Direct coherent tradeoffs

2415. Lesson identity

Module

5.15 — Creative direction

Lesson order

2

English lesson title

Direct coherent tradeoffs

Spanish lesson title

Dirigir decisiones coherentes entre alternativas

Academic type

Integration

Schema type

Practical

Estimated time

40–50 minutes

This lesson turns the player promise from the previous lesson into a decision filter. You will compare alternatives, make a deliberate choice, and record what you rejected and why.

2416. Learning objective

After this lesson, you can accept or reject competing game alternatives by applying the creative brief, the player promise, and player evidence, then justify the decision in a concise decision record.

2417. Why this matters

A creative direction is not a list of preferences. It is a repeatable way to decide what belongs in the game and what does not. AI can produce many plausible implementations, assets, and interactions; plausibility is not evidence of fit. A coherent project accepts alternatives that reinforce the intended player experience and rejects attractive alternatives that pull the game in another direction. This protects the player promise when scope, polish, and experimentation compete.

2418. Prior knowledge

You should already be able to state the player promise from Lesson 5.15 L1, distinguish the intended player experience from a feature list, and describe the relevant player evidence or test observation. You should also be able to identify the constraint being protected: tone, clarity, tension, pacing, agency, scope, or another criterion in the creative brief.

2419. Core concept

Coherence is a decision relationship, not a visual style.

An alternative is coherent when it strengthens the player promise under the project’s real constraints. An alternative is incoherent when it may be impressive or functional on its own but weakens, distracts from, or contradicts that promise.

Use three separate judgments:

1. Fit: Does this alternative reinforce the intended player experience?
2. Evidence: What observation, test result, or clearly stated requirement supports the choice?
3. Cost: What does accepting it consume in scope, clarity, performance, production time, or future flexibility?

A decision is not automatically good because it has the lowest cost. It is not automatically good because it is novel. It is good when its fit is strong enough, its evidence is credible enough, and its cost is justified by the promise it protects.

2420. Mental model

The coherent tradeoff test

Use the model in this order: promise → player change → evidence → cost → rejection. The final rejection matters because a direction becomes actionable only when the team knows which plausible alternative it will not pursue.

A concise decision record can use this format:

Choose: [alternative] Because: [player promise and observable effect] Evidence: [specific evidence or requirement] Cost accepted: [tradeoff] Reject: [alternative] Reason for rejection: [coherence, evidence, or cost problem]

2421. Concrete example

Suppose the player promise is: “The player should feel clever and exposed while making meaningful delivery decisions under pressure.” Two interaction alternatives are being considered for a delivery route:

• Alternative A: A route preview shows every patrol position and marks the safest path.
• Alternative B: The player gets a short, imperfect signal about nearby risk and must inspect the environment before committing.

In a hypothetical playtest, three of five players using Alternative A followed the marked path without inspecting the environment. With Alternative B, two of five players missed the signal and described the resulting danger as arbitrary, while two others inspected their surroundings and changed route after noticing it.

Alternative A may improve immediate clarity, but it removes much of the exposure and turns the decision into route selection from complete information. Alternative B preserves uncertainty and asks the player to interpret evidence, but it may be too obscure if the signal is difficult to read.

A coherent decision record might be:

Choose: Alternative B, with a clearer signal and a short explanation on first use. Because: It preserves the feeling of making a risky, informed delivery decision rather than selecting a solved route. Evidence: The brief requires cleverness and exposure. The observed player evidence is mixed: Alternative A discouraged inspection for three of five players, while Alternative B prompted inspection for two players but was missed by two others. This supports preserving Alternative B’s interpretation step while qualifying the choice with a clarity revision. Cost accepted: Additional clarity work and a more demanding tutorial explanation. Reject: Alternative A. Reason for rejection: Complete route information weakens uncertainty and reduces the significance of the player’s judgment.

The choice is not “minimal interface versus helpful interface.” It is a tradeoff between complete certainty and meaningful interpretation.

2422. AI-native workflow

Use AI to expand and interrogate the decision, not to outsource the decision itself.

1. Give the AI the player promise, the relevant creative constraint, and the alternatives. Do not ask it to choose immediately.
2. Ask it to describe how each alternative changes player behavior, clarity, tension, and scope.
3. Ask it to identify assumptions and missing evidence for each alternative.
4. Compare the response against the brief and actual player evidence. Mark observations as supported, inferred, or unknown.
5. Make the decision yourself and ask the AI to challenge the rationale by presenting the strongest case for the rejected option.
6. Revise the decision record only when the challenge reveals a real contradiction, missing evidence, or unacknowledged cost.

A useful prompt is:

“Here is the player promise, the creative constraint, and two alternatives. Compare their effects on player behavior and experience. Separate evidence from inference, identify the largest cost of each, and argue for the strongest rejected option. Do not make the final decision.”

Do not treat fluent AI analysis as player evidence. It is a hypothesis generator and critique partner; the brief and observed player response remain the decision sources.

2423. Common mistake

Mistake: choosing the most feature-rich or visually impressive option and calling the result creative direction.

This confuses local quality with project coherence. A richer option can add noise, dilute the central interaction, increase learning burden, or consume time needed for a more important part of the player experience. If you cannot state what the option protects and what you are willing to give up for it, you have not made a tradeoff yet.

2424. Guided practice

Create a decision record for the following case.

The player promise is: “The player should feel clever and exposed while making meaningful delivery decisions under pressure.” The team is choosing between two presentation approaches for a risky route:

• Option A — Full tactical overlay: The screen displays route safety, known hazards, estimated travel time, and a recommended path.
• Option B — Sparse situational cues: The screen displays only a recent warning and a rough direction; the player must combine that information with what they observe in the world.

Use these supplied case notes:

• In a five-player trial of Option A, three players immediately followed the recommended path without inspecting the environment.
• In a five-player trial of Option B, two players missed the warning and later described the danger as arbitrary.
• In that same trial, two other players paused, inspected the environment, and changed route after noticing the warning.
• The team concludes that Option B creates more meaningful decisions. This is an interpretation, not a direct observation.

Treat the player promise as a requirement and the numbered trial results as limited observations from this hypothetical case.

Copy and complete this text-labeled worksheet; do not use color alone to distinguish evidence states:

Complete this scored practical for 10 points:

1. Classify the player promise, each trial result, and the team conclusion as a requirement, observation, or inference. (2 points)
2. For each trial result, state what decision question it is relevant to and one limitation of the evidence. (2 points)
3. State the player behavior each option encourages and which part of the promise it strengthens or weakens. (1 point)
4. Make a provisional choice. Cite at least two supplied observations and explain whether each one changes the choice or merely qualifies it. (2 points)
5. Complete the decision-record format, including the largest accepted cost and a specific reason for rejecting the competing option. (2 points)
6. State one condition under which the rejected option should be reconsidered. (1 point)

A response earns full credit only when its choice follows from the classified evidence rather than treating the brief, an inference, or a small trial result as conclusive on its own.

2425. Validation / evidence

Your work is sufficient when it contains all of the following:

• One accepted alternative and one explicitly rejected alternative.
• A direct connection to the player promise rather than a generic claim about quality.
• At least one observable player effect for both alternatives.
• A clear distinction between evidence and assumption.
• One acknowledged cost of the accepted choice.
• A rejection reason that identifies a coherence, evidence, or cost problem.
• A condition that could cause the rejected option to be reconsidered.

A strong decision record can be read by another developer without a meeting and still communicate what direction to implement and what tempting direction to avoid.

2426. Key takeaways

• Creative direction is the disciplined selection and rejection of alternatives.
• Judge alternatives by their effect on the player promise, not by novelty or isolated polish.
• Separate observed evidence from inference and preference.
• Every accepted direction has a cost; state the cost instead of hiding it.
• A specific rejection makes the direction usable by the rest of the team and by AI collaborators.

2427. Next lesson

The next lesson is 5.16 — Capstone, where you will carry this creative direction into the capstone and demonstrate coherent project-level decisions.

2428. Knowledge check