> ## Documentation Index
> Fetch the complete documentation index at: https://fracturestudio.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# How Fracture Studio Scores and Evaluates Your Argument

> Understand the 0–100 audit score and its dimensional breakdown so you know exactly where your argument stands and what to fix first.

Every audit produces an overall score from 0 to 100, plus a breakdown across the parts of an argument that decide whether it survives. The overall score tells you at a glance how much work your draft needs; the dimensional breakdown tells you *where* to start.

## The Overall Score

| Range | Meaning |
| - | - |
| 90–100 | Resilient under close questioning |
| 75–89 | Strong with fixable pressure points |
| 60–74 | Usable but vulnerable |
| 40–59 | Major revision needed |
| 11–39 | The argument breaks under pressure |
| 0–10 | Not enough of an argument to evaluate |

## The Breakdown

The score separates distinct dimensions so you can see which area needs attention first, rather than relying on a single number to guide your revision. Depending on the mode you choose, these dimensions include argument strength, assumption safety, logic, and rhetoric.

> [!NOTE]
>
> Citation completeness is treated as a verification task, not as automatic proof that an argument is weak. Reasoning is scored first, so a draft with strong logic will not be dragged into the low ranges simply because sources still need checking.

## How It Is Calibrated

The scoring scale is designed to reflect the actual state of your reasoning — not a mechanical penalty system.

* A clear, coherent, well-warranted draft can score in the 80s or 90s even if its citations still need work.
* The engine does not push a draft into the 20s or 30s simply because sources still need verification.
* Very low scores (0–10) are reserved for drafts that are not meaningful arguments, are mostly unsupported assertion, or collapse in their central reasoning.

> [!TIP]
>
> After each revision cycle, compare the score breakdowns across audit runs side by side. Watching individual dimensions move — or stay flat — tells you precisely which changes strengthened your argument and which ones still need more work.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.