@smackdown.kitty first REAL day out - charts and stats

Today the first real flags by @smackdown.kitty were made for the data gathered on yesterday 12th July comment self vote activity.

The worst of the technical difficulties have been solved, and after some tweaking the flag percentage algorithm is almost exact to counter a self vote - where we have enough SP that is! Even 100k SP is not enough to counter whale up votes 😅

We're still looking for delegations - get in touch

Top 20 self voted comments

Reminder of conditions: top 20 self voted comments, by users with over 1000 SP, ordered by pending payout of self vote contribution only.

Voter	Num top 20 SV comments	Total self votes worth	Self vote ratio
@dang007	7	$159.946	53.85%
@mrwalt	6	$47.937	90%
@pal	4	$29.388	60%
@sigizzang	1	$24.964	100%
@greatpath	2	$19.61	60%

For individual details check out smackdown.kitty on steemd.com

Note that some flags were not dealt because the payout window had closed already (thanks @demotruk for pointing this out)

All self votes

Please note: all stats are from 12 July 2017 only. Please do not jump to conclusions but look at it within it's context as a day snapshot.

There were 2655 unique users self voting at least one of their comments

Again, on comments only and from users above 1000 SP. We do not take into account voting percentage, but this is obviously in the function of pending payout (reward) also factoring their SP.

There are two main stats of interest for all users.

1. What percentage of their comment votes were for themselves
2. What was is the total pending payout for all self votes on comments

Summary stats for pending payouts per user

Edit: I forgot to do these when originally posting

For all comment self voters, here are the summary stats. Note that they are not per comment but per user for the cumulative sum on the pending payout of their comment self votes.

• Average: $0.635
• Median: $0.004

When the average is this far above the median, it suggests the data is skewed heavily to the right, i.e. there are probably a lot more low end values than high. This is consistent with the distribution of SP across Steemit (many more minnows than whales) and so maybe is not that interesting.

It does tell us that the top 20 results are extreme, and that over half of self voters do not even take more than half a cent out of the reward pool each. Practically no more than $5.31 would be paid out because of self voting between all 1326 lower half voters.

I also want to highlight that we do not even count minnows below 1000 SP in these data.

Graph of pending payouts per user

Observe how long the tail is! You can barely see the top, it fades off so quickly.

Histogram - number of users in each comment self vote percentage group

Bins are 10% wide, so the first bin is users voting for their own comment 10% of the time, and the top bin is users voting for their own comments more than 90% of the time.

Bin	Num users (bin)	Num users (cumulative)
< 10%	85	85
10% - 20%	141	226
20% - 30%	204	430
30% - 40%	190	620
40% - 50%	109	729
50% - 60%	373	1102
60% - 70%	196	1298
70% - 80%	103	1401
80% - 90%	99	1500
90% - 100%	1155	2655

It's clear that of the users who self vote their comments (and have > 1000 SP) 43.5% are doing it vastly most of the time, i.e. 90% of the time or more.

The 90% - 100% bin is by far the most populous.

I was almost surprised to find so few of comment self voting users (who have > 1000 SP) only doing it very rarely or sometimes. *17% of all comment self voters (i.e. 454 users) use self votes comment a third or less of their total comment votes.

So what we can see is that if you self vote your comments, we are 56.6% likely to do it almost all the time.

Histogram - number of users in each comment self vote pending payout group

Bins go up in multiples of 10

Bin	Num users (bin)	Num users (cumulative)
< $0.10	2205	2205
$0.10 - $1.00	278	2483
$1 - $10	138	2621
$10 - $100	32	2653
$100 - $1000	2	2655

This turned out as expected, that the majority of self votes are of low reward / payout. This is because of the so-called long tale of SP distribution across all Steemit accounts (there are lots of minnows),

Again the disclaimer - this is just for one day's statistics and shouldn't be taken necessarily indicative of the general trends

We are going to keep making these stats and probably keep the bigger trend analysis to a weekly event instead of daily. I just wanted to show our preliminary results.

What's up pussycat?

Stats ideas

I'm not currently recording the SP of users but I think this would be useful for getting a better idea of the lay of land, so to speak. It's very likely most comment self voters will have lower levels of SP, because the behavior is likely spread across the user base. But it would be interesting to confirm that.

There was also an idea today in chat by some users (especially @ financialcritic ) suggesting to use the payout as a fraction of user SP to create the top 20 list. We'll think about this further.

Reactions

We have gotten the attention of @dang007 who has flagged this post. @transisto has created @sadkitten for a yet unknown purpose in it's life, but for the moment it is being used to counter these. We're debating whether or not this is the thing to do (I'm personally undecided) but for the moment we might use it this way.

We've also had a few chats with people in the # steem-coop-general room today which have been really good as the action was happening. Join in if you want 🙂

Tomorrow is another day! 😸