Of course you are correct that adding votes to a witness improves the ranking but only at the expense of other lower ranking witnesses. Let's say that I remove all my votes for the top 20 and only vote for 5 that are not on that list. Those five witnesses will improve their position but will push down another five who are not on my list. In the meantime the top 20 can vote trade and amplify their influence.
I am with you on distributing the voting but I don't agree that we should be advocating for using only a few votes.
The current scheme is not representative of the actual stake if you do not use all of your votes. I better setup would be to redestribute your voting weight e.g. if I have 100 vests and vote for 1 block producer my vote is worth 100 but if I vote for 2 then my vote is worth 50. Our current governance model favors those that use all of their votes. That's why the amount of votes matter.
There is no 'pushing down' in backup witnesses, they get assigned blocks related to the votes they have, not the position. More votes for one means a bit less rewards for all the others and a bit more for him.
The idea that more votes split the weight wouldn't solve anything, as that'd affect everyone. Theoretically, everyone could even have unlimited votes and the system would work fine. I won't get into the details about that now though because I'm pretty annoyed by your way of trying to explain me I'm wrong when you don't grasp half of what you talk about.
I am not trying to annoy you, I apologize if I did. When I have the time I will make a post about this (including the math behind it). Thank you for taking the time to respond.
You've been moving the goal post along the way. Your initial claim that the amount of votes given right now would matter is still not true.
#20 has 60M Hive voting for him. That's twice of what's locked up on binance. There is no current threat that requires piling votes there, it only reinforces the results we had before which were mostly decided by a few big stakeholders as you correctly mentioned.
So what is the effect of switching a vote exactly?
Removing a vote from a top witness: doesn't influence anything at all, as long as it doesn't make it drop out and someone else get in
Adding a vote to a top witness: doesn't change anything except making it harder for backups to replace them
Removing a vote from a backup witness: lowers the rewards for that one, raises the rewards for all other backups
Adding a vote to a backup witness: raises the rewards for that one, lowers the rewards for all other backups
If you use more votes, you actually lower the influence you have on the individual witnesses you vote for. So if you really only want to support a few and fill it up with others, you actually lower your support for the important ones. And in the situation where an attack isn't to be expected it doesn't make sense either. One could spread them at the top of the top to secure their position when necessary, but during times like this where we're reasonably safe that just reinforces decisions by other big stakeholders as said a few times by now.
--
Regarding the number of votes and stuff, @raycoms did an analysis recently and I've seen numbers for other models too. Nothing would really improve the situation, as anything goes in two directions - if it's easier for the community to prevent a supermajority, a big stakeholder can also block hardforks easier. In the end stake is defining and the only real security measure is a good distribution so buying amounts big enough to be a threat is unreasonably expensive.
I agree with you on the fact that we need to re-align the witness votes and that currently we do not have any probable threats (maybe I should have started with that).
I don't disagree. For now let's leave it at that and thank you again.