Mark Boucher: 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
86.2%z 4.52 - 2
76.8%z 3.98 - 3
70.1%z 3.91 - 4
63.6%z 3.84 - 5
49.4%z 3.68 - 6
47.7%z 3.65 - 7
47.5%z 3.63 - 8
47.4%z 3.63 - 9
46.7%z 3.58 - 10
46.6%z 3.58 - 11
46.5%z 3.57 - 12
46%z 3.54 - 13
45.8%z 3.52 - 14
45.8%z 3.52 - 15
45.4%z 3.50 - 16
45%z 3.48 - 17
44.7%z 3.46 - 18
44.7%z 3.46 - 19
44.6%z 3.45 - 20
44.4%z 3.43 - 21
44%z 3.40 - 22
43.9%z 3.40 - 23
43.3%z 3.36 - 24
43.2%z 3.35 - 25
43%z 3.34 - 26
42.6%z 3.32 - 27
42.6%z 3.32 - 28
42.3%z 3.29 - 29
42.2%z 3.29 - 30
41.8%z 3.26 - 31
41.7%z 3.26 - 32
41.3%z 3.23 - 33
41.1%z 3.22 - 34
41%z 3.21 - 35
40.9%z 3.21 - 36
40.9%z 3.21 - 37
40.9%z 3.20 - 38
40.8%z 3.20 - 39
40.5%z 3.18 - 40
40.3%z 3.17 - 41
40%z 3.15 - 42
39.9%z 3.14 - 43
39.8%z 3.14 - 44
39.8%z 3.13 - 45
39.5%z 3.12 - 46
39.3%z 3.11 - 47
39.3%z 3.10 - 48
39.2%z 3.09 - 49
39%z 3.08 - 50
39%z 3.08