Mark Kerr: 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
91.3%z 5.08 - 2
73.4%z 3.94 - 3
56.1%z 3.76 - 4
53.2%z 3.72 - 5
46.6%z 3.58 - 6
46.1%z 3.54 - 7
45.1%z 3.48 - 8
44.4%z 3.43 - 9
43%z 3.35 - 10
43%z 3.34 - 11
42.5%z 3.31 - 12
42.3%z 3.30 - 13
42.2%z 3.29 - 14
42.2%z 3.29 - 15
42%z 3.28 - 16
41.1%z 3.22 - 17
40.5%z 3.18 - 18
40.4%z 3.17 - 19
40.2%z 3.16 - 20
40%z 3.15 - 21
39.9%z 3.14 - 22
39.7%z 3.13 - 23
39.7%z 3.13 - 24
39.5%z 3.12 - 25
39.4%z 3.11 - 26
38.6%z 3.06 - 27
38.1%z 3.02 - 28
37.6%z 2.99 - 29
37.6%z 2.99 - 30
37.6%z 2.99 - 31
37.5%z 2.99 - 32
37.3%z 2.98 - 33
37.1%z 2.96 - 34
36.9%z 2.95 - 35
36.6%z 2.93 - 36
36.6%z 2.92 - 37
36.4%z 2.91 - 38
36.3%z 2.91 - 39
36.3%z 2.91 - 40
36.1%z 2.89 - 41
36%z 2.89 - 42
35.9%z 2.88 - 43
35.8%z 2.88 - 44
35.7%z 2.87 - 45
35.7%z 2.87 - 46
35.5%z 2.86 - 47
35.5%z 2.85 - 48
35.4%z 2.85 - 49
35.3%z 2.84 - 50
35.2%z 2.84