Are you still running your no-whale-votes experiment? I just checked, and I control posting keys for over 1 GV. Am I spoiling the experiment?
For what it's worth, my bot-clients normally only add about $0.30-$0.50 to a post, but now it's nearly $3.00: https://steemd.com/japanese/@kumada90/5g1wam
So that part of your experiment seems to be very effective.
So your army fits the "whale-like-group" definition. You have the power. Wish you are voting wisely.
When no perceived whale voting is going on I don't need to down-vote, no matter whether the experiment is "running" or not.
Let's see if the system can still work without whales up-voting.
Well, my bot targets somewhere around 80 votes per day so that each vote is supposed to be below whale-level. They're spread around to a great deal of generally-unpopular content. I think I have one of the most socially-responsible bots out there. I suppose I could increase the number-of-votes target to spread the votes out even more.
With 80 votes per day you are only voting at half strength of a whale voting 40. Starting at 1 GV that still fits abit's definition, though not by a huge margin.
So my question is: what's most important? Keeping the strength of each vote small? Or keeping the total number of rshares^2 low? Because if the former, I'll just spread the votes out on more posts. If the latter, I'd need to let some voting power go to waste.
Aren't they the same?
If you're voting on unpopular contents, likely I won't notice them..
More posts is fine IMO (although I can't speak for abit). My downvoting bot will ignore whale votes when they only apply a small vote strength to an individual post. n^2 applies at the level of individual posts. So for example, a whale voting full power on one empty post, say for 10 "vote points", would generate a reward of 100, but voting 10% power on ten empty posts for 1 "vote point" each would generate a total reward of 10. In the former case the whale's influence is extreme and crowds out others (say those voting with 1 point maximum) from having any meaningful influence at all. In the latter cases it is not. (Unscaled hypothetical units, and not exact.)
(nesting limit)
Oh, I meant total rshares^2 in all of steem. I meant
total_reward_shares2.finally see some stats re this experiment! does it mean that your clients are giving out more money without earning as much? :) the latter part is my guess - as you cannot expect many big votes to come (even if they do, they come with a curse! lol). and, yes, I think spreading more votes is a correct strategy. :)
Yeah, they're certainly giving out more money. Probably at least 3 times as much. It's hard to say anything about earning more or less money at this point. There are so many moving parts, and the revenue has never been constant. It depends on my vote stalkers, Curie's policies, my vote weights, the current accuracy of my bot's model. My revenue hasn't been stellar in the past 2 days, but my bot also hasn't yet learned that the world has changed. If the "experiment" runs long enough, it will be interesting to see how my bot adapts to it.
thanks for reply. actually I am more curious about how the 7-day window is gonna affect curating strategies... while the experiment is temporary, this change has been confirmed (though HF date postponed to Mar 21). your thoughts?
Yeah, it's going to be interesting. I think 7 days is going to make curation slightly more difficult. Really, just automated curation. Manual curation will be what it always has been.
Any adaptive automated curation system like mine will now have to wait much longer for feedback. Right now, my bot checks each post 2 days after it was posted, records its payout, and learns from it. Once HF17 goes into effect, it will have to wait 7 days to see the final payout. The nice thing is that most votes will probably still come in fairly early in a post's lifetime, so I'll be able to partially learn something at the 1 or 2 day mark anyway, and then update it once the post finally pays out at 7.
I don't think current witnesses will let the 7-day change pass as is.
@biophil
that's interesting to know. thanks!