A dynamic asset selection system that adapted to market conditions, promoting instruments with momentum, demoting those without. It served its purpose. We moved on.
The Adaptive Watchlist was a system that automatically adjusted which trading instruments to monitor based on market momentum. Instruments showing consistent movement were promoted; stagnant ones were demoted. The goal was to focus attention and capital, where the market was actually doing something.
It worked. It taught us about market dynamics, data pipelines, and the limits of rule-based selection. But as ZETA evolved into a more capable LLM-driven agent, the watchlist's role was absorbed into the broader system. Rather than maintaining a separate component, we archived it.
Rule-based systems are brittle. Momentum thresholds that worked in one market regime failed in another. This insight pushed us toward LLM-driven decisions, the core of what became ZETA.
Self-hosted data pipelines are worth the effort. The watchlist's data flow (broker API → Python → Teable) became the template for ZETA's architecture.
Knowing when to archive is as important as knowing when to build. The watchlist wasn't a failure. It was a stepping stone. Learn. Adapt. Build.
The Adaptive Watchlist is no longer running. Its lessons live on in ZETA and in how we approach every new project.
Back to L.A.B.