AI Agents Are Quietly Rewriting Prediction Market TradingAI Agent 正在悄悄改写预测市场交易格局
Autonomous AI agents running 24/7 on Olas protocol are giving retail traders a disciplined, strategy-driven edge on Polymarket — early traction shows 4,200+ trades in one month, with individual returns up to 376%.在 Olas 协议上运行的自主 AI Agent 正让散户在 Polymarket 上拥有全天候、策略驱动的优势——早期数据显示单月执行 4200+ 笔交易，单笔最高收益达 376%。
The Setup
Prediction markets have exploded from niche forecasting tools into a $44B+ annual trading arena. But there’s a dirty secret: retail traders are increasingly losing to automated players who never sleep. Enter Olas and its flagship product Polystrat — an autonomous AI agent that trades Polymarket 24/7 on behalf of human users.
Valory AG co-founder David Minarsch’s core thesis: the intelligence locked inside modern AI models hasn’t yet been fully translated into financial markets. Prediction markets, built on probabilistic forecasting, are the perfect testing ground.
Key Takeaways
- Polystrat launched February 2026: executed 4,200+ trades on Polymarket in its first month, with returns as high as 376% on individual positions.
- The edge is behavioral, not just analytical: humans lose focus, sleep, and panic. Agents don’t. Disciplined execution of a strategy is often more valuable than the strategy itself.
- Olas is building an “agent economy”: a decentralized ecosystem where user-owned AI agents perform useful tasks and generate crypto returns — users self-custody and own the agent, not the protocol.
- Retail vs. automation: prediction market volume is ~$44B annually, concentrated in Kalshi (CFTC-regulated) and Polymarket (crypto-native). Both increasingly dominated by automated players.
Why It Matters
Rex has been watching Polymarket for a while. The emergence of user-owned AI agents as trading tools is the exact intersection of two of his core themes: AI Agent infrastructure and prediction markets as a financial primitive.
The key insight from Minarsch: this isn’t just about performance — it’s about access. Retail traders can now delegate execution to a disciplined agent, removing the emotional and timing disadvantages that kill most manual prediction market strategies.
This is also a proof-of-concept for the broader “agent economy” thesis: AI agents that generate economic value for their owners on-chain, fully autonomous, 24/7. If this works in prediction markets, the template applies everywhere.
What to watch:
- Polystrat’s sustained performance over 3-6 months (early 376% returns are exciting but need durability)
- Whether Olas/Valory can expand beyond Polymarket to Kalshi and other venues
- Regulatory response: CFTC already watches Kalshi closely — automated trading agents are a new variable
- Competing agent protocols entering prediction market space
背景
预测市场已经从小众预测工具爆发成年交易额超过 440 亿美元的金融赛道。但有个不太好说的现实：散户越来越跑不过不眠不休的自动化玩家。Olas 协议和它的旗舰产品 Polystrat 来了——一个在 Polymarket 上 24/7 代替用户交易的自主 AI Agent。
Valory AG 联创 David Minarsch 的核心判断：现代 AI 模型里锁着的智识，还没有充分转化到金融市场里。而预测市场天然就是概率预测的战场，是最好的试验场。
关键要点
- Polystrat 于 2026 年 2 月上线：首月在 Polymarket 执行超 4200 笔交易，单笔最高盈利 376%。
- 优势在于行为纪律，不只是分析能力：人会分心、睡觉、恐慌，Agent 不会。严格执行策略，往往比策略本身更值钱。
- Olas 在构建”Agent 经济”：用户自托管并拥有 Agent，Agent 在链上自主执行任务并产生收益，完全去中心化。
- 散户 vs. 自动化：预测市场年交易额约 440 亿美元，主要集中在 Kalshi（受 CFTC 监管）和 Polymarket（加密原生）。两个平台上自动化玩家占比越来越高。
为什么重要
Rex 一直在关注 Polymarket。用户自有 AI Agent 作为交易工具这件事，正好打在他两个核心主题的交叉点：AI Agent 基础设施 + 预测市场作为金融原语。
Minarsch 点出的关键洞察：不只是性能提升，而是准入平等。散户现在可以把执行交给一个有纪律的 Agent，消除了大多数手工预测市场策略失败的根源——情绪波动和时机判断。
这也是更宏大的”Agent 经济”命题的概念验证：AI Agent 在链上为所有者产生经济价值，全自主、24/7。如果这在预测市场成立，这套模板可以复制到任何地方。
值得关注：
- Polystrat 在 3-6 个月内的持续表现（早期 376% 的回报令人兴奋，但需要时间验证）
- Olas/Valory 能否从 Polymarket 扩展到 Kalshi 和其他平台
- 监管动向：CFTC 本来就盯着 Kalshi，自动化交易 Agent 是新变量
- 其他 Agent 协议入场预测市场赛道
Related Essays相关文章
Join Rex's Lab on Telegram 加入 Rex's Lab
Crypto · AI · Investing — raw thinking, before it becomes a tweet. 加密 · AI · 投资 — 推文之前，更原始的思考。