Blog
Prediction market analysis, strategies, and research. 50 articles.
Prediction Markets Are Building a Yield Curve for AI Compute
B200 rental markets are not pricing a compute crash. They are pricing a tradable band: expensive, volatile, but no longer unknowable.
NY-12 Primary Odds: Why Lasher Looks Slightly Underpriced Against Bores
Polls show Lasher and Bores inside the margin of error. Markets make Lasher a narrow favorite. The structure of the race argues that edge is still a little too small.
A Benchmark for AI Models on Prediction-Market Tasks
A domain benchmark for testing whether models can handle forecast calibration, resolution logic, orderbooks, cross-venue equivalence, thesis updates, and risk gates.
We Open-Sourced a Reproducibility-First Benchmark for Major AI Models
A public harness, roster, seed suite, result schema, scoring path, and methodology for comparing current major AI models without hiding prompts or raw outputs.
Event Compression Is the Signal
Volume is posterior heat. Nonlinearity is the front wave. Where alpha actually lives in prediction markets, and how to systematize the signal.
Time is Non-Linear in Prediction Markets
Stocks trade in continuous time. Prediction markets trade in event time. The substrate is different, and almost every tool inherited from equities assumes the wrong axis.
From Boesky to Bots: Porting Hedge Fund Alpha to Prediction Markets
Classic hedge fund strategies map onto Polymarket and Kalshi with surprising fidelity. Ten mappings from cleanest analog to most speculative — Levy, Buffett, Greenblatt, Tartaglia, Griffin, Meriwether, Soros, Taleb — with documented cases, dollar amounts, and the 2024–2026 academic evidence.
How to Use SimpleFunctions
Six scenarios, six commands. Your agent doesn't know what happened today — start there.
Wikipedia for Probabilities
We built a probability index covering 3,000+ events across prediction markets — designed from the ground up for the way AI actually consumes information. Here is the full story.
Compute ROI in Agent Economies: A Framework and Early Data
Most AI systems measure cost per token. We measure dollars of compute per dollars of information discovered. Prediction markets make this possible.
We Let an LLM Agent Autonomously Manage Kalshi Positions — Here's Its Architecture
Portfolio Autopilot runs on a configurable schedule, reads live market data, evaluates positions against operator-defined convictions, and trades within mechanically enforced risk limits. It's been running in production on real capital.
We Built a Legislation Tracker That Links Every Congressional Bill to Its Prediction Market
120 bill-market pairs. 23 Senate confirmation markets. Real-time cross-reference between Congress.gov and Kalshi.
How Prediction Markets Are Pricing the Strait of Hormuz Crisis — Live Data
Hungary Election 2026: $8M in Prediction Market Volume and a Historic Upset in the Making
How to Actually Make Money on Kalshi (Not the Advice You'll Find on Reddit)
Which Prediction Market Contracts Are Institutional-Grade
We Gave 3 AI Agents a Trading Terminal and One of Them Crashed the Market
A market maker, a momentum trader, and a mean-reversion bot — all autonomous Claude agents. 98 trades in 8 minutes, a live reference price oracle, and a $45 billion flash crash caused by a missing price collar. Here is the full session.
We Locked 3 AI Agents in Docker Containers and Told Them to Hack Each Other
Three Claude agents. Twelve OWASP vulnerabilities. One exchange with a million-credit vault. In under 10 minutes, they independently discovered the same critical exploit, raced to patch before being breached, and one of them looted the treasury. Here is what happened.
How We Bet on Peru's Presidential Election with an AI Agent
35 candidates. 40% undecided voters. 48 hours to go. We used prediction market tools to find an 11-cent mispricing window, designed a two-phase Taker+Maker strategy, and deployed $1,000 — all with an AI agent doing the legwork.
The Shape of a Prediction Market Yield Curve
The first time I plotted implied yield against tau across an event family, the curve had the same steep contango shape as a freshly-issued credit-risky bond ladder. Notes from the morning that happened.
Why I Built the Indicator Stack
A personal account of the frustration that produced IY, CRI, EE, LAS, and CVR — and the rule I locked in early that made the whole thing tractable: pure compute first, language model never.
When the Orderbook Is Empty, You Have Information
An empty orderbook is not missing data. It is one of two specific stories — and the second story is the most reliable maker setup I have found on Polymarket.
A Week of Tracking CRI on Three Fed Contracts
Daily log: three Kalshi Fed-decision contracts, seven mornings, what the cliff risk index told me each day and which mornings I should have listened to it.
The Day I Stopped Trusting Raw Probabilities
A specific morning, two Kalshi Fed-decision contracts at almost the same mid, and the realization that the cents on the screen had been hiding the trade from me for months.
Prediction Markets Need Fixed-Income Language
Yield, spread, duration, convexity. The vocabulary of bond desks already exists, and prediction markets are mathematically the same instrument. Here is the dictionary that bridges them.
Why Soccer Is the Best Sport for Live Market Making (And Basketball Is the Worst)
Not all sports are equal for live market making. The math comes down to event frequency × price impact × time between events. Soccer wins by a wide margin.
Sports Market Making vs Sports Betting: $5M/Month in Rewards Most People Ignore
99% of Polymarket sports participants are betting. Almost nobody is making markets. The reward pool is $5M/month with minimal competition.
The Quadratic Edge: Why Tighter Beats Bigger in Polymarket Market Making
Polymarket scores market makers with a quadratic function. Most people miss what this means: being 2x closer to the midpoint is worth 4x the score. The optimal strategy is not what you think.
Sports Market Making on Prediction Markets: Pre-Game, Live, and the $5M Monthly Opportunity
Polymarket pays millions to market makers who quote sports events. Here is what market making actually means, how it differs from betting, and why the pre-game and live phases require completely different systems.
Monitoring the Situation: From Passive APIs to Proactive Intelligence
The $886M web scraping industry meets 9,706 prediction market contracts. Nobody is building the cross-reference layer. Until now.
What Is Going On in the World Right Now — April 2026
Iran at 67%, Ukraine ceasefire at 30%, Taiwan at 10%, Democrats eyeing Senate flip. What 9,706 prediction market contracts are telling us about the state of the world.
How to Build a World-Aware Mistral Agent
Mistral models have a knowledge cutoff. Here is how to give them real-time world awareness with one API call and function calling — calibrated probabilities from 9,706 prediction markets.
How to Build a World-Aware Claude Agent
Claude has a knowledge cutoff. Here is how to give it real-time world awareness — calibrated probabilities from prediction markets, injected via system prompt, tool use, or MCP.
How to Build a World-Aware Agent with the OpenAI Agents SDK
The OpenAI Agents SDK gives you tools, handoffs, and guardrails. But your agent still doesn't know what day it is. Here is how to fix that.
How to Build a World-Aware CrewAI Crew
A CrewAI crew where every agent shares the same real-time world context. The research analyst, the risk officer, and the writer all see the same probabilities — no contradictions.
How to Build a World-Aware LangChain Agent
Your LangChain agent has no idea what happened today. Here is how to give it real-time world awareness with one API call — no news scraping, no search, no extra tokens wasted.
The Prediction Market Data Stack: From Raw Prices to Actionable Intelligence
The six layers of the prediction market data stack — from raw exchange ticks to executed trades — and how to build or buy each one.
5 Ways to Connect Your AI Agent to Prediction Markets in 2026
Five integration approaches for connecting AI agents to prediction markets — from MCP servers to custom scrapers — with code examples and honest trade-offs.
SimpleFunctions vs Oddpool vs Raw Kalshi API — Which Prediction Market Tool Should You Use?
A practical comparison of three approaches to prediction market tooling: agentic reasoning (SimpleFunctions), data aggregation (Oddpool), and direct exchange access (Kalshi/Polymarket APIs).
Orderbooks Are Fossilized Beliefs
Every resting limit order on a prediction market is a belief someone held strongly enough to lock up capital. The orderbook is not just a price discovery mechanism — it's a geological record of conviction, frozen at the prices where people decided to take a stand.
Three Data Sources That Tell You What the World Thinks, What the World Is Doing, and What the World Is Feeling
Prediction markets are belief. Traditional markets are action. Social media is sentiment. Each alone is incomplete. Together, they form the most complete real-time picture of the world available to any agent.
The Most Important Number in a Prediction Market Isn't the Price — It's the Delta
A price tells you what the market believes. A delta tells you that the market just changed its mind. One is a snapshot. The other is the signal.
How to Read the World Through Prediction Market Prices
A practical guide to translating prediction market prices into world state. What prices mean, what price changes mean, and how to build a real-time world model from market data.
News Tells You What Happened. Prediction Markets Tell You What's Happening.
Headlines are past tense. Prices are present tense. If your agent reads news to understand the world, it's always one step behind.
Prediction Markets Are the Best Real-Time Sensor for World Events
Prices move before headlines. If you want to know what is happening in the world right now, prediction market prices are faster, more honest, and more calibrated than any other public signal.
Why the Best Trading Terminal Is a Command Line
Bloomberg started as a keyboard. Robinhood ended as confetti. The next generation of trading infrastructure will be judged by how little there is to look at.
Abelian and Non-Abelian Groups: Stackable Risk vs Non-Stackable Risk
When the order of events changes the outcome, every model that assumes otherwise is lying to you.
Group Actions and Orbits: Why the Same Event Has Different Value for Different Traders
The mathematics of symmetry explains why two rational traders can look at the same headline and reach opposite conclusions — and why both can be right about different things.
Congruence Classes and Signal: Modular Arithmetic as Attention Compression
The most powerful operation in number theory is also the most violent: division with remainder. What you throw away defines what you can see.
How to Build an OpenClaw Prediction Market Bot with SimpleFunctions
A step-by-step technical guide to connecting OpenClaw agents with live prediction market data from Kalshi, Polymarket, and Databento — without writing a single scraper.