Matt Servitto: 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
84.5%z 4.38 - 2
81.1%z 4.10 - 3
70.2%z 3.91 - 4
62.6%z 3.83 - 5
61.3%z 3.81 - 6
46.1%z 3.54 - 7
45.8%z 3.52 - 8
45.7%z 3.52 - 9
45.1%z 3.48 - 10
44%z 3.41 - 11
43.2%z 3.35 - 12
43%z 3.34 - 13
42.2%z 3.29 - 14
42.1%z 3.28 - 15
41.4%z 3.24 - 16
41.4%z 3.24 - 17
41.4%z 3.24 - 18
41.2%z 3.22 - 19
40.4%z 3.17 - 20
40.3%z 3.17 - 21
40.1%z 3.15 - 22
39.8%z 3.13 - 23
39.4%z 3.11 - 24
39%z 3.08 - 25
38.6%z 3.06 - 26
38.3%z 3.04 - 27
38.2%z 3.03 - 28
38.2%z 3.03 - 29
38.1%z 3.03 - 30
38.1%z 3.02 - 31
38%z 3.02 - 32
37.9%z 3.01 - 33
37.9%z 3.01 - 34
37.9%z 3.01 - 35
37.8%z 3.01 - 36
37.5%z 2.99 - 37
37.5%z 2.98 - 38
37.4%z 2.98 - 39
37.4%z 2.98 - 40
37.2%z 2.97 - 41
37.1%z 2.96 - 42
37.1%z 2.96 - 43
36.9%z 2.95 - 44
36.7%z 2.93 - 45
36.4%z 2.92 - 46
36.4%z 2.91 - 47
36.1%z 2.90 - 48
36.1%z 2.90 - 49
36.1%z 2.90 - 50
36%z 2.89