Marc Duret: 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
89%z 4.75 - 2
86.3%z 4.53 - 3
82.4%z 4.21 - 4
74.3%z 3.95 - 5
68%z 3.89 - 6
64.2%z 3.84 - 7
60.1%z 3.80 - 8
58.9%z 3.79 - 9
55%z 3.74 - 10
51.3%z 3.70 - 11
50.5%z 3.69 - 12
50.4%z 3.69 - 13
49.7%z 3.68 - 14
48%z 3.67 - 15
47.2%z 3.62 - 16
46.5%z 3.57 - 17
45.8%z 3.53 - 18
45.8%z 3.53 - 19
45.7%z 3.52 - 20
45.4%z 3.50 - 21
44.3%z 3.43 - 22
43.9%z 3.40 - 23
43.5%z 3.38 - 24
43.4%z 3.37 - 25
43.2%z 3.36 - 26
43.2%z 3.35 - 27
42.9%z 3.34 - 28
42.6%z 3.32 - 29
42.6%z 3.32 - 30
42.5%z 3.31 - 31
42.3%z 3.30 - 32
41.5%z 3.25 - 33
41.1%z 3.22 - 34
40.6%z 3.19 - 35
40.6%z 3.19 - 36
40.5%z 3.18 - 37
40.3%z 3.17 - 38
40.3%z 3.17 - 39
40.3%z 3.16 - 40
40%z 3.15 - 41
40%z 3.15 - 42
40%z 3.15 - 43
39.8%z 3.14 - 44
39.8%z 3.14 - 45
39.7%z 3.13 - 46
39.6%z 3.12 - 47
39.4%z 3.11 - 48
39.4%z 3.11 - 49
39%z 3.09