split,arm,n,coverage_50_raw,coverage_90_raw,width_90_raw,scale_50,scale_90,coverage_50_conformal,coverage_90_conformal,width_90_conformal test (mid),E_single_llm,80,0.4625,0.8875,0.5620513641013618,1.2643313611844822,1.6294725791155371,0.525,0.925,0.9158472858576518 test (mid),F_independent,80,0.5125,0.8875,0.589534685295042,1.2611647423036356,1.584302968391311,0.6125,0.95,0.9340015518825725 test (mid),G_debate_r2,80,0.475,0.8875,0.6022991979979221,1.2644742287419792,1.5481787277678398,0.6375,0.9375,0.9324668060920134 test (mid),J_cost_matched,80,0.475,0.8625,0.5608868161548359,1.2432409161671234,1.5323539200936531,0.525,0.9375,0.859477111463711 holdout 2026,E_single_llm,60,0.4666666666666667,0.8,0.49688479586848483,1.2643313611844822,1.6294725791155371,0.6333333333333333,0.8833333333333333,0.8096601498471171 holdout 2026,F_independent,60,0.5166666666666667,0.8333333333333334,0.5077470175934654,1.2611647423036356,1.584302968391311,0.5833333333333334,0.9,0.8044251071651625 holdout 2026,G_debate_r2,60,0.5333333333333333,0.8333333333333334,0.5277165599452268,1.2644742287419792,1.5481787277678398,0.5833333333333334,0.8833333333333333,0.8169995523980222 prospective,E_single_llm,40,0.425,0.75,0.5579650700737118,1.2643313611844822,1.6294725791155371,0.6,0.9,0.9091887817893924 prospective,F_independent,40,0.5,0.8,0.5741003394033062,1.2611647423036356,1.584302968391311,0.575,0.925,0.909548871871117 prospective,G_debate_r2,40,0.45,0.825,0.5707408543553014,1.2644742287419792,1.5481787277678398,0.575,0.925,0.8836088497809206