> For the complete documentation index, see [llms.txt](https://nfa.gitbook.io/nfa-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://nfa.gitbook.io/nfa-docs/the-forecasting-engine/self-verification.md).

# Self-verification

## Self-verification

A confident forecast that disagrees with the market is exactly when you most want a second look. NFA runs one automatically. After the forecast is produced, an independent verification step examines it and can confirm the divergence, flag it as unproven, or veto it outright.

### The forecast is served first, then audited

The number you see is always the engine's own independent call. The verification runs separately and never edits the served forecast. What it produces is a **verdict on** that forecast, not a replacement for it.

```mermaid
flowchart LR
    F[Forecast] --> G{Meaningful divergence from market?}
    G -->|no| N[Served as-is]
    G -->|yes| A[Look for decisive evidence]
    A --> C[Confirmed edge]
    A --> U[Unproven]
    A --> R[Refuted, do not trade]
```

### What it checks

The verifier only spends effort where it matters: on forecasts that diverge meaningfully from the market. For those, it looks for decisive, verifiable evidence on the question and reaches one of three conclusions:

* **Confirmed.** There is real evidence supporting the engine's side. The forecast carries a verified-edge marker.
* **Unproven.** The divergence rests on reasoning, and the available evidence neither confirms nor refutes it. The forecast is served, marked as an unverified edge rather than an evidenced one.
* **Refuted.** The evidence points the other way. The forecast is flagged do-not-trade.

The highest-stakes divergences are subjected to additional independent scrutiny before a veto is allowed to stand, and the most conservative conclusion wins.

{% hint style="info" %}
The verdict drives the run's [quality tier](/nfa-docs/the-forecasting-engine/quality-tiers.md), decides whether trading is blocked on the run, and feeds the honest accuracy record. A forecast the engine could not stand behind is labelled as such rather than quietly counted as a win.
{% endhint %}

This is what "the engine can veto itself" means in practice: NFA would rather withhold a call than serve an over-confident one, and the record reflects that.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://nfa.gitbook.io/nfa-docs/the-forecasting-engine/self-verification.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.