Any ideas on a better quantitative metric which one can easily get data for?
Not really. It's just really hard to try to prove what you're saying with the data available. The total number of kills can be a function of time played as much as anything. You can still get 100+ kills with a 0.5 K/D ratio if you're flying for 100 hours a month. The graph just isn't very informative with respect to pilot skill.
enlighten us then...
I'll try. Just to give a personal example by hopping into the Way Back Machine, there was a time where myself (flying as Calamari), ManeTMP, and Vulcan constituted an overwhelming majority of the Typhoon kills in the game. This wasn't because we were particularly expert in the Typhoon, but rather because literally nobody flew Typhoons back then. I'm just making up these numbers, but let's say that between the three of us, we produced 500 kills per tour in the Tiffie. Since so few people flew the plane, the total number of kills was, say, 800 - meaning that the three of us represented 62.5% of the total kills by that plane. This is without regard to K/D.
Now let's say more people started flying the Typhoon. Suddenly the number of kills doubles in a given tour, but Calamari, ManeTMP, and Vulcan continue to fly as much as they did before. They wrack up another 500 kills between them, but now the overall number of kills by Typhoon has increased from 800 to 1600. Now the three players represent 31.25% of the total number of Typhoon kills. Nothing has changed skill-wise between them, and if anything they've gotten better with experience, yet they represent a smaller percentage of the overall number of kills.