table
3. •Work Item Analysis (all 4 categories)
4. •SAF Distribution
5. •UUCP Calculation
6. •TCF Assessment (with TFactor breakdown)
7. •ECF Assessment (with EFactor breakdown)
8. •UCP Calculation
9. •Effort Estimation (productivity factor: 18-22 hours/UCP typical)

Key Concepts

v4.0 Complete Work Coverage: Unlike traditional UCP (50-60% coverage), v4.0 captures ALL development work across 4 categories with differentiated multipliers.

Specificity Adjustment Factor (SAF): Rewards high-quality specifications, reduces credit for vague requirements. Ranges 0.70x (no spec) to 1.00x (excellent spec).

Framework Leverage: Accounts for framework-provided functionality (FastAPI routing, SQLAlchemy ORM, etc.). Typically 0.7 for modern frameworks.

Azure DevOps Hierarchy:

• •Epic (Strategic) � Feature (Functional) � User Story (Use Case ) � Task (Implementation)
• •Count User Stories for UCP, not Epics or Tasks
• •If only Features available: apply 1.8x calibration

Transaction Weighting:

• •0.5 for framework operations (HTTP routing, ORM queries, SDK calls)
• •1.0 for custom logic (business rules, workflows, algorithms)

Available Resources

Scripts

• •

  scripts/count_transactions.py  Parse acceptance criteria, count weighted transactions, classify complexity

  bash

  python scripts/count_transactions.py --file user_story.txt --verbose
  

• •

  scripts/calculate_ucp.py  Calculate UCP from UAW, UUCW, TCF, ECF

  bash

  python scripts/calculate_ucp.py --uaw 24 --uucw 562.4 --tcf 1.40 --ecf 0.86 --framework-leverage 0.70
  

References

• •references/tcf-factors.md  Complete TCF factor definitions (T1-T20), rating criteria, examples
• •references/saf-methodology.md  SAF 5-dimension rubric, impact calculations, quality assessment
• •references/azure-devops-hierarchy.md  Work item hierarchy, calibration factors, decision matrix
• •references/git-repository-analysis.md  Code-based UCP analysis methodology

Assets

• •assets/templates/ucp-report-template.md  Complete report structure for filling in calculated values

Output Format

Generate comprehensive UCP Analysis Report with:

Section 1: Executive Summary (key metrics table) Section 2: Actor Summary (classification table, UAW calculation) Section 3: Work Item Analysis (4 categories, transaction counts, SAF scores, UUCW calculations) Section 4: SAF Distribution (quality assessment) Section 5: UUCP Calculation (UAW + UUCW) Section 6: TCF Assessment (factor ratings, TFactor, TCF formula) Section 7: ECF Assessment (factor ratings, EFactor, ECF formula) Section 8: UCP Calculation (Adjusted UCP, Framework Leverage, Realistic UCP) Section 9: Effort Estimation (productivity factor, estimated hours)

Include interpretations: What TCF/ECF values mean, accuracy notes, comparison to traditional UCP (v3.0 vs v4.0).

Decision Matrix for Input Sources

Common Patterns

Pattern 1: Azure DevOps with User Stories

1. •Export User Stories with acceptance criteria
2. •Count transactions per story using scripts/count_transactions.py
3. •Calculate SAF for each story (typically 0.65-0.85)
4. •Calculate UUCW per category
5. •Assess actors, TCF, ECF
6. •Generate final UCP report

Pattern 2: Azure DevOps with Features Only

1. •Read references/azure-devops-hierarchy.md
2. •Extract Key Deliverables from Features
3. •Estimate transactions per deliverable
4. •Apply 1.8x calibration factor
5. •Add 25-30% for untracked infrastructure
6. •Note �20% margin of error in report

Pattern 3: Git Repository Analysis

1. •Read references/git-repository-analysis.md
2. •Analyze codebase structure (API routes, services, components)
3. •Identify use cases from code organization
4. •Count transactions (framework 0.5x, custom 1.0x)
5. •Calculate UUCW with 100% coverage
6. •Most accurate result (�5-10%)

Important Reminders

1. •Always use User Stories over Features  95% accuracy vs 80% accuracy
2. •v4.0 counts ALL work  Stories, Tasks, Bugs, Tests (not just stories)
3. •SAF rewards quality  Well-specified work gets full credit (1.0x), vague work gets reduced credit (0.7x minimum)
4. •Check for missing categories  Infrastructure, monitoring, logging, deployment often untracked
5. •Framework leverage is real  Modern frameworks provide 60-80% of functionality, account for it
6. •TCF typically 0.95-1.15  Values >1.30 indicate exceptional complexity
7. •ECF typically 0.70-0.90  Favorable teams ~0.85, challenging environments ~0.70
8. •Productivity factor 18-22 hrs/UCP  For moderate challenges, adjust based on ECF

Token Efficiency Notes

• •Scripts handle deterministic calculations (transaction counting, UCP math)
• •References contain detailed rubrics (TCF 20 factors, SAF 5 dimensions)
• •Template provides report structure (fill-in-the-blanks)
• •SKILL.md focuses on workflow and decision-making (when to use what)

This progressive disclosure keeps core instructions concise while making detailed reference material available when needed.