1% Author: Simon-Pierre Boucher — simon-pierre.boucher@uqo.ca2% generated by scripts/20_evaluate.py — do not edit3\begin{tabular}{lrrrrrrrrr}4\toprule5Arm & {Mean $|e|$} & {MAE (CAD)} & {MdAPE} & {RMSE (log)} & {P90 $|e|$} & {P95 $|e|$} & {Bias $\bar e$} & {$|e|>\log 1.25$} & {N} \\6\midrule7A (hedonic) & \num{0.191} & \num{88962} & \num{0.126} & \num{0.286} & \num{0.412} & \num{0.641} & \num{0.014} & \num{0.270} & \num{96935} \\8A (hedonic spatial) & \num{0.208} & \num{76181} & \num{0.154} & \num{0.279} & \num{0.417} & \num{0.552} & \num{-0.078} & \num{0.359} & \num{8988} \\9B (hierarchical) & \num{0.182} & \num{84925} & \num{0.115} & \num{0.281} & \num{0.402} & \num{0.645} & \num{0.026} & \num{0.242} & \num{96935} \\10C (catboost) & \num{0.177} & \num{82153} & \num{0.111} & \num{0.278} & \num{0.375} & \num{0.640} & \num{0.021} & \num{0.226} & \num{96935} \\11C (gbm) & \num{0.179} & \num{83227} & \num{0.114} & \num{0.278} & \num{0.388} & \num{0.630} & \num{0.023} & \num{0.238} & \num{96935} \\12C (gbm spatial) & \num{0.176} & \num{65412} & \num{0.128} & \num{0.258} & \num{0.358} & \num{0.537} & \num{0.050} & \num{0.241} & \num{8988} \\13C (xgboost) & \num{0.180} & \num{85232} & \num{0.112} & \num{0.286} & \num{0.392} & \num{0.668} & \num{0.055} & \num{0.231} & \num{96935} \\14D (avm) & \num{0.171} & \num{80059} & \num{0.105} & \num{0.270} & \num{0.373} & \num{0.622} & \num{0.010} & \num{0.218} & \num{96935} \\15D (avm spatial) & \num{0.161} & \num{57681} & \num{0.105} & \num{0.248} & \num{0.345} & \num{0.527} & \num{0.045} & \num{0.204} & \num{8988} \\16\midrule17E (single llm) & \num{0.166} & \num{76060} & \num{0.107} & \num{0.250} & \num{0.354} & \num{0.508} & \num{0.013} & \num{0.237} & \num{80} \\18E (single llm small) & \num{0.182} & \num{90125} & \num{0.104} & \num{0.287} & \num{0.406} & \num{0.634} & \num{0.038} & \num{0.273} & \num{300} \\19F (independent) & \num{0.159} & \num{76764} & \num{0.091} & \num{0.240} & \num{0.314} & \num{0.499} & \num{0.003} & \num{0.212} & \num{80} \\20F (independent small) & \num{0.176} & \num{86596} & \num{0.100} & \num{0.281} & \num{0.417} & \num{0.572} & \num{0.035} & \num{0.257} & \num{300} \\21G (debate r2) & \num{0.158} & \num{73765} & \num{0.087} & \num{0.243} & \num{0.307} & \num{0.493} & \num{-0.004} & \num{0.225} & \num{80} \\22G (debate r2 small) & \num{0.180} & \num{88861} & \num{0.106} & \num{0.283} & \num{0.413} & \num{0.602} & \num{0.028} & \num{0.280} & \num{300} \\23H (debate skeptic) & \num{0.158} & \num{73923} & \num{0.092} & \num{0.242} & \num{0.322} & \num{0.493} & \num{-0.002} & \num{0.225} & \num{80} \\24I (placebo) & \num{0.157} & \num{75095} & \num{0.093} & \num{0.239} & \num{0.321} & \num{0.497} & \num{0.004} & \num{0.212} & \num{80} \\25J (cost matched) & \num{0.163} & \num{75760} & \num{0.118} & \num{0.249} & \num{0.347} & \num{0.497} & \num{0.009} & \num{0.250} & \num{80} \\26\bottomrule27\end{tabular}28