Fitness

wouldn’t that paint a pretty distorted picture?

It can still paint an accurate picture, if understood within scope. You can't outrun a bad diet with normal amounts of exercise. Even if it is true that you can outrun a bad diet with exercise that exceeds a particular threshold, it's still a useful finding that under that threshold, the correlation drops to near zero.

Anecdotally, I know that my appetite adjusts naturally to my activity level, for the most part. The exception was during periods of my life when I was running more than 20 miles (32km) per week, where my calorie consumption hit a plateau even when my calorie expenditures exceeded that. I also tended to lose weight when backpacking in forests and mountains, but part of that was diet, too (of eating only what I packed in).

So if it turns out there is an amount of exercise that allows for a bad diet, it might very well be such a high activity level that it isn't relevant to broad population-wide public health principles.