Intelligence layer, not a trading bot

PredictorIQ

Evaluate signal quality in prediction markets

PredictorIQ is a decision intelligence layer built on prediction markets. We don't predict outcomes or automate trades. Instead, we evaluate whether price movements, capital flows, and structural changes are worth trusting in noisy, fast-moving environments.

Prediction markets compress time, information asymmetry, and economic feedback into a uniquely demanding environment, making them an ideal proving ground for signal evaluation and system robustness.

Request Access See How It Works

Current stage: working prototype with live market coverage (Polymarket, Kalshi, Myriad, Limitless Exchange).

What PredictorIQ Does

We evaluate signal quality through four core methods. These capabilities serve both individual investors (seeking mispricing insight) and institutional users (seeking early risk signals).

Option-Anchored Mispricing Analysis

Pricing anchor

For markets with price-based payoffs, we map them to equivalent option structures and use Black–Scholes–style probability estimates tied to real option market implied volatility.

This gives us an independent fair probability estimate to compare against the prediction market's implied probability, revealing potential mispricing without subjective forecasting.

Wallet Behavior & History Analysis

Capital quality

We continuously track wallet behavior on both Yes and No sides, including historical performance, consistency, position changes, and entry/exit timing.

The goal isn't to label "smart money." Instead, we determine whether one side is primarily driven by historically more informative participants.

Anomaly & Outlier Detection

Statistical signal

Statistical methods identify wallets that deviate significantly from their own historical baseline or overall market behavior, including early positioning, unusual concentration, or coordinated activity.

These signals indicate potential information strength, not claims about motivation or insider knowledge.

Short-Horizon Evaluation (15-min)

Rapid resolution

For very short-duration markets (e.g., Polymarket 15-minute contracts), we analyze late-window price dislocations, liquidity structure, and short-term mean reversion patterns.

Transaction costs are explicitly incorporated. Opportunities surface only when post-fee expected value remains positive, distinguishing emotional mispricing from genuine information updates.

A demo repository is available showing our approach and capabilities. View demo repository on GitHub .

Who It's For

PredictorIQ serves two distinct user groups with different objectives, but both rely on the same underlying signal evaluation methods.

Individual Investors

Core question: Is this price movement noise, or does it reflect genuine informational edge or mispricing?

• Evaluate whether observed price changes are driven by noise, emotion, or historically informative behavior
• Identify potential mispricing through option-anchored fair value estimates
• Understand capital positioning and wallet quality on each side
• Assess short-horizon opportunities with explicit cost analysis

Institutional Users

Core objective: Early risk signaling, not trade execution.

• Track probability structure changes across clustered, related markets
• Monitor capital concentration, reallocation patterns, and behavioral shifts
• Detect potential risk regime shifts when structural changes and capital behavior reinforce each other
• Receive explainable early-warning signals tied to specific markets and time windows

How It Works

PredictorIQ uses an agent-native architecture, not as a marketing feature, but because prediction markets demand speed, frequency, and continuity that exceed manual analysis capabilities.

Agent Architecture

System Agents

Continuously generate mispricing signals, behavioral analysis, anomaly detection, and risk indicators. These agents operate 24/7 across multiple markets and venues.

User Agents

Represent individual or institutional users. They consume signals, execute decisions, and provide feedback, enabling personalized signal filtering and action workflows.

This architecture is driven by market constraints (speed, continuity, scale), not marketing positioning. The core product remains signal evaluation and risk insight.

Step 1

Monitor

System agents continuously track markets, prices, wallet activity, and structural changes across venues.

Step 2

Evaluate

Apply option-anchored pricing, wallet analysis, anomaly detection, and cost-adjusted opportunity assessment.

Step 3

Signal

Surface high-quality signals with clear explanations, confidence levels, and supporting evidence.

Step 4

Learn

Track signal outcomes, refine models, and improve evaluation methods based on real market feedback.

Why Prediction Markets

Prediction markets are our proving ground, not our final form. They offer a uniquely compressed environment for testing signal quality and system robustness.

Time Compression

Markets resolve quickly (minutes to months), providing rapid feedback loops for signal validation and model refinement.

Information Asymmetry

Participants have varying information quality, making signal evaluation critical and measurable through market outcomes.

Economic Feedback

Real capital at stake creates genuine economic incentives, producing verifiable, backtestable data for system improvement.

Long-term vision: The methods we develop for prediction markets (evaluating signal quality under noise, tracking behavioral patterns, detecting structural anomalies) apply to broader decision intelligence problems. Prediction markets are where we prove the system works.