Patrick Prosser: 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
83%z 4.26 - 2
74.4%z 3.96 - 3
65.2%z 3.86 - 4
48.2%z 3.67 - 5
44.7%z 3.45 - 6
43.4%z 3.37 - 7
42.6%z 3.31 - 8
42.5%z 3.31 - 9
42.3%z 3.30 - 10
42.2%z 3.29 - 11
42.1%z 3.29 - 12
42%z 3.28 - 13
42%z 3.27 - 14
41.8%z 3.26 - 15
41.7%z 3.26 - 16
41.7%z 3.26 - 17
41.5%z 3.25 - 18
41.4%z 3.24 - 19
41.4%z 3.24 - 20
41.2%z 3.22 - 21
41.2%z 3.22 - 22
41%z 3.22 - 23
40.7%z 3.19 - 24
40.5%z 3.18 - 25
40.4%z 3.17 - 26
40.3%z 3.17 - 27
40.3%z 3.16 - 28
40.1%z 3.15 - 29
39.8%z 3.13 - 30
39.7%z 3.13 - 31
39.4%z 3.11 - 32
39.2%z 3.10 - 33
39.1%z 3.09 - 34
39.1%z 3.09 - 35
39%z 3.08 - 36
38.8%z 3.07 - 37
38.7%z 3.07 - 38
38.6%z 3.06 - 39
38.5%z 3.05 - 40
38.4%z 3.05 - 41
38.3%z 3.04 - 42
38.2%z 3.03 - 43
38%z 3.02 - 44
37.9%z 3.01 - 45
37.8%z 3.00 - 46
37.7%z 3.00 - 47
37.7%z 3.00 - 48
37.6%z 3.00 - 49
37.4%z 2.98 - 50
37.3%z 2.98