> For the complete documentation index, see [llms.txt](https://midashand-1.gitbook.io/midaspredict/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://midashand-1.gitbook.io/midaspredict/the-basics/how-midaspredict-works.md).

# How MidasPredict Works

MidasPredict is a prediction market. You buy and sell shares whose value depends on the outcome of a future event. If you're right, your shares are worth $1. If you're wrong, they're worth $0.

If you've used a trading platform before, the basics will feel familiar. The difference is that you're not trading tokens — you're trading on *outcomes*.

***

## Shares and Prices

Every market on MidasPredict asks a question about the future. Take a simple example:

> **"Will Bitcoin exceed $150,000 by December 31, 2026?"**

This market has two tradable shares: **YES** and **NO**.

* If Bitcoin exceeds $150K by the expiry date, YES shares resolve to **$1.00** and NO shares resolve to **$0.00**.
* If it doesn't, NO shares resolve to **$1.00** and YES shares resolve to **$0.00**.

**Share prices reflect probability.** If YES is trading at $0.35, the market collectively estimates roughly a 35% chance the event happens. If you think that probability is too low, buying YES is a bet that you know something the market doesn't.

### The Math Is Simple

You buy 100 YES shares at $0.35 each. Your cost: **$35**.

* **If YES wins:** Your shares are worth 100 × $1.00 = **$100**. Profit: **$65**.
* **If NO wins:** Your shares are worth 100 × $0.00 = **$0**. Loss: **$35**.

That's it. Your maximum loss is always what you paid.

***

## Three Ways to Participate

### Daily Markets

The core engagement engine. Daily Markets are recurring, time-boxed prediction markets on observable variables — crypto prices, commodities, macro indicators — resolving on a fixed daily cadence.

**Examples:** BTC/USD closing price above $X at 16:00 UTC, gold spot price at London PM fix, Fed funds rate expectations.

Because each Daily Market instance shares identical structure with previous days, participants build familiarity and return with minimal friction. Liquidity concentrates in the current day's contract.

### Featured Markets

Higher-profile, event-driven markets around elections, championships, policy announcements, and other time-specific events with longer time horizons. These drive new-user acquisition through social sharing and media coverage.

**Examples:** U.S. election outcomes, FIFA World Cup match results, major central-bank rate decisions, crypto governance votes.

### Creator-Led Markets

Any user who has earned the [Midas Catalyst badge](/midaspredict/for-creators/midas-catalyst.md) can create a market on any well-defined, resolvable event. Creators earn 0–2% of traded volume. See [Creating a Market](/midaspredict/for-creators/creating-a-market.md).

***

## How Prices Move

When more people buy YES, the YES price goes up. When more people buy NO, the NO price goes up. Prices adjust automatically — you don't need to find a counterparty.

MidasPredict uses a pricing engine called **LS-LMSR** (Liquidity-Sensitive Logarithmic Market Scoring Rule) that manages liquidity behind the scenes. What this means in practice:

* **You can always trade.** Every outcome in every market is tradeable from the moment it launches — no waiting for a counterparty.
* **One shared pool per market.** All outcomes share a single pool of collateral. Buying YES doesn't just move YES — every other outcome's price adjusts coherently in the same trade.
* **Liquidity grows with activity.** The more a market is traded, the deeper the pool becomes and the tighter the spreads get. Small markets stay tradeable; big markets get deeper automatically.

> 💡 **Think of share prices as probabilities.** YES at $0.65 means roughly 65% chance. If you disagree, that's your edge.

<details>

<summary><strong>Advanced (Optional): LS-LMSR Mechanics</strong></summary>

MidasPredict uses the **Liquidity-Sensitive Logarithmic Market Scoring Rule (LS-LMSR)**. It's an adaptive bonding curve with several useful properties:

**Cost function:** `C(q) = b · ln(Σ e^(q_i/b))`, where `q` is the vector of outstanding shares per outcome and `b` is the liquidity parameter.

**Adaptive depth:** `b = α · Q`, where `Q` is the total depth of the market. Early on, `Q` is small and the curve is sensitive — small trades nudge prices noticeably. As volume builds, `Q` grows, `b` scales with it, and the pool deepens automatically. Spreads tighten and large trades glide through without blowing up odds.

**Unified pool:** All outcomes share a single collateral pool — there's no fragmentation between YES and NO (or between A/B/C/D in multi-outcome markets). Under the hood, LS-LMSR behaves like a softmax over outcomes, so the distribution stays mathematically consistent at all times.

**Bounded loss:** The creator's seed liquidity caps the maximum loss the AMM can take. Traders can't drain more value than was provided.

**Why not a constant-product AMM?** A x·y=k AMM (like Uniswap v2) spreads capital evenly across price ranges and was designed for token swaps where both sides are continuously valued. LS-LMSR is purpose-built for outcome markets: collateral only needs to fund the winning side at resolution, not both sides simultaneously, so the math is more capital-efficient for fixed-outcome bets.

**Payouts can deviate slightly from $1.00 per share.** Because LS-LMSR keeps trade slippage inside the pool (rather than extracting it as a fee), deep and balanced markets can settle slightly *above* $1.00 per winning share at resolution. Conversely, thin or heavily lopsided markets can settle slightly *below* $1.00. Most middle-of-the-road markets land near $1.00. 100% of the pool's collateral is redistributed to winning shares at resolution — none of it is extracted by the protocol.

</details>

***

## How Markets Resolve

Every market moves through the same stages:

**Created → Pending → Active → Resolved → Dispute Period → Closed**

1. **Resolved:** Once the observation window ends, the oracle posts an outcome. Most markets (\~95%) resolve automatically via aggregated data feeds.
2. **Dispute Period:** A 24–48 hour challenge window opens. Any user can dispute the outcome by posting a bond. The Judge Team reviews and either approves the change or rejects the dispute.
3. **Closed:** If no dispute succeeds, the outcome is final. Winners redeem $1.00/share, losers get $0.00. Unclaimed winnings remain available for 90 days.

> ⚠️ **The dispute period is the one chance to correct a bad resolution.** Once a market is Closed, the outcome can't be changed.

For the full lifecycle, dispute mechanics, and void resolution, see [Market Lifecycle & Disputes](/midaspredict/knowledge-base/market-lifecycle-and-disputes.md).

***

## Multi-Outcome Markets

Not every question is YES/NO. MidasPredict also supports markets with multiple options.

**Example:** *"Which AI company will have the highest market cap by end of 2026?"*

* Option A: Nvidia
* Option B: Microsoft
* Option C: Google
* Option D: Other

All options share a single LS-LMSR pool, so every trade updates the whole distribution coherently. Buy Option A, and A ticks up while B, C, and D all tick down in the same motion. Share prices across all options sum to approximately $1.00, and only the winning option pays out at resolution.

***

## You Don't Have to Wait

You can **sell your shares at any time** before the market closes, at the current market price. This lets you:

* Take profit early if prices have moved your way
* Cut losses if your view changes
* React to breaking news or new information

You're never locked in.

***

## FAQ

<details>

<summary>Can I lose more than I put in?</summary>

No. Your maximum loss is the amount you paid for your shares. There are no margin calls or liquidations.

</details>

<details>

<summary>How fast do trades confirm?</summary>

Sub-100ms pre-confirmations via the Espresso sequencer on LitVM, with sub-cent gas fees.

</details>

<details>

<summary>What tokens can I use to trade?</summary>

zkLTC (Litecoin on LitVM), stablecoin (USDC), and approved partner tokens. See [Supported Tokens & Collateral](/midaspredict/knowledge-base/supported-tokens-and-collateral.md).

</details>

<details>

<summary>Who creates the markets?</summary>

Daily Markets and Featured Markets are curated by the MidasPredict team. Creator-Led Markets allow anyone to create and monetize custom markets. See [Creating a Market](/midaspredict/for-creators/creating-a-market.md).

</details>

<details>

<summary>What if a market's resolution is disputed?</summary>

Every resolved market enters a 24–48 hour dispute period. Any user can challenge the outcome by posting a dispute bond — the Judge Team reviews and may update the outcome. See [Market Lifecycle & Disputes](/midaspredict/knowledge-base/market-lifecycle-and-disputes.md) for full detail.

</details>


---

# 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://midashand-1.gitbook.io/midaspredict/the-basics/how-midaspredict-works.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.
