7:54/3:43 = 474/223 = 2.12
log(2) 2.12 = ~1.09 ### 2 because it's distance to 2x distance ###
2000/1500 = 1.33
207 x 1.33^1.09 = 282 = 4:42 (what JI should can do... El G had a 4:41 in him)
That 2000 world best may not be as great as you think. Why would it when the holders took way more cracks at the main distances?
So, you say this wouldn't work for distances closer together? Only ones far apart, like 2x? Well let's figure the ratio log for mile and 1500
223.13/206.00 = 1.083
1609.3/1500 = 1.073
log(1.07) 1.08 = 1.13
From sprints up toward distance, this log value tends to decrease toward 1. So a good prediction for an equivalent 2000 should be a range calculated using values from 1.09 to 1.13. This would mean
206 x 1.33^1.09 = 4:41 < El G's 2000 potential < 206 x 1.33^1.13 = 4:44