Peter Roskam: raw computed results
The raw ranking from the production similarity calculation (face-embedding centroid cosine → gender-cohort z-normalization). The web-opinion crown, curation, and fusion re-rank are removed.
- 1
85.6%z 4.47 - 2
64.9%z 3.85 - 3
64.3%z 3.85 - 4
60.7%z 3.81 - 5
48.2%z 3.67 - 6
47.8%z 3.65 - 7
47.6%z 3.64 - 8
46.4%z 3.57 - 9
46.3%z 3.56 - 10
45.8%z 3.52 - 11
45.6%z 3.51 - 12
44.9%z 3.47 - 13
44.5%z 3.44 - 14
44.2%z 3.42 - 15
43.6%z 3.38 - 16
43.5%z 3.38 - 17
43.3%z 3.36 - 18
42.6%z 3.31 - 19
42.3%z 3.30 - 20
42.2%z 3.29 - 21
41%z 3.21 - 22
41%z 3.21 - 23
40.6%z 3.19 - 24
40.6%z 3.19 - 25
40.4%z 3.17 - 26
40%z 3.15 - 27
40%z 3.15 - 28
39.9%z 3.14 - 29
39.9%z 3.14 - 30
39.6%z 3.12 - 31
39.3%z 3.10 - 32
38.9%z 3.08 - 33
38.8%z 3.07 - 34
38.7%z 3.07 - 35
38.7%z 3.06 - 36
38.6%z 3.06 - 37
38.4%z 3.05 - 38
38.4%z 3.05 - 39
38.3%z 3.04 - 40
37.9%z 3.01 - 41
37.9%z 3.01 - 42
37.8%z 3.01 - 43
37.8%z 3.01 - 44
37.8%z 3.00 - 45
37.8%z 3.00 - 46
37.6%z 3.00 - 47
37.1%z 2.96 - 48
37.1%z 2.96 - 49
37%z 2.96 - 50
36.4%z 2.92