Thanks for this proposal. I think something of this sort could end up being a very elegant solution to our qualms. We just need to ensure that changes that differ from native stake weighting preserve the EV of native stake weighting.
With this proposed modificatiion I believe there is a flaw. Say the mint rate is 1 pokt per relay, there are 1000 nodes with 15k staked, and 1000 relays per day. The total pocket minted per day would be 1000 pokt (each node receives one relay, and it mints one pokt). Then, say we implement this change, and the number of nodes halves to 500 nodes, each with 30k pokt. Now, the total number of relays each day is still 1000, but each node receives 2 relays.
Per this, each node would mint 4 pokt per day, bringing the total mint of the network to 2k pokt a day. I might be missing something, but this appears to happen.
I think there might be something there with modifying the pokt per relay parameter in accordance with the average servicer stake, but I still need to put more thought into that. I like the direction of this though–finding a very simple modification to consensus that serves as an ailment.