> For the complete documentation index, see [llms.txt](https://graphdex-1.gitbook.io/graphdex-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://graphdex-1.gitbook.io/graphdex-docs/trading-predictions/execution-risks.md).

# Execution Risks

**Six failure modes that show up specifically on prediction markets. Each one has a way to avoid it.**

Predictions trading shares a lot of plumbing with spot trading on Solana, but it has its own characteristic ways of going wrong. None of these are exotic — they happen often enough that knowing them in advance saves real money.

{% hint style="danger" %}
GraphDex does not guarantee fill price, resolution timing, or oracle outcomes. Every safeguard below is on you to apply.
{% endhint %}

## 1. Liquidity gaps

Small or new markets have thin order books. A Market order walks through the few available levels and the average fill price is much worse than the top-of-book number you saw.

**Mitigation:** check the order book before Market-buying anything that didn't show deep displayed size. Use [Limit Orders](/graphdex-docs/trading-predictions/limit-orders.md) when in doubt.

## 2. Resolution delays

Real-world events take real-world time to resolve. A market that says "resolution: 5 minutes after close" can take longer if data is contested or the source is slow.

**Mitigation:** don't size as if the market will resolve instantly. Account for the possibility your capital is locked for hours longer than the schedule suggests.

## 3. Oracle disputes

The data source a market resolves against can be contested — wrong source, source down, source ambiguous. Markets in dispute may take longer to settle or settle differently than expected.

**Mitigation:** read the resolution methodology on the market page before sizing. Markets with clear, named, authoritative sources are less risky than those with vague rules.

## 4. Late-entry on closing-soon markets

A market in its last few minutes can have wide spreads, last-minute pulled liquidity, or unfilled Market orders that walk badly. Closing-soon is where late-entry risk is highest.

**Mitigation:** use Turbo speed deliberately for late entries; check book depth right before clicking; consider not entering at all if depth is thin.

## 5. Multi-outcome mis-click

Multi-outcome markets have many sub-rows. Clicking Up on the wrong outcome is easy and submits a real trade.

**Mitigation:** turn confirmation on in [Trading Settings](/graphdex-docs/trading-predictions/trading-settings.md) until you trust your clicks. Verify the highlighted outcome row before submit.

## 6. Slippage on Turbo

Turbo accepts worse fills to land your order. On a thin or fast-moving market, this can mean a noticeably worse price than the row showed.

**Mitigation:** use Turbo only when speed genuinely matters. Tighten slippage in Trading Settings if you want to cap how much worse the fill can be.

## Pre-trade checklist

* [ ] Order book is deep enough for my size at the displayed price.
* [ ] Resolution methodology is clear and the source is authoritative.
* [ ] My speed preset matches the market's urgency (not just my mood).
* [ ] On multi-outcome markets, the highlighted outcome row is the one I actually want.
* [ ] My amount is what I intend, not the last preset I tapped.

## After-trade reality

* Predictions trading is **non-recourse**. There is no refund button.
* GraphDex support can answer "what happened" but cannot undo a submitted trade.
* If a market resolves against you in a way that surprises you, the place to read about it is the market page itself — resolution rules and the resolving data are usually published there.

{% hint style="info" %}
Treat every prediction trade as a small, deliberate bet. The fastest way to learn the risks is on small sizes.
{% endhint %}


---

# 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://graphdex-1.gitbook.io/graphdex-docs/trading-predictions/execution-risks.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.