OK, I see, you're suggesting that curation rewards be independent of relative voting time, but still leave curation rewards linear. It's an interesting idea. It leaves an incentive to vote (to get curation rewards), but it removes the incentive to try to be the first to upvote content that's expected to be popular.

In theory, of course, the current system was designed to encourage early voting and discourage late voting, as late voters get much less curation rewards. The idea was that this would tend to put a natural cap on how much posts would get upvoted and also to encourage upvoting of content that everyone would like, not just the voter.

But I would agree that in practice, this has mostly failed, in part because many people don't know how to maximize under those rules. And for those that do, it's become clear that it is easier to optimize via a bot set to vote for popular authors versus the potential losses associated with not voting enough manually. In theory, a dedicated manual voter who wants to maximize his rewards can beat the bots on returns, but I think it's just too much work.

Bots setup to vote before the 5 minute were also supposed to discourage automated voting of popular authors by bots, since it was supposed to generate a race towards the bottom on curation rewards (with bots voting ever earlier to capture more of a shrinking curation pie), but I don't see this occurring either (admittedly, I've only done eyeball observations using hivestats.io, I haven't spent any time analyzing all the curation data, but from what I've seen, I think I'm correct on this). Again, it's probably mostly just laziness, even by the bot operators.

Maybe it should be no surprise: in a wealthy society, people often pass up opportunities for money that require effort. Generally speaking, the larger stakeholders on Hive are probably in higher income brackets. And OTOH, small stakeholders aren't really that incentivized to work harder either, because the curation rewards are much smaller.

Doing some quick napkin math, I guess this shouldn't be surprising: a stakeholder with 100K USD worth of hive (a substantial stakeholder, somewhere between 400-500K HP right now), at best they might be able to get about a 20% APR by careful and consistent manual voting. They can probably increase that a little by adding in a bot to help. Alternatively, they can probably get 10% or so just by using a bot at the moment. So full time manual curation is only going to net you around 10K USD per year (vs just using a voting trail). That's a lot of work for the return (for someone who has 100K they can put into such a volatile investment as crypto). So for those who do manually curate, it's going to be driven more by considerations beyond just curation rewards.

Your idea is probably worth trying on a trial basis, but it's probably contentious enough that it wouldn't be easy to get consensus on it. I suppose this is one of those cases where something like SMTs would be nice as a way to do a trial run.