Seems like sound logic.
I wonder whether this may lead to some kind of cascading effect, where the best time to leave is just after your accumulated rewards are paid out. So maybe when that payment event is triggered it provokes a few (or a lot) of nodes to leave? Just poking at the potentially sensitive aspects of this mechanism, I’m not really sure how significant this will be in real life. Maybe having payouts staggered between nodes is less sensitive to mass exodus?
What happens to accumulated reward that are not paid out because a node departs unexpectedly? Do the unclaimed rewards return to the section pool of rewards?
One thing that isn’t clear to me is why the motive to ‘stay around’ leads to ‘permanent data’. It would be possible to drop unpopular data and if it’s not requested during the time the node holds it then there is no consequence for the node.
Overall I like the idea in the quoted chain of logic, it seems quite elegant.
Yes, good to have this reminder, the design is not permanent it’s a way to get started on testing.
Hopefully we can see the results of these simulations sometime, sounds fascinating.