== Decision rule { "n_seeds": 5, "pass_counts": { "W1_shared_curve": 5, "W2_close_to_fair_oracle": 4, "W2b_beats_both_controls": 5, "W3_offset_sign": 2 }, "verdict": "partly supported", "per_seed": [ { "seed": 20261101, "errors": { "pooled": 2.8448754574456245, "no_pooling": 8.342624068989247, "complete_pooling": 3.7176530525388816, "oracle": 3.3727911532973667 }, "fleet_curve_rmse": 0.5207373682883977, "per_site_mean_curve_rmse": 0.9882655332689357, "sites_with_right_offset_sign": 4, "offset_slope": 0.4971675745030259, "W1_shared_curve": true, "W2_close_to_fair_oracle": true, "W2b_beats_both_controls": true, "W3_offset_sign": false, "control_beats_pooled": [] }, { "seed": 20261102, "errors": { "pooled": 3.6050158120195577, "no_pooling": 5.891215262030696, "complete_pooling": 4.629821858801246, "oracle": 2.802581549391389 }, "fleet_curve_rmse": 0.5410980886685299, "per_site_mean_curve_rmse": 0.8122598673301972, "sites_with_right_offset_sign": 6, "offset_slope": 0.5299135756171323, "W1_shared_curve": true, "W2_close_to_fair_oracle": false, "W2b_beats_both_controls": true, "W3_offset_sign": true, "control_beats_pooled": [] }, { "seed": 20261103, "errors": { "pooled": 3.275526300028456, "no_pooling": 8.35834485812222, "complete_pooling": 3.379542404987649, "oracle": 3.914105843206484 }, "fleet_curve_rmse": 0.5929028543844737, "per_site_mean_curve_rmse": 0.796296990729369, "sites_with_right_offset_sign": 3, "offset_slope": -0.0616482785646142, "W1_shared_curve": true, "W2_close_to_fair_oracle": true, "W2b_beats_both_controls": true, "W3_offset_sign": false, "control_beats_pooled": [] }, { "seed": 20261104, "errors": { "pooled": 3.0813928595561695, "no_pooling": 11.522233911636112, "complete_pooling": 3.565011715085662, "oracle": 4.0733155009694455 }, "fleet_curve_rmse": 1.3149433399188393, "per_site_mean_curve_rmse": 1.2812692783687702, "sites_with_right_offset_sign": 6, "offset_slope": 0.6716367082077541, "W1_shared_curve": true, "W2_close_to_fair_oracle": true, "W2b_beats_both_controls": true, "W3_offset_sign": true, "control_beats_pooled": [] }, { "seed": 20261105, "errors": { "pooled": 3.385389513511528, "no_pooling": 4.132149484042927, "complete_pooling": 3.9285850245907485, "oracle": 3.9974678928073843 }, "fleet_curve_rmse": 0.6188661145087989, "per_site_mean_curve_rmse": 0.6235776770915895, "sites_with_right_offset_sign": 4, "offset_slope": 0.271414970331408, "W1_shared_curve": true, "W2_close_to_fair_oracle": true, "W2b_beats_both_controls": true, "W3_offset_sign": false, "control_beats_pooled": [] } ] } == Heating error per seed (points of effective capacity) seed 20261101: penalty 1000; {'pooled': 2.84, 'no_pooling': 8.34, 'complete_pooling': 3.72, 'oracle': 3.37} seed 20261102: penalty 1000; {'pooled': 3.61, 'no_pooling': 5.89, 'complete_pooling': 4.63, 'oracle': 2.8} seed 20261103: penalty 1000; {'pooled': 3.28, 'no_pooling': 8.36, 'complete_pooling': 3.38, 'oracle': 3.91} seed 20261104: penalty 300; {'pooled': 3.08, 'no_pooling': 11.52, 'complete_pooling': 3.57, 'oracle': 4.07} seed 20261105: penalty 1000; {'pooled': 3.39, 'no_pooling': 4.13, 'complete_pooling': 3.93, 'oracle': 4.0} == Coefficient-space error (root mean square over the 8 turns) seed 20261101: {'pooled': 1.87, 'oracle': 2.37, 'complete_pooling': 2.1, 'no_pooling': 5.21} seed 20261102: {'pooled': 1.99, 'oracle': 2.3, 'complete_pooling': 2.29, 'no_pooling': 4.33} seed 20261103: {'pooled': 1.99, 'oracle': 2.58, 'complete_pooling': 2.17, 'no_pooling': 5.51} seed 20261104: {'pooled': 2.26, 'oracle': 4.06, 'complete_pooling': 2.2, 'no_pooling': 8.41} seed 20261105: {'pooled': 2.46, 'oracle': 3.22, 'complete_pooling': 2.63, 'no_pooling': 4.3} == Offset sign: substations of 8 with the injected sign, per estimator (post hoc: all) seed 20261101: sign {'pooled': 4, 'oracle': 3, 'no_pooling': 4}; slope {'pooled': 0.5, 'oracle': 0.35, 'no_pooling': -0.17}; |injected offsets| [0.9, 0.02, 0.25, 1.58, 3.63, 0.79, 1.11, 1.71] seed 20261102: sign {'pooled': 6, 'oracle': 8, 'no_pooling': 3}; slope {'pooled': 0.53, 'oracle': 1.0, 'no_pooling': 0.42}; |injected offsets| [0.62, 0.61, 3.85, 1.83, 1.45, 2.29, 2.33, 0.71] seed 20261103: sign {'pooled': 3, 'oracle': 5, 'no_pooling': 4}; slope {'pooled': -0.06, 'oracle': 0.26, 'no_pooling': 1.26}; |injected offsets| [1.8, 1.04, 0.73, 0.98, 0.95, 0.69, 2.68, 1.2] seed 20261104: sign {'pooled': 6, 'oracle': 7, 'no_pooling': 3}; slope {'pooled': 0.67, 'oracle': 1.47, 'no_pooling': 1.57}; |injected offsets| [0.97, 0.74, 2.91, 0.41, 1.04, 0.69, 1.07, 1.35] seed 20261105: sign {'pooled': 4, 'oracle': 5, 'no_pooling': 3}; slope {'pooled': 0.27, 'oracle': 0.66, 'no_pooling': 0.94}; |injected offsets| [0.33, 1.6, 2.56, 3.72, 1.05, 2.47, 0.85, 2.12] == Fair oracle capacity ratio (fitted over true) per turn seed 20261101: 1.03 0.98 1.05 0.99 0.84 1.03 1.00 1.38 seed 20261102: 1.03 1.02 1.09 0.99 0.98 1.08 0.87 0.79 seed 20261103: 1.13 0.81 1.23 0.99 0.82 1.01 1.05 0.94 seed 20261104: 1.16 1.06 1.05 1.50 1.23 1.21 1.05 1.18 seed 20261105: 1.02 1.08 0.85 0.92 0.83 1.08 1.10 1.05