Code-agent ordering: TrackerAgent > SetRaceAgent > EconomyAgent

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Two heuristic code agents (TrackerAgent and SetRaceAgent) outperform most tested LLMs.claim0.799
  Calibration that conditional logic can beat cost-efficient LLMs in this setting.
• SetRaceAgentmethod0.787
  deterministic code agent that greedily pursues quartet completion, bidding aggressively on near-complete sets
• TrackerAgentmethod0.753
  deterministic code agent that maintains perfect information from observable events and makes greedy decisions conditioned on card counts and estimated wealth
• Code agents operate on structured data with exact arithmetic, while LLMs must parse natural-language observations and track state across turns; some failures may partly reflect numerical parsing or working-memory limitationsclaim0.738
  discussion of potential confounds
• TrackerAgent and SetRaceAgent have TC tightness τ ≈ 0.2–0.25, looser countersfinding0.735
  Code agents trade bargaining precision for acquisition pressure.
• Active inference agent with learnable preferences developed a strict preference for goals (score +) when the Frisbee location was encountered first, becoming a goal-seeking agent.finding0.721
  Figure 5.4 and text.
• Within-agent score standard deviation suggests deck order matters more than seat position.claim0.718
  Observation from variance analysis, though not tested as hypothesis.
• EconomyAgentmethod0.717
  deterministic code agent that models resource economy, tracking money flows and exploiting cash-poor opponents