Results for “statistical-arbitrage”

Orchestrate token-heavy research, coding, and testing by delegating bounded tasks to cheaper subagents while reserving Claude Fable for architecture, synthesis, and final review.

Benchmarks language model agents on the C2C multi-agent negotiation task, reporting win rate across starting positions.