Thanks @richhorn.

Yes, this Steemit experiment is very unique and, at the same time, fairly complex, considering all the algorithms that go into deciding the payment distribution. It seems to me that the system is too easy to "game" for the whales, especially when considering that they (the whales) are very likely, IMO, to have two or more accounts with high Steem Power.

One solution to the unbalanced up-voting system that I've just recently pondered, but which might be too drastic of a tweak for the majority of the Steemit community to accept, is making the "accumulated time spent on a post" the weighing factor of how much Steem gets distributed to any one post, getting rid of the up-vote altogether (curation rewards being weighted to go mostly to the early readers of any post that accumulates a lot of combined reading time). It's just too easy with Steemit's current system to up-vote an article that we think will get a lot of up-votes, without really liking the content AND, most importantly, without PENALTY for up-voting something that we really don't value. With this system of voting by time spent on the post, at least we now have to put some effort/ energy into sending the message that we like it, as opposed to only having to take the second required to click on an up-button. I think people would become more inclined to start focusing more on the content that they really like and, in the process, distributing Steem according to actual value, instead of perceived popularity (due mostly to gaming the current system, IMO).

One problem I see with this is that very long posts may receive higher payouts, regardless of content quality, due to the reader having to spend more time to finish "consuming" the content. Furthermore, high quality content can be relatively small in terms of word count, so the time spent on a good post can be short. However, I don't think this would be too much of an issue in the long-run since people are far less likely to read all the way through poor quality content and they will, of course, be less likely to read future content from authors with sub-par posts; hence, the majority of accumulated time spent will be on the higher quality content - where it belongs.

On the flip side, it might encourage posters to cover their topics more thoroughly, which should lead to higher quality content, one would think.

Added after an edit

As far as Steem Power (SP) goes: every X SP can be an additional 1 second added to read time for every Y time that goes by while still on a post (call this product the "reading points") - with some capped time limit on any one post. There, IMO, should be a capped limit on how many reading points any one account can vote towards a post and a severe drop off on curation payout beyond a set "accumulated reading points" on a post.

As an example:

We may decide to go with an additional 1 second worth of "reading points" per 10 seconds spent on a post (call it "amplifying period"), per 1,000 SP (1/1,000th of a second per 10 seconds spent on a post, at 1 SP) - capping the total reading points at 12 hours worth of seconds (43,200 reading points) to any one account.

So, if an account worth 30,000 SP were to spend 15 minutes on a post, it would have benefited that post:

[(15 minutes * {60 seconds per minute/ 10 second per "amplifying period"}) * {(1/1,000th of a second per 1 SP) * (30,000 SP)}] = 90 * 30 = 2,700 reading points.

Now, what all these values should be, from the "amplifying period" to the increase in voting time per SP to the capped limits, would likely require some experimentation and tweaking, but I'm convinced that this system would be much fairer to the minnows and still fair to the whales, since higher SP still has its obvious benefits, albeit perhaps not as attractive as with the current payout system.