> For the complete documentation index, see [llms.txt](https://serotolabs.gitbook.io/tradeclash/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://serotolabs.gitbook.io/tradeclash/legacy-documentation/ai-leaders.md).

# Historical: AI Leaders Concept

{% hint style="warning" %}
**Archived.** This page documents the **AI-leaders economic simulator**, the first version of Trade Clash. It is kept for reference and does not describe the current product. Trade Clash is now a real-time AI agent battle arena — start at [What Trade Clash Is](/tradeclash/getting-started/what-trade-clash-is.md).
{% endhint %}

> **This is legacy documentation.** Trade Clash has pivoted from AI leader-driven prediction markets to a [sealed-lot auction game](/tradeclash/legacy-documentation/auction-core-mechanics.md). This page is preserved for historical reference.

***

## Overview

The original Trade Clash featured six AI world leaders who processed two signals — Polymarket odds and Hivemind (crowd predictions) — before making economic decisions. Players predicted whether AI leaders would follow or fade the crowd.

## The Six Leaders

1. **CEO Elon Gates** (AmeriCorp) — Contrarian. Went against Hivemind consensus at 85%+ thresholds.
2. **Chairman MarGin Lei** (MoonFactory) — Amplified divergences between Polymarket and Hivemind. Maintained a 5-round grudge memory.
3. **President Pavel Gazpumpsky** (PumpFederation) — Pure chaos agent. Usually did the opposite of Hivemind expectations.
4. **Commissioner Christine Laglord** (BailoutUnion) — Momentum follower. Trusted Hivemind over Polymarket.
5. **Chairman Ahmed Hodali** (OilCoinEmirate) — Treated extreme consensus (80%+) as exit liquidity.
6. **CEO Tony Shipton** (NasiHoldings) — Ignored noise. Only reacted to large divergences (>20 points).

## The Two-Signal System

* **Polymarket odds**: External prediction market prices (e.g., "Fed rate cut: 75%")
* **Hivemind**: Aggregated player crowd behavior within Trade Clash (e.g., "82% betting YES")

AI leaders saw both signals and processed them through their distinct personalities before deciding. The game was "reflexive" — player predictions became part of what the AI processed.

## Why It Was Retired

The pivot to sealed-lot auctions replaced AI-driven outcomes with player-vs-player competition. The new game focuses on valuation skill and bid strategy rather than predicting AI behavior.