0.7 Parcheesi Assignment

Overview

Should We Trust the AI Move?

AI tools can generate useful recommendations, but generated output is not automatically correct and is not evidence by itself. In this assignment, you will evaluate one AI-style Parcheesi recommendation and create a UML activity diagram showing how an AI tool, a human decision-maker, and an authoritative source work together.

The final decision remains with the human. Your diagram will show the AI producing a recommendation, the player identifying claims that need checking, an authoritative rule providing evidence, and the player deciding whether the recommendation should be Accept, Revise, Verify Further, or Reject.

User Story

As a Parcheesi player, I want to evaluate an AI-recommended move before using it so that I do not assume a generated recommendation is valid.

Acceptance Criteria

Task List

  1. Read the Game Situation
  2. Read the AI Recommendation as Claims
  3. Check the Authoritative Rule
  4. Choose Among Four Human-Judgment Outcomes
  5. Plan the Three Responsibilities
  6. Create the Diagram in Violet
  7. Trace the Evidence Path
  8. Save, Submit, and Revise
AI Output Is a Starting Point

Evaluate whether the recommendation is supported by checked evidence, not whether it sounds confident.

  1. 1. Read the Game Situation

    Begin with a small game state. You do not need to model the whole board.

    Step 1 — Read the supplied object diagram

    Object diagram showing redPawn on cell 2 and bluePawn on cell 5.
    Starting situation: Red is on Cell 2 and Blue is on Cell 5.

    Step 2 — Add the die roll

    For this situation, Red rolls a 3.

    1. Start at Cell 2
    2. Count three cells forward: 3, 4, 5
    3. Notice that Blue is already on Cell 5
    This Is the Situation the AI Will Discuss

    Treat the generated statements below as claims to evaluate, not as facts merely because they were generated.

    Go to top

  2. 2. Read the AI Recommendation as Claims

    Evaluate this AI-style recommendation:

    Move the Red pawn three spaces to Cell 5. The Blue pawn should be sent to Start. This is always the best move.

    Step 1 — Separate the recommendation into claims

    1. Where does the AI say Red should move?
    2. What does the AI say should happen to Blue?
    3. What does always the best move claim?

    Step 2 — Decide which claims require evidence

    Ask: What source could actually support or contradict this statement?

    Confidence Is Not Evidence

    Words such as should and always can sound certain. Certainty in wording does not prove a claim.

    Go to top

  3. 3. Check the Authoritative Rule

    For this course example, use the supplied rule card as the authoritative source for the simplified landing behavior.

    Simplified Course Rule

    A pawn moves forward the number of cells rolled. If a pawn lands on a cell already occupied by a pawn of another color, move the existing pawn to its Start cell.

    Step 1 — Check the destination claim

    1. Red begins on Cell 2
    2. The roll is 3
    3. Determine whether Cell 5 is supported

    Step 2 — Check the Blue-pawn claim

    1. Blue is already on Cell 5
    2. Red would land on Cell 5
    3. Determine what the supplied rule supports

    Step 3 — Check “always best” separately

    Ask whether the supplied rule says anything about the best strategic move in every possible situation.

    Evidence Can Support Only Part of an Answer

    Some claims may be supported while another claim remains unsupported. That is why Revise and Verify Further are useful choices.

    Go to top

  4. 4. Choose Among Four Human-Judgment Outcomes

    OutcomeUse it when...
    AcceptThe relevant claims are adequately supported for the current purpose
    ReviseUseful parts can be kept, but unsupported or inaccurate parts need correction
    Verify FurtherYou do not yet have enough evidence
    RejectThe recommendation is contradicted or not sufficiently supported

    Step 1 — Review a small decision pattern

    Generic activity diagram that asks whether more checking is needed before choosing an evidence-supported response.
    Verification may lead to more checking before a final response is chosen.

    Step 2 — Keep the human responsible

    The AI generates the recommendation. The rule provides evidence. The Player compares the evidence and chooses the final response.

    Go to top

  5. 5. Plan the Three Responsibilities

    Step 1 — Read a generic responsibility flow

    Activity diagram showing AI Tool generating a recommendation, Player identifying a claim, Authoritative Source providing evidence, and Player comparing evidence and making the final judgment.
    AI generates output; a human checks claims against evidence and makes the final judgment.

    Step 2 — Translate the roles to this assignment

    PartitionResponsibility
    AI ToolProduce the move recommendation
    PlayerIdentify claims, compare evidence, and make the judgment
    Parcheesi RulesProvide the authoritative rule evidence for this course example
    The Player Appears More Than Once

    The flow can leave Player to check a rule and return to Player for comparison and judgment.

    Go to top

  6. 6. Create the AI-and-Human-Judgment Diagram in Violet

    This is the main diagram you will submit.

    Step 1 — Create the three named partitions

    1. AI Tool
    2. Player
    3. Parcheesi Rules

    Step 2 — Use this partial diagram as a scaffold

    Partial activity diagram showing AI Tool recommendation, Player claim identification, Parcheesi Rules evidence checking, and Player evidence comparison. The learner must add the judgment decision and four outcomes.
    The scaffold reaches evidence comparison. You must add the judgment decision and outcome paths.

    Step 3 — Add the human judgment decision

    After Player compares the recommendation with the evidence, add a decision that determines what should happen to the AI recommendation.

    Step 4 — Add the four available outcomes

    1. Add a path for Accept
    2. Add a path for Revise
    3. Add a path for Verify Further
    4. Add a path for Reject
    5. Keep all judgment actions in the Player partition

    Check Your Work

    • AI Tool, Player, and Parcheesi Rules are represented
    • The AI does not make the final judgment
    • The Player identifies a claim
    • Parcheesi Rules supplies evidence
    • The Player compares evidence
    • A decision follows the evidence check
    • All four outcomes are available
    • The Player owns the final outcome
    • The diagram fits on one Violet screen
    Go to top

  7. 7. Trace the Evidence Path

    Step 1 — Start with AI output

    1. AI Tool produces the recommendation
    2. Player identifies its separate claims

    Step 2 — Follow the evidence check

    1. Move to Parcheesi Rules
    2. Use the simplified rule as evidence
    3. Return to Player
    4. Compare each important claim with what the evidence supports

    Step 3 — Follow your selected outcome

    Choose the outcome your evidence supports and be prepared to explain why. Refer to the supplied rule rather than to the AI's confidence or wording.

    Checkpoint Summary

    You modeled AI output as information that requires human review, verified claims against an authoritative source, and kept the final Accept/Revise/Verify Further/Reject judgment with the human decision-maker.

    Go to top

  8. 8. Save, Submit, and Revise the Diagram

    Use the same concise submission process you practiced in the earlier Parcheesi assignments.

    © 2026 Northcentral Technical College

    Go to top