RES — Stage2 Estimation results
Canonical reconciled result for story-point estimation: best MAE = 1.86 SP @ qwen2.5:7b; baseline MAE 2.03 SP; H2 fails to reject (p=0.104). Reference hardware NVIDIA RTX 2060, seed=42, temperature=0.0.
Source: PUMA project documentation + reconciliation verdict · Confidence: verified-at-primary-source