Just a couple of comments...
1) If your upgrade path does NOT include a 2080 Ti or whatever Ampere card will be at the top, then these gaming benchmarks are useless. It's like comparing a GTX 950 with a RTX Titan on a Athlon 3000G. The bottlenecks go both ways.
2) If you own a RTX 2080 Ti, you sure as hell ain't gonna consider buying any of these two CPUs. If you got the money for just the RTX 2080 Ti, you're better off just getting a 2070 super and you'll have enough leftover money to jump straight to an i9 or Ryzen 9.
3) What all these benchmarks show, is that at that point, when the benchmark ran, the 1600AF was faster. That is true. What they don't show, is how the 9100F vs 1600AF on the long run. How about a 8h/day typical workload sim? For many people, that's the scenario. How about how easy it is to build a system with these two CPUs? System stability, virtualization benchmarks. You talk "productivity" but seriously, 7zip, Cinebench and Blender? That's all the productivity you can do with a PC?
Disclaimer: I decided to go with an i3 9100F, after having 5 of my friends coming over to my place, because their Ryzen systems would act up on daily basis. RAM compatiblity issues, overclocking where they shouldn't have... I think Intel's K and non-K philosophy actually makes sense. What's the use of having an unlocked CPU if you have no idea what to do with it? That, right there, is 80% of the people. That's why big integrators don't use overclocking chips, nor do their BIOSes allow overclocking. Simple as that