It's a false dichotomy. The reason is "legit 50% better" is not something we currently know how to measure.
So yes, it trends towards more coarse grained measures (years of work doesn't have to be one) at the expense of the "best" performers.
If you see lots of your best performers leaving for more money elsewhere you have 2 options, raise everyone up, or come up with better coarse grained measures.
I'm of the opinion that if you can come up with good objective measures for performance of software devs, you should immediately stop doing whatever made up business problem you are currently working on, and switch to judging software devs. You could literally charge whatever you want.
So yes, it trends towards more coarse grained measures (years of work doesn't have to be one) at the expense of the "best" performers.
If you see lots of your best performers leaving for more money elsewhere you have 2 options, raise everyone up, or come up with better coarse grained measures.
I'm of the opinion that if you can come up with good objective measures for performance of software devs, you should immediately stop doing whatever made up business problem you are currently working on, and switch to judging software devs. You could literally charge whatever you want.