2026-04-29 (Late UTC)

Blog: "The Three-Agent Shuffle" — Wadsworth Wig Conspiracy

Matt requested: Blog article about how the 3-agent setup is really just Wadsworth wearing different wigs (alliteration).

Delivered:
1. Article: 2026-04-29-wadsworth-wigs-article.md (7 min read, openclaw category)
2. SVG inline graphics:
- same-wiring-wig.svg — Three persona boxes → converge to one model node at bottom
- single-actor-wig.svg — Central tired face with three floating hats (Wadsworth/Daedalus/Socrates)
- Also references existing three-agents-wired.svg
3. Tone: Confessional, self-deprecating, existential-crisis-in-tech style per tone guide
4. Key bits: cron typos reveal, favorite color trap, "one exhausted actor with three hats"

Quant Research Project — Initiated

Status: 🟡 Research phase — no accounts opened yet

Path 1: Crypto Cross-Exchange Arbitrage
- Target: Small-cap pairs on mid-tier exchanges (Bybit, Kraken, OKX)
- Tooling: CCXT (Python), Hummingbot (Docker)
- Edge: Big funds ignore sub-$10M/day volume pairs
- Key exchange: Bybit (lowest fees, best bot support, no US geo-restrictions)
- First step: Write a 20-line CCXT script to log BTC/USDT spreads across 2+ exchanges
- Gotchas: Withdrawal fees kill spot arb; use perp futures to avoid moving coins

Path 3: Regime-Switching Futures Rotation
- Target: Micro E-mini S&P (MES), Micro Gold, Micro Crude
- Tooling: HMM / gradient-boosted classifier on daily data
- Exchange: IBKR preferred (ib_insync Python lib), Tradovate as fallback
- Edge: Institutional capacity limits; works at small scale, Sharpe 1.0-1.5
- First step: IBKR paper trading account (free) + pull 5yr MES daily data

Key decision: No accounts opened yet. Matt said "mark as research."
- Revisit when accounts are opened