You are talking about the severity of a bottleneck , the term bottleneck describes the underlying limiting behavior. I am simply indicating to you the definition of the word bottleneck . I wouldn't say that a 3% bottleneck is something to worry about but it fits the definition nonetheless just the same as a 15% bottleneck which obviously would be a greater cause for concern.
Showing a couple of games where a 6core cpu doesn't substantially limit performance does not refute my conditional statement . I would say for most people and most games a 9600 would be fine but that is a subjective assessment which an individual would need to determine based on their own use case . It does not mean that in certain games you would not have problems .
https://www.gamersnexus.net/hwreviews/3407-intel-i5-9600k-cpu-review-vs-2700-2600-8700k
please direct your attention to the data on far cry 5 for one example of what i am talking about . I have noticed similar behavior in other games in my own testing of a 9600k vs higher thread count cpus .
The article linked is also a good illustration of why 1% lows ,as you quoted, do not tell the whole story of end user performance. Frametime plots are more reliable generally but the way they are presented in the toms articles are too condensed as too be difficult to assess.