Peter Benchley: 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
84.2%z 4.36 - 2
79.4%z 4.01 - 3
66.7%z 3.87 - 4
61.6%z 3.82 - 5
52.7%z 3.72 - 6
51.6%z 3.71 - 7
51.3%z 3.70 - 8
47.9%z 3.66 - 9
47.7%z 3.65 - 10
47.5%z 3.64 - 11
47.5%z 3.63 - 12
47%z 3.60 - 13
46.6%z 3.58 - 14
45.9%z 3.53 - 15
45%z 3.47 - 16
44.2%z 3.42 - 17
43.5%z 3.38 - 18
43.4%z 3.37 - 19
43.2%z 3.35 - 20
43%z 3.34 - 21
43%z 3.34 - 22
42.9%z 3.33 - 23
42.5%z 3.31 - 24
42.1%z 3.28 - 25
41.9%z 3.27 - 26
41.6%z 3.25 - 27
41.6%z 3.25 - 28
41.6%z 3.25 - 29
41.6%z 3.25 - 30
41.2%z 3.23 - 31
40.9%z 3.21 - 32
40.6%z 3.19 - 33
40.3%z 3.17 - 34
40.1%z 3.15 - 35
39.8%z 3.14 - 36
39.8%z 3.13 - 37
39.6%z 3.12 - 38
39.5%z 3.12 - 39
39.3%z 3.10 - 40
39.3%z 3.10 - 41
39.1%z 3.09 - 42
38.9%z 3.08 - 43
38.3%z 3.04 - 44
38.3%z 3.04 - 45
38.2%z 3.03 - 46
38.2%z 3.03 - 47
38.1%z 3.03 - 48
37.8%z 3.00 - 49
37.7%z 3.00 - 50
37.6%z 3.00