It's pointless to propose reason against arguments that are solely based on biased outliers.
Honest example that are kind of embarrassing:
I do or have done many Fortitude things in real life. Some still regularly. Specifically, I've been in three fist-fights this year. I've had my ass handed to me in one of them, to the degree that I thought I received permanent damage to my eardrum. Categorically, I've failed 33% of the time, yes? But, I tried 100% of the time in those encounters to resolve the situation with Fortitude things.
In my kwoon, I've met a lot of individuals that have no other observable Fortitude skills, except their martial art. They're good in set piece strikes and blocks, but when advanced students slip in something unexpected they fail Fortitude rolls very quickly, especially when they're just admins or couch potatoes the other six days a week. They lack a body of experience that makes them an adept at Fortitude things despite having ranks in Fighting with Hands.
When I'm not getting my ass kicked, I do system engineering and architecture. I have mad skills, but I'm terrible at math. Like, hardly competent at some traditional maths*. This makes a significant number of development tasks cumbersome for me. I know a fuck-lot of kernel-level Linux and can dizzy most people with my grasp of network protocols and regular expression. But, I can't idly wrap my mind around using those skills to design something new without a quiet place and a lot of personal time (taking 10).
Some people will read those statements above and isolate one sentence to dismiss the idea that skills equal overall competency in a particular Attribute. Others might see that I might be better in one category than another. Ultimately, fuck you for implying your view point is better than mine because of one particular edge case. It's just a game.
* - Eugenia Cheng is a heroine for being able to explain categorical mathematics. Check her out.