Matt Good: 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
91.1%z 5.04 - 2
88%z 4.67 - 3
82%z 4.18 - 4
81.3%z 4.12 - 5
73.4%z 3.94 - 6
67.1%z 3.88 - 7
55.5%z 3.75 - 8
54.6%z 3.74 - 9
47.7%z 3.65 - 10
47.1%z 3.61 - 11
46.1%z 3.54 - 12
46.1%z 3.54 - 13
45.5%z 3.51 - 14
45.3%z 3.49 - 15
44.1%z 3.41 - 16
43.9%z 3.40 - 17
43.8%z 3.40 - 18
43.8%z 3.39 - 19
43.6%z 3.38 - 20
43.4%z 3.37 - 21
43.2%z 3.36 - 22
42.9%z 3.33 - 23
42.8%z 3.33 - 24
42.5%z 3.31 - 25
41.6%z 3.25 - 26
41.4%z 3.24 - 27
41.4%z 3.24 - 28
41.3%z 3.23 - 29
41.2%z 3.23 - 30
40.7%z 3.20 - 31
40.7%z 3.19 - 32
40.5%z 3.18 - 33
40.2%z 3.16 - 34
40.2%z 3.16 - 35
39.8%z 3.14 - 36
39.8%z 3.13 - 37
39.7%z 3.13 - 38
39.7%z 3.13 - 39
39.3%z 3.10 - 40
39.3%z 3.10 - 41
39%z 3.08 - 42
39%z 3.08 - 43
38.9%z 3.07 - 44
38.8%z 3.07 - 45
38.6%z 3.05 - 46
38.4%z 3.04 - 47
38.3%z 3.04 - 48
38.3%z 3.04 - 49
38.2%z 3.03 - 50
38.1%z 3.03