# ChatGPT Prompt - Anchor Test (Steelman)

## Intent

Produce a doctrine-grade steelman poster that formalizes the anchor test criteria while preserving visual clarity.

## Primary Prompt

Use this prompt in ChatGPT web image generation:

```
Create a doctrine-style version of the anchor-test bridge metaphor.
Scene: bridge expansion into fog with a clearly grounded section and an active frontier section.
Add concise checklist-like criteria on signs or labels: observable, testable, predictive, falsifiable, decision-relevant.
Tone: rigorous and fair, not hostile, not anti-science.
Visual style: high-contrast infographic-meets-editorial poster.
Priority: readability first, symbolism second.
Avoid: tiny unreadable text, visual clutter, logos, copyrighted characters.
Output: one high-resolution 4:5 image suitable as a concluding doctrine post.
```

## Negative Constraints

- No accidental logos or copyrighted characters.
- No unreadable tiny text.
- No style drift from concept intent.

## Variant Prompts

### Variant A: Cleaner

```
Limit criteria labels to three maximum.
Increase whitespace and simplify bridge details.
Keep one dominant question callout.
```

### Variant B: More aggressive

```
Keep all criteria labels visible.
Add stronger visual contrast between anchored and unanchored sections.
Preserve non-hostile academic tone.
```

### Variant C: Minimal

```
Bridge into fog with one compact criteria stamp:
"observable, testable, predictive".
Single question callout near frontier.
```

## Run Log

- Date: 2026-06-25
- Model: ChatGPT image generation
- Best result: b93dc108-bfee-4ec0-ad8f-fbd1fd163b25.png
- Why chosen: Strongest steelman of the anchor frame and best doctrine-close image.