If what you're saying is correct (and it's not), then, for example, a wild betta x a fancy one would have "better" fins and color than continuing to line breed fancy bettas x fancy bettas (from the same line) for desired characteristics.
Out-crossing to a wild fish would basically require starting from scratch in terms of selecting for a stabilizing the desired characteristics for that line of fish.
The foundation of your confusion seems to be the notion that the characteristics that wild fish exhibit are the ones that aquarists also seek - That wild fish are necessarily "better and stronger." Decades of "fancy" angelfish (discus, livebearer, etc, etc.) line breeding would argue otherwise, as aquarists seek fish lines that look nothing like wild fish.
Bright orange discus and platies with Mickey Mouse patterned tails have goofy made up names too
How? Read and listen...
Combining fish with unrelated bloodlines doesn't only mean the resulting offspring will be better and stronger.... It pretty much guarantees it. Any undesirable traits and flaws would be picked up in the culling process...that's just how it works. Any serious breeder knows that working with the same bloodlines and using the same formula long term is a recipe for failure. It's actually almost the complete opposite to this guy's explanation of how breeding bloodlines work. No new genes= no new improvements.
Normally only a hand full out of the thoudsands of fry even make it past a few days and or weeks, and in most cases very few ever even survive to adulthood. So I'm not sure how much weight his whole generations of inbreeding in the wild bit holds.. but the chances would be slim at best under most circumstances. In captivity however, sure they inbreed like heck it's been happening for ever which has led to some of worse monstrosities in some species ever seen. Why defend that? Why uphold regression?
The notion that inbreeding a species in captivity for generations assures the best results as opposed to using new blood is just plain baffling.