(Real life) Kill ratios are grossly misrepresentative of the quality of a jet as a real adversary has never been found. A well-trained, numerically equal, adversary. Not to say they are bad jets.
What is the Top Gun school / Red Flag etc. kill ratio of the F-15 for example? Probably classified and a lot lower.
I get what you are saying, but IMO real life kill ratios matter quite a bit. Those reflect the wars that have been fought in the last 40+ years. Do they reflect true peer on peer conflict with the best from the first world? No, but they reflect reality. The days of the king sending out his best knight to deul the other kings best knight have passed some time ago.
The real world scores of the F-15, F-16 and Sea Harrier are impressive by ANY measure and are important to the debate. They might not be peer on peer data, but are the best we have for real data. Yes exercises data helps too, but we can not miss the data we have from real conflicts.
We have learned tons of lessons from ground, sea and air conflict since 1945. By your reasoning they should not count because they were not against a peer foe?
These posts tend to focus to focus on weapon vs. weapon (F-35 vs PAK etc), but a better focus is on doctrine, support and the whole picture. We may (hopefully) never know how gen 5 planes fight each other in a real peer war. My money is still that the best "system" will have the edge. If that is the F-35 supported by tankers, AWACS, jammers, and the best training and support you can buy, it will be pretty good. US doctrine is not based on a "fair" knife fight.