# Wisdoms From Underdog

This page captures the reusable lessons from the Underdog racing exercise.

## 1. Big Ideas Need Small Doors

An ambitious idea is not a problem. The problem is trying to enter it through
the biggest possible door.

The small door for Underdog was a time-trial physics lab.

## 2. Do Not Confuse More Features With More Game

More menus, currencies, unlocks, and rivals do not automatically create more
fun.

The question is:

```text
What does the player feel every five seconds?
```

For a racing game, that means steering, braking, speed, mistakes, recovery, and
track readability.

## 3. Make Invisible Systems Visible

The first car moved, but the player could not see the steering intention.

Visible front tires changed that.

The same applies to AI work:

```text
If the system matters, surface it.
```

Checkpoints, tests, screenshots, handoff notes, and HUD values are all ways to
make hidden work visible.

## 4. A Boring Test Map Can Teach You. A Boring Game Map Will Lose You.

The oval was useful when physics was fragile. Once the game needed racing vibe,
the map had to carry more meaning.

Good maps create decisions:

- brake now or risk the curb
- cut the corner or protect traction
- prioritize exit speed or entry speed
- recover after grass without rage quitting

## 5. Difficulty Is a Conversation

Difficulty should not be a secret punishment.

It should tell the player:

- what pressure they chose
- why the rival feels faster
- how to improve
- when to lower or raise pressure

## 6. Tests Are Not Only for Coders

The best non-technical test is still powerful:

```text
Can I open it, use it, see it, and explain what happened?
```

That question caught the rendering crash.

## 7. The Handoff Is Part of the Product

`README.md`, `GEMINI.md`, smoke notes, and screenshots are not leftovers. They
are how the next session starts without rediscovering everything.

## 8. Suffering Is Only Valuable When It Produces Signal

Hard games can be satisfying. Hard workflows can be productive.

But pain without feedback is just waste.

The harness converts suffering into signal:

```text
What failed?
Where?
How do we know?
What is the next smallest fix?
```

## Final Wisdom

Build the smallest thing that can teach you the next truth.

Then listen to it.
