The macro question of how to (a) consistently support a high performance creative environment, and (b) objectively measure the performance of knowledge workers [0] must be the most significant challenge for leaders at all levels!
On performance measurement I agree - over time all managers build a nose for it with some heuristics - if it quacks like a duck, walks like a duck etc etc.
I suspect that with high performers "class is permanent, form is temporary" and this is why the challenge is how to create (hard) and then maintain (harder?) a creative culture. Personally, I increasingly put more focus on the way the individual interacts with the environment (and how the environment is configured for them) for enabling high performance.
What I get from your comment is that you're looking for people who are willing to be creative and challenge existing solutions which I agree are two hallmarks of challenger environments. I feel this is correct, but also have no proof.
I find the 'free X' stuff trivial (and possibly a signal of entitlement) but the underlying point is that touching anything that could negatively impact the creative culture should be done carefully and after a lot of thought. People always react with more anger to a minor thing being removed, than a major thing being added.
Ultimately, if 'free stuff' is a signal of an environment that values developers contributions less then there can be no surprise if developers leave when the environment becomes less focused on enabling them. I think the point you're getting at is that if the environment is less inclusive of inventing new things/trying new things then a person whose driven to try new things will inevitably move on to the next challenge. Again, I do sort of agree but have no proof - I probably feel that people tend to be more sticky due to human loyalty etc.
Moving on a bit from what you actually said ... I was picking at something I thought you implied which was the existence of the "10x performer" and them being able to pick and choose. This is something, I'm really cautious about. I feel like the narrative of the 10x risks becoming a really negative one where we're saying "there are rock stars that are 10x better who can pick and choose and float to whatever the best environment is". Even if this is true due to PR, I don't think this is objective reality. I see good people, and even great people, but team performance is far more valuable to me than individual performance and it's factors are complex. Even for personal performance individuals do wax and wane depending on the project/culture/situation and the idea of an order of magnitude of better performance(!) - I've yet to see it on a consistent basis ... except for me of course!
[0] This forum is highly focused on developers, but I see the same pattern across multiple groups, from UX/Design out to Finance/HR. We depend on individual and team contribution, but at best find some 'objective' measures for individual contribution and wholly flawed subjective ratings (ie likeability ratings) for team contribution.