Can 3D/4D Spatial Representation Beat Test-Time Compute? Zero-CoT Three.js Scene Graphs Outperform 345,000-Token Reasoning on 2D Grids and Eliminate Spatial Drift

In our companion study (Scene-Code-as-Topology), we showed under symmetric inference settings that compiling a 2D Gomoku grid into executable Three.js 3D/4D procedural scene graph code overcomes 1D token-stride and diagonal blindness (23W-3D-4L across 30 games). In this companion study (Paper II), we introduce an Asymmetric Compute-vs-Representation Handicap Benchmark (Group 4, 10 games, 205 plies on DeepSeek-V4.1-Flash-1017), pitting Agent A (2D ASCII Matrix with Full Thinking ENABLED, consuming 345,052 reasoning tokens at 28.08s/move) against Agent B (Three.js 3D/4D Scene Code with Thinking DISABLED, consuming 0 reasoning tokens at 5.52s/move).Despite a 0 vs. 345,052 reasoning-token handicap, Zero-CoT Agent B defeats Full-CoT Agent A by 6 Wins to 4 Losses (60.0% win rate, including a 6-game winning streak from Game 4 to Game 9 and 3/5 wins as second-mover White), while running 5.09x faster and reducing occupied-cell hallucinations by 94.1% (3 vs. 51 retries). Furthermore, we formalize Disconnected Spatial Drift (18 vs. 0 in Group 4; 31 vs. 1 across 20 DeepSeek games) and the Input-Length vs. Thinking-Time Inversion Law. Across all 40 games (924 moves in Groups 1-4), Scene-Code-as-Topology achieves 29W-3D-8L (78.4% decisive win rate, two-tailed exact binomial p = 0.00077 < 0.001).Open-Source Repository: https://github.com/pathetic777/scene-code-as-topology-gomoku Primary Archival DOI (CERN Zenodo): https://doi.org/10.5281/zenodo.23765598