CVaR-tail active residual correction / results.json
Mechanism confirmed, baseline not beaten
1{
2 "seed": 2816,
3 "beta": 0.99,
4 "mechanism_checks": {
5 "cvar_quantile_check": {
6 "predicted_eta_quantile": 2.244101818870541,
7 "observed_eta": 2.2443537786481667,
8 "abs_error": 0.0002519597776258209,
9 "direct_cvar": 2.569699060237259,
10 "rockafellar_at_eta": 2.569666528962227
11 },
12 "ridge_shrinkage_check": {
13 "lambda": [
14 0.0,
15 0.01,
16 0.1,
17 1.0,
18 10.0
19 ],
20 "predicted_ratio": [
21 1.0,
22 0.9900990099009901,
23 0.9090909090909091,
24 0.5,
25 0.09090909090909091
26 ],
27 "observed_ratio": [
28 1.0,
29 0.9900990099009901,
30 0.9090909090909091,
31 0.5,
32 0.09090909090909091
33 ],
34 "max_abs_error": 0.0
35 },
36 "tail_region_check": {
37 "epsilon": [
38 0.5,
39 1.0,
40 2.0
41 ],
42 "predicted_fraction": [
43 0.38292492254802624,
44 0.6826894921370859,
45 0.9544997361036416
46 ],
47 "observed_fraction": [
48 0.38288,
49 0.681245,
50 0.954455
51 ],
52 "max_abs_error": 0.001444492137085862
53 }
54 },
55 "mini_experiment": {
56 "true_cvar": 1.655386962453665,
57 "runs": {
58 "random": [
59 {
60 "calls": 24,
61 "cvar_abs_error": 0.005825732289025609,
62 "global_mse": 0.0009243312249660243,
63 "tail_mean_error": 0.005825732289025609
64 },
65 {
66 "calls": 40,
67 "cvar_abs_error": 0.000731938163846424,
68 "global_mse": 0.0009532694097385896,
69 "tail_mean_error": 0.000731938163846424
70 },
71 {
72 "calls": 56,
73 "cvar_abs_error": 6.259185885060248e-05,
74 "global_mse": 0.00098975709215039,
75 "tail_mean_error": 6.259185885038043e-05
76 },
77 {
78 "calls": 72,
79 "cvar_abs_error": 0.0003108659759230026,
80 "global_mse": 0.0010007334963034686,
81 "tail_mean_error": 0.00031086597592278054
82 },
83 {
84 "calls": 88,
85 "cvar_abs_error": 0.0016970746152977778,
86 "global_mse": 0.0009365463697149309,
87 "tail_mean_error": 0.0016970746152979999
88 }
89 ],
90 "global_uncertainty": [
91 {
92 "calls": 24,
93 "cvar_abs_error": 0.028632310517711534,
94 "global_mse": 0.0008670792450797092,
95 "tail_mean_error": 0.028632310517711757
96 },
97 {
98 "calls": 40,
99 "cvar_abs_error": 0.00491547321271768,
100 "global_mse": 0.0008398343501442326,
101 "tail_mean_error": 0.00491547321271768
102 },
103 {
104 "calls": 56,
105 "cvar_abs_error": 0.0039830504544564604,
106 "global_mse": 0.0008402996863935622,
107 "tail_mean_error": 0.0039830504544564604
108 },
109 {
110 "calls": 72,
111 "cvar_abs_error": 0.003349616857913462,
112 "global_mse": 0.0008416172592643351,
113 "tail_mean_error": 0.00334961685791324
114 },
115 {
116 "calls": 88,
117 "cvar_abs_error": 0.0029859180973683674,
118 "global_mse": 0.0008425347666401883,
119 "tail_mean_error": 0.0029859180973683674
120 }
121 ],
122 "tail_residual": [
123 {
124 "calls": 24,
125 "cvar_abs_error": 0.008794593977745402,
126 "global_mse": 0.000835577839607,
127 "tail_mean_error": 0.008794593977745402
128 },
129 {
130 "calls": 40,
131 "cvar_abs_error": 0.0019135625095341169,
132 "global_mse": 0.0008423661401131514,
133 "tail_mean_error": 0.0019135625095341169
134 },
135 {
136 "calls": 56,
137 "cvar_abs_error": 0.0009636376286337178,
138 "global_mse": 0.0008422630745070769,
139 "tail_mean_error": 0.0009636376286337178
140 },
141 {
142 "calls": 72,
143 "cvar_abs_error": 0.0011553427163835117,
144 "global_mse": 0.000980876594478739,
145 "tail_mean_error": 0.0011553427163835117
146 },
147 {
148 "calls": 88,
149 "cvar_abs_error": 0.0011785595135815896,
150 "global_mse": 0.0010473948185824036,
151 "tail_mean_error": 0.0011785595135815896
152 }
153 ]
154 }
155 }
156}