I've noticed that now that the validator bug has been fixed, the current voting system has become a kind of race to the bottom. It looks like the best strategy for delegates is to have the least amount of votes while remaining a validator. This seems a less than optimum way to drive security and demand for semux.
An idea I've thought about is what would happen if instead of all the validators getting the same reward (i.e. block frequency), there were several levels. For example, if the 100 validator positions were divided into roughly thirds.
Level 1 (validators 71-100) could have the same setup as is used now, with any delegate capable of reaching at least the 100 position getting the same rewards as any other validator up to the 71st position. Having the same rewards across all positions in level 1 would maintain some stability at this lowest entry point level.
Level 2 (validators 30-70) would have rewards scaled (the fun level
), so that 69 would get slightly more blocks than 70, and 68 slightly more than 69, and so on, up to position 30. This scaled reward would eliminate the race to the bottom in terms of votes. It would instead make it a race to the top, increasing security, and increasing demand for semux.
Level 3 (validators 1-30) would be similar to level 1, in that all the validators get the same reward, but unlike level 1 the uniform reward is equivalent to the highest reward achievable in level 2. Having the top level rewards the same would help prevent centralization of semux votes at the very top positions.
I know it's probably too late to change now, and I also may be blind to a fatal flaw in this kind of scenario. But, I think it's interesting to think about how this kind of setup could operate from a game theory perspective in a way to drive demand for semux instead of people trying to get the minimum amount of votes while remaining a validator.
Just my 2 satoshis
