OK, so in short a takeover would have to involve compliance among all 17 of the top witnesses, and the 1 month delay in votes would allow time for this to be noticed and acted upon. Still not rock steady, but certainly not without merit, and with the backup of a lot of illiquid Hive. That certainly induces some more confidence.
I think something like this has to be relayed to the average inquisitive user in a way that's easily broken down and easy to understand. A sort of Hive-for-dummies type of explanation. Let's face it... this is relatively new technology and will fly over most people's heads (my own inclusive).
And remeber something else: the top positions are based on votes which can move with three clicks and no transaction fee.
This can change the order pretty quickly. If there is any indication of some kind of corruption, the order will change, new witnesses will be in the top 20 and any coalition built will be useless.
So in a sense there are 2 layers of votes that work here: those from the witnesses themselves and those from the users who vote the witnesses. Therein lies one of the original issues I mentioned, however: the possibility of buying out those who vote for witnesses (or planting a very powerful account to vote for such witnesses), plus the witnesses themselves. However, that additional layer, plus 85% majority consensus requirement, does indeed decrease the likelihood of this occurring.
This makes me think that the number of votes a witness receives should outweigh the power of a single heavyweight voter, if that makes sense. For instance, picture this scenario: some whale creates an account and purchases 5,000,000 HP, so the weight of their vote will essentially exceed all others combined and render them insignificant. Then they'd have the power to choose whichever witnesses (which could even be newly created puppet accounts) will help them based on the weight of their vote, and retain those witnesses, even though these witnesses would only have that single huge vote. Unless another whale comes in within a 1 month time frame, they'd essentially own the top witnesses. That's the power of stake at work.
If I'm understanding the voting weighing mechanism correctly, does that scenario seem like something that's actually possible, even if highly unlikely? If it is, then I wouldn't be convinced that it's a 100% decentralized platform, but as close as it can get without significant changing the voting system.
The idea of weighting lots of smaller votes slightly higher than their combined big vote has been discussed and rejected because ultimately it is prone to Sybil attack (bots).
When you start to do the maths on buying witness votes, the numbers pretty much show this is uneconomic.It was possible pre Hive, because of the centralised stake of Steemit and the collusion with the exchanges. Today I think we've closed this route down completely.
That's the kind of insight only someone experienced can provided. Thanks, that makes sense.
I'd be interested to see that discussion. I like accounting of numbers lol (you can see my posting history for proof).
What about this type of scenario: creating 17 of one's own puppet accounts (I believe one requires around 6,000 HP to claim a new account token) --> voting them in as witnesses --> increasing, say, the HBD savings APR to an outrageous figure (i.e. 100%) through votes using those puppet accounts --> taking advantage of this mechanism for only a few years --> switching HBD to Hive and selling it off for fiat/BTC/other over the course of a few years to avoid a sudden price crash.
I'm sure you guys have discussed this exact type of event, but it would be good to know that there's a fail-safe for this scenario, aside from creating another new fork on the blockchain.
It takes 63.4m HP worth of votes to scrape into the top 20. You need more than 80m to displace 3rd place. You need to fill all positions from 20 to 3.
63.4m HP is $25m worth of Hive. You simply can't buy that amount on the open market quickly: the price would start rising so dramatically we'd all love you. Couple that with the modest earnings a top 20 witness makes ($5k per month) and the considerable investment all the top witnesses have in Hive's future as a project, corrupting them for money is very hard.
You would need to deploy a whole lot of money to do this, in a way that we didn't notice and hope you get to keep any rewards.
And if you really are successful and a majority of us really don't like what you're doing, we'll fork off. We've done it before.
Now this is getting very clear as to why it's economically unfeasible, especially with the figures you provide. Even if it's done insidiously so as not to alter the price too much, it'll likely be noticed over time, thus making it a far riskier maneuver. The risk itself would be a deterrent for any whale to try to perform this, especially since the eventual selling off would potentially induce price crashes, thus defeating the purpose of their investment.
Now, isn't also true that there is a ceiling to how much HP can be owned by a single user? If the ceiling is less than the minimum to break into the Top 20 witness position, then the witness positions are actually completely protected, thus alleviating the concern I had about puppet accounts / buying out witnessed, and consequently decentralized.
There's no limits in theory or in code to HP by a single account. But there are very real practical limits to how much can be accumulated and how quickly by any single entity.