// Track

Prediction Market Mechanics

CLOB integration from first principles

The Polymarket CLOB integration layer demystified — FOK vs GTC, USDC.e collateral, EOA signing vs proxy wallets, balance guards, and the semantic matching patterns for consensus-based calibration. Everything you need to build a prediction market bot that doesn't silently fail.

6 lessons~47 min total

Lessons are shown in recommended order. Complete them in sequence for the best experience — or jump to any lesson.

Lesson 263·9 min read

CLOB Architecture and Order Flow

How Polymarket's Central Limit Order Book actually works. Maker, taker, orderbook depth, the balance-allowance endpoint, and why thin books change how execution logic has to be written.

Lesson 264·8 min read

FOK vs GTC Execution Semantics

Fill-or-Kill attempts first for maker fee savings, Good-Til-Canceled is the fallback. The HE-06 bug: status=live means pending on the book, not filled. How to distinguish resting orders from actual fills.

Lesson 265·7 min read

USDC.e vs Native USDC on Polygon

Polymarket collateral is bridged USDC.e at 0x2791Bca1f2de4661ED88A30C99A7a9449Aa84174, NOT native USDC at 0x3c499c542cEF5E3811e1192ce70d8cC03d5c3359. Sending the wrong one equals $0 balance forever. How to verify before funding.

Lesson 266·8 min read

Signature Types and Proxy Wallets

signature_type=0 means EOA is the funder — simple. signature_type=1 means proxy wallet for magic/email logins. signature_type=2 is a different proxy flavor. When to use each, and the funder address trap that bites every new Polymarket integration.

Lesson 267·7 min read

Balance Guards and Derived-Address Validation

Every EVM bot must derive its EOA from the configured private key, log it at startup, and compare it to an expected-address env var. The Hermes wallet mismatch cost an hour of debugging — a five-line startup check would have caught it in one.

Lesson 268·8 min read

Semantic Market Matching for Calibration

The Metaculus/Manifold calibrator pattern. Embedding-based matching with all-MiniLM-L6-v2, a 0.60 cosine threshold, semantic reranking of top candidates. Why it works 3.8% of the time for political markets and what that means for calibration strategy.