Earlier quoted context omitted.
If you have a hand of cards you want to optimize it makes sense to keep discarding and redrawing your worst cards. At the beginning your worst cards are probably below average, so any new replacement is statistically likely to be better. But even when you reach the point where all your cards are above average it can make sense to keep redrawing a small percentage, since that's the only way to get more aces. Of course…
I kind of see that, but it completely ignores the cost of hiring (financial and in lost productivity in the team that has to onboard the new hire) and the loss of knowledge and experience from firing as well.
Not defending, just explaining.