Schedule Comparison & Scoring Manual

This document details the mechanics, logic, and scoring formulas used by the compare command in the ScheduleGate.

Overview

The comparison engine ("Delta Engine") takes two schedule files—a Previous version (Baseline) and a Current version—and benchmarks them to quantify stability and reliability.

Command Reference

./schedulegate compare <previous_file> <current_file> [flags]

Flag	Description
--html <path>	Generate an HTML comparison report (auto-opens in browser)
--csv <path>	Append comparison results to a CSV history database
--detailed	Include row-by-row task-level change detail in the HTML report
--customer <name>	Customer name for report metadata
--project <id>	Project ID for report metadata
--pct-format <format>	Percent complete scale: "0-100" (MS Project default) or "fraction" (0.0–1.0, Primavera exports)
--date-locale <locale>	Date parsing priority for ambiguous dates: "US" (MM/DD first, default) or "EU" (DD/MM first)

1. Stability Score

The Stability Score (0-100) is a composite index derived from three weighted pillars.

Pillar A: Schedule Stability (40% Weight)

Measures how well the schedule holds its dates.

• Metric: Finish Variance (Difference in Finish Date between versions).
• Penalty Condition: A task is penalized if its Finish Variance is > 2.0 days.
• Weighting:
  • Standard Task: 1.0 point.
  • Milestone: 2.0 points (Milestone slippage is weighted double).
• Scoring Formula:
  • Penalty = (Weighted Count of Slipping Tasks / Total Tasks) * 100
  • Score = MAX(0, 40 - Penalty)
  • Interpretation: If 40% of your tasks slip by more than 2 days, you lose all 40 points in this pillar.

Pillar B: Duration Reliability (30% Weight)

Measures "Task Bloat"—the tendency for tasks to expand in duration rather than finishing.

• Metric: Duration Delta (Current Duration - Previous Duration).
• Penalty Condition: A task is penalized if its duration grows by > 10%.
  • Example: A 10-day task becoming 12 days (+20%) is penalized.
• Scoring Formula:
  • Bloat % = (Count of Bloated Tasks / Total Tasks) * 100
  • Penalty = Bloat % * 1.5
  • Score = MAX(0, 30 - Penalty)
  • Interpretation: This is a harsh penalty. If 20% of your tasks bloat, you lose all 30 points (20 * 1.5 = 30). This encourages realistic initial durations.

Pillar C: Scope Churn (30% Weight)

Measures the volatility of the scope (tasks added or removed).

• Metric: Task Churn (New Tasks + Deleted Tasks).
• Scoring Formula:
  • Churn % = ((New Tasks + Deleted Tasks) / Total Tasks) * 100
  • Penalty = Churn % * 2.0
  • Score = MAX(0, 30 - Penalty)
  • Interpretation: High churn indicates poor planning or changing requirements. 15% churn results in a 0 score for this pillar.

---

2. Friction Index ("Ghost Tasks")

The Friction Index identifies "Ghost Tasks"—tasks that are haunting the schedule.

• Definition: A Ghost Task is a task where:
  1. Planned Start Date is in the past (before the status/data date).
  2. % Complete is still 0%.
• Meaning: These tasks are "sliding" to the right. They should have started but haven't. They represent immediate execution bottlenecks.
• Botttleneck Identification: The CLI aggregates these Ghost Tasks by their top-level WBS (Work Breakdown Structure) to show you exactly which phase of the project is stuck.

3. Change Metrics

The CLI reports raw counts for context:

• New Tasks: Tasks present in Current but not Previous (matched by Task ID).
• Deleted Tasks: Tasks present in Previous but not Current.
• Modified Tasks: Tasks where Name, Duration, Percent Complete, or Dates have changed.
• Duration Bloat: The specific count of tasks that triggered the Pillar B penalty (>10% growth).

FAQ

Q: Why is my Reliability Score 0? A: This usually means >20% of your tasks have increased their duration. Note that if you act on "Ghost Tasks" by simply extending their duration instead of starting them, you will improve stability (dates might hold if you extend others) but you will likely crash your Reliability score.

Q: Does it handle renamed tasks? A: The comparison is based on Task ID — the value from the Task ID / ID column (the sequential row number in MS Project's left-hand column). If only a Unique ID column is present, that value is used instead. If you rename a task but keep the ID, it counts as "Modified". If you delete a task and add a new one with a new ID, it counts as "Deleted" + "New" (Churn).

4. CSV History Database (--csv)

Use --csv to append one summary row per comparison run to a flat CSV file (separate schema from the assess DCMA CSV). Pair with --customer and --project to track multiple programs across comparison cycles.

schedulegate compare prev.xlsx curr.xlsx \
  --customer Acme --project ST0410 --csv compare_history.csv \
  --date-locale US

• Append behavior: If the file does not exist, a header row is written first; each subsequent run appends a new row.
• One row = one comparison cycle: scores and change counts only (no task-level deltas).

Column	Description
Customer	From --customer
Project	From --project
Previous Schedule	Previous file basename (no extension)
Current Schedule	Current file basename (no extension)
Report Date	Timestamp when the row was written
Tool Version	CLI version string
Overall Score	Composite stability score (0–100)
Stability Score	Pillar A (max 40)
Reliability Score	Pillar B (max 30)
Scope Churn Score	Pillar C (max 30)
Total Tasks	Unique non-summary tasks across both versions
New Tasks / Deleted Tasks / Modified Tasks / Unchanged Tasks	Change counts
Ghost Tasks	Ghost task count
Duration Inflated Count / Duration Inflated Pct	Pillar B bloat metrics
Churn Pct	(New + Deleted) / Total Tasks * 100

5. Detailed Task Analysis (--detailed)

When using the --detailed flag with the --html option, the report includes a row-by-row breakdown of every significantly changed task.

Symbology Legend

Symbol	Meaning	Impact Type	Description
⊕	New Task	Scope Churn	Task was added in the new version.
×	Deleted	Scope Churn	Task was removed from the schedule.
☒	Delayed	Stability	Finish date slipped significantly (> 2 days).
←	Pulled In	Stability	Task is finishing earlier than planned.
🐢	Bloated	Reliability	Duration increased by > 10%.
👻	Ghost Task	Reliability	Planned start in past, 0% complete.
📝	Modified	Neutral	Minor changes (name, small date shifts).
□	Unchanged	Neutral	No significant changes.

How to Interpret the Table

The detailed table allows you to audit the score. If you see a lot of ☒ symbols, you know exactly which tasks are dragging down your Stability score. If you see many 🐢, you know Reliability is suffering due to duration inflation.