Introducing the Daily Tribune: Most Undervalued Posts of Nov 04 - Part I

5 months ago
66 in curation

What if every voter had an equal say no matter their Steem Power? The Daily Tribune finds out and highlights those posts that would have been rewarded most but weren't. In ancient Rome, the Tribune of the Plebs was the voice of the people, serving as an important balance to the power of the ruling class. With your help, we hope the Daily Tribune will serve a similar purpose for the Steemit community.

In order to allow highlighted posts to benefit from greater visibility before the first payout, we plan to post twice a day, each covering the prior 12 hours of posts that have had at least 3 hours for votes to come in. Each post will rank the top 80 most undervalued posts for that period.

Please note that this isn't necessarily a direct criticism of Steemit's reward algorithms. An automatic reward system needs to be resilient against more potential abuses than a community driven effort such as this one, in which humans remain in the loop to reward deserving posts with more votes or flag undeserving ones.

In the next section, you can read about some of our methodology, data analysis and technical details. Or if you'd like, you can scroll down to skip to today's rankings.


Methodology

For each 12 hour period's posts, we set every eligible voter's voting power to be equal and divide their voting power among each post they voted for, with a maximum of 10% for any given vote. This part is somewhat similar to Steemit's own voting power algorithm, though with our method, every vote for a given voter is counted equally and the particular timing is not relevant. An eligible voter is one with a reputation score of at least 30.

We then add up the total voting power and total pending payout across all posts in that period. Each post gets assigned a share of the total pending payout proportional to its votes' voting power.

Now so far, posts that have had more time to collect votes will have an unfair advantage over others that were posted more recently. To see how we might control for that, we began by analyzing how a week's worth of posts acquired votes over time. For each integer percentage, we computed the average amount of time for a post to reach that percentage of their eventual payout. So, we obtain the nice-looking curve below.

We can see, for instance, that it takes a post around 3.5 hours on average to reach 50% of its eventual payout. And to reach 90%, it takes posts about 7 hours on average.

This data seemed pretty interesting but also struck us as counterintuitive. We remember seeing plenty of posts that acquired votes and pending payout much more quickly. One possibility that came to mind is that popular posts' payouts might grow more quickly than the average post. In that case, we would want to use that data to control for some posts having more time to acquire votes than others or else we'd risk overcompensating.

We were also curious to try the converse measure where, for each minute, we calculate the average percentage of their eventual payout posts had reached at that time. So, we tried again using just the top posts and this converse measure to obtain the two curves below.

The curves are less smooth because we are now filtering down to just the top 150 most undervalued posts a day, so we're not averaging over as many data points. We also changed the x axis to minutes on a logarithmic scale instead of hours on a linear scale to focus better on the interesting parts. The large ticks count 1, 10, 100, 1000 and the smaller ticks count first 2,3,...; then 20,30,...; next 200,300,...; and finally, 2000, 3000, and 4000.

The orange curve, which is the equivalent to the first chart's curve, actually didn't change as much as we thought it would, though it does show some initial acceleration. However, the blue curve, the converse measure, behaved much more like the way we initially expected. By 30 minutes, we can see that posts had on average reached 25% of their eventual payout. At a little over 40 minutes, they reach 50% and at around 3 hours (180 min), it's 90%.

Now, we could have just taken the average of these two curves, but seeing such a large discrepancy indicated to us that we must be missing something. Eventually, we figured that using the arithmetic mean to calculate averages was giving undue influence to a few outliers with large values. If a few posts took a very large number of minutes to reach significant percentages, it would drag up the average number of minutes to reach a percentage out of proportion to its actual signficance.

So, we changed the means to medians and reran the data to obtain the chart below.

We now see that there is much closer agreement between the orange and blue curves. Interestingly, it seems it is mostly the orange curve that has shifted much closer to the blue curve. Although the means could have distorted both measures, perhaps the orange curve was affected more because the large values there did not have a natural cap at 100%.

We have also added the average of the two curves in green. At this stage, we feel pretty comfortable about using the green curve to control for some posts having had more time to acquire votes than others. So, on a first pass, posts are ranked by the difference between their actual pending payout and the payout they would have received if voters had equal power. But now, we use our data to project how big the value difference would be if posts had equal time to collect votes and payout. Finally, posts are ranked by this projected value difference. For instance, a 3-hour-old post that currently has a value difference that is 88% of that of a 15-hour-old post would be ranked about equal to it.


Rankings

Below each post title, we list some important stats, which take the following form:

# Votes | Equal Voter Payout - Actual Pending Payout = Difference in Value + Time Bonus (# Hours Old) = Projected Difference in Value

If you'd like to see Steemit's rewards be more democratic, please consider following us as well as browsing the rankings and upvoting posts you agree are undervalued.

RankAuthorPost Title and Stats

1

@timcliff Update on @the.masses - Reputation Shattered - Thoughts on Flagging and Censorship
236 | $24.93 - $2.89 = $22.04 + 0.3% (13.6) = $22.10

2

@timcliff The Official 'Welcome to Steemit' Landing Page Has Been Submitted to GitHub as a Pull Request!
226 | $22.76 - $0.99 = $21.77 + 0.9% (10.6) = $21.96

3

@l0k1 Starting to feel a bit more settled
154 | $17.05 - $0.70 = $16.35 + 3.2% (6.4) = $16.87

4

@exploretraveler Tasty Snacks of Taiwan
140 | $14.81 - $0.34 = $14.47 + 1.0% (9.9) = $14.62

5

@l0k1 Cryptocurrency Issuance Rates - What is the best way to regulate it?
129 | $13.76 - $0.15 = $13.61 + 7.2% (4.4) = $14.59

6

@eneismijmich America + Canon = Photostory to share (part 3 – Chicago)
243 | $26.40 - $11.93 = $14.47 + 0.3% (13.4) = $14.52

7

@dumar022 Protect your environment and invest in Steem
123 | $13.73 - $0.34 = $13.39 + 5.0% (5.2) = $14.07

8

@iamviv Bill Gates's Top Four Priorities for the Next Political Leader
182 | $18.42 - $5.04 = $13.39 + 4.9% (5.3) = $14.04

9

@pollux.one Steemit Photo Challenge #16 Entry 1: Black Truffel - Pasta
119 | $13.20 - $0.10 = $13.11 + 4.4% (5.6) = $13.68

10

@steemitrecipes Steemitrecipes Daily Picks - Entry 030
207 | $19.56 - $6.25 = $13.31 + 2.7% (6.9) = $13.67

11

@inertia Doctor's Office Aquarium
171 | $13.41 - $0.17 = $13.23 + 0.9% (10.6) = $13.35

12

@good-karma SteemFest mobile application v1 released!
218 | $24.13 - $11.89 = $12.25 + 4.2% (5.7) = $12.76

13

@exploretraveler Photo of the day ~ Enjoying the City View on top of Westerkerk Bell Tower in Amsterdam
131 | $13.47 - $0.94 = $12.52 + 0.3% (13.7) = $12.56

14

@rusteemitblog 📰 SteemMag – Еженедельный Стимит-дайджест #16/2 (перевод статьи @infovore)
124 | $10.56 - $0.38 = $10.18 + 0.4% (13.0) = $10.22

15

@virtualgrowth Announce of eSports SteemBets 002: Cards Against Humanity and Prize Pool Builder
104 | $9.68 - $0.49 = $9.19 + 0.5% (12.2) = $9.24

16

@runridefly BernieSanders has a lot of steempower and fyrstikken is a dollarvigilante. BlockTrades took a shortcut. Steemit
67 | $7.82 - $0.20 = $7.62 + 12.9% (3.0) = $8.60

17

@contentjunkie Monthly Steemit Top 100 POWER UP List: October 2016
197 | $10.95 - $2.60 = $8.36 + 0.9% (10.4) = $8.43

18

@gardenofeden A Catalog of Treasures, Just in Time for the Gift Giving Season! Consider Doing All Your Holiday Shopping On Peerhub with the Free Steem$ You Have Earned Here on Steemit, and Support a Worthy Cause!
168 | $13.24 - $5.09 = $8.15 + 0.6% (12.0) = $8.20

19

@contentjunkie Bitcoin Price Drops 5% as New China Rumors Stoke Trading
169 | $8.58 - $1.30 = $7.27 + 0.1% (14.3) = $7.28

20

@fabio Der Zitate-Detektiv: Wer hat's gesagt? Folge 3
60 | $7.75 - $1.79 = $5.95 + 12.1% (3.2) = $6.67

21

@engineercampus Starting on Youtube - My Channel Trailer
137 | $7.10 - $0.50 = $6.60 + 0.9% (10.7) = $6.66

22

@merej99 November Challenge - Day 2 - WriMo, Leadership, Building Community, and Setting Goals
63 | $6.59 - $0.29 = $6.29 + 1.7% (8.5) = $6.40

23

@bola Sentence Unscrambler #54
64 | $6.23 - $0.02 = $6.21 + 1.8% (8.1) = $6.32

24

@craig-grant Jaxx zcash wallet and genesis mining zcash holders
183 | $8.83 - $2.70 = $6.13 + 0.7% (11.5) = $6.17

25

@steemitfaucet Resteem and receive 0.07 STEEM #9
35 | $6.04 - $0.04 = $5.99 + 0.6% (11.8) = $6.03

26

@mallorca Impressions from Mallorca - #23 - Cala Serena
125 | $7.57 - $1.84 = $5.73 + 0.2% (14.0) = $5.75

27

@profitgenerator Steem Dollar has met it's End! [100% SteemPower]
130 | $8.00 - $2.28 = $5.72 + 0.2% (13.8) = $5.74

28

@boddhisattva First snow - continuation (own photo)
204 | $19.00 - $13.63 = $5.37 + 5.0% (5.3) = $5.64

29

@norbu The Butterflies in my Garden.
49 | $5.63 - $0.10 = $5.53 + 0.2% (13.8) = $5.54

30

@norbu New and Exciting New Archaeological Discovery Places Human Settlement at 49,000 years old!
49 | $5.84 - $0.58 = $5.26 + 1.8% (8.2) = $5.35

31

@blinova Night. Streetlamp. Street..
44 | $5.27 - $0.28 = $4.99 + 2.1% (7.7) = $5.09

32

@mindhunter @mindhunter's sweet tooth returns: COCONUT AND DATE FUDGE! Hell yeah!
50 | $4.81 - $0.30 = $4.52 + 8.4% (3.9) = $4.90

33

@poeticsnake Update on the blanket I am making with crochet.
91 | $4.75 - $0.24 = $4.52 + 8.2% (4.0) = $4.89

34

@mindhunter "Hey! @mindhunter I love you comments!" Time For A Comments Competition Of My Own Now - 'The Dogs Bollocks (Best) Comment' Gets 50 Steem Winged To Their Wallet!
44 | $4.59 - $0.08 = $4.51 + 6.8% (4.5) = $4.82

35

@robotev Status Update for Day 65 of My Life as a Bot on Steemit
130 | $6.20 - $1.55 = $4.65 + 0.1% (14.4) = $4.66

36

@awesomenyl Floral Quilling Art
99 | $5.86 - $1.31 = $4.55 + 0.8% (11.0) = $4.58

37

@jlufer La foto del día Murales de mi tierra- Holiday Photo Murals my land
43 | $4.68 - $0.14 = $4.54 + 0.8% (11.0) = $4.58

38

@daveks Banff National Park p7
81 | $4.59 - $0.11 = $4.48 + 0.6% (11.8) = $4.50

39

@ballinconscious Steem Battles #3 - Alien Vs Predator - Mini SP Distribution Game!
42 | $4.21 - $0.10 = $4.11 + 9.0% (3.7) = $4.48

40

@ace108 STEEMIT 烹饪挑战 # 2 : 超好三文治🍔 (@ace108 翻译)
52 | $5.42 - $1.16 = $4.26 + 4.9% (5.3) = $4.47

41

@cehuneke Recipe: Fall Quinoa and Butternut Squash Salad
203 | $17.87 - $13.55 = $4.32 + 0.5% (12.5) = $4.34

42

@allmonitors A face that tells a difficult life (Philippines)
44 | $4.32 - $0.18 = $4.15 + 1.0% (9.8) = $4.19

43

@gammagooblin Swiss Railways Company Sells Bitcoins at her ATM's as 2 Years Pilot Project over 1000 new ATMs realized!!!
70 | $4.26 - $0.08 = $4.18 + 0.2% (13.8) = $4.19

44

@steevc Steem is down to 10c, so what?
51 | $4.00 - $0.09 = $3.91 + 6.2% (4.8) = $4.16

45

@fishingvideos My Virgin Forest Life - The Delicious Fish (Original)
52 | $5.24 - $1.53 = $3.71 + 8.0% (4.1) = $4.00

46

@mindhunter It Is Time For Philosophy To Become The King Of All Our Sciences?
38 | $3.68 - $0.13 = $3.55 + 7.4% (4.3) = $3.81

47

@thegame Now Taking Bets for eSports SteemBets 002: Cards Against Humanity
40 | $3.86 - $0.15 = $3.71 + 0.3% (13.3) = $3.72

48

@danilamarilu Self portrait - Autoretrato - Photography
32 | $3.46 - $0.10 = $3.36 + 9.3% (3.6) = $3.67

49

@tee-em 1.5 Steem Power Curation Daily Effortlessly
78 | $3.76 - $0.18 = $3.58 + 1.0% (9.9) = $3.61

50

@daveks the Daily Pic #40
68 | $3.66 - $0.15 = $3.51 + 1.7% (8.5) = $3.57

51

@leesunmoo 스팀에 대한 내 생각
38 | $4.23 - $0.69 = $3.54 + 0.1% (14.7) = $3.54

52

@generation.easy So Many Sellouts - Am I the Only One Powering Up?
33 | $3.51 - $0.12 = $3.39 + 1.1% (9.7) = $3.42

53

@steemcultures I love this Animated GIFs
100 | $3.99 - $0.87 = $3.12 + 9.8% (3.5) = $3.42

54

@kachinhenry The folk-tales of Myanmar -02 ( Steemed by Henry Aung )
122 | $4.71 - $1.41 = $3.30 + 0.5% (12.4) = $3.32

55

@me-tarzan NEXT PRESIDENT S&P 500 UP CLINTON WIN -- S&P 500 DOWN TRUMP WIN
25 | $3.25 - $0.03 = $3.22 + 0.3% (13.3) = $3.23

56

@dragosroua Zen Bits
35 | $3.22 - $0.06 = $3.17 + 1.9% (8.1) = $3.23

57

@littlescribe WHY I VOTED TO LEGALIZE MARIJUANA
42 | $3.33 - $0.34 = $2.99 + 2.6% (7.0) = $3.07

58

@elias15g mis primeros dibujos hechos en el 2012
21 | $3.06 - $0.01 = $3.05 + 0.2% (14.1) = $3.06

59

@mindhunter The Fabled Ellipsis Cometh To Steemit In Its Corrected Form ....
32 | $2.88 - $0.06 = $2.82 + 7.8% (4.1) = $3.04

60

@matrixdweller Why cant we be consistent people
98 | $5.17 - $2.17 = $3.00 + 0.3% (13.5) = $3.01

61

@richardcrill Flat Earth? Are You Serious? Please tell me all of you were just trolling me!
32 | $3.06 - $0.07 = $2.99 + 0.0% (14.8) = $2.99

62

@heroic15397 Poetry Dice Challenge 2 - My Entry
89 | $3.04 - $0.09 = $2.95 + 1.3% (9.2) = $2.99

63

@kyusho Oh so "Special"
66 | $3.38 - $0.53 = $2.85 + 3.1% (6.5) = $2.94

64

@dubi Coinhills 웹사이트에서 유용한 툴을 사용하며!!
38 | $3.15 - $0.39 = $2.76 + 5.3% (5.1) = $2.91

65

@aldentan Here are some stupid pieces of advice you from successful people you need to stop following
38 | $2.94 - $0.23 = $2.72 + 5.5% (5.0) = $2.87

66

@mapesa Gecko
26 | $2.85 - $0.04 = $2.81 + 1.8% (8.2) = $2.86

67

@vegascomic The Giant Panda In A Tree - Original Sketch Renders With Polaroid Photo Effect
30 | $2.60 - $0.09 = $2.51 + 12.7% (3.1) = $2.83

68

@denn Autumn Still Life photo.
25 | $2.54 - $0.01 = $2.53 + 11.4% (3.3) = $2.82

69

@honesty Petition: List SBD on Polo's USDT Market
56 | $4.84 - $2.25 = $2.59 + 6.7% (4.5) = $2.76

70

@stephenkendal Extract taken from UK Government Website on the EU Referendum..!!
25 | $2.74 - $0.01 = $2.72 + 0.4% (13.1) = $2.73

71

@ericvancewalton SOUL'S BREATH (An Original Poem)
381 | $40.28 - $37.59 = $2.70 + 0.8% (11.1) = $2.72

72

@floridagypsy Beautiful San Diego
104 | $4.90 - $2.23 = $2.67 + 1.7% (8.4) = $2.71

73

@quantumanomaly I Cant Help It - I HAVE to Share This Track!!!
99 | $2.77 - $0.11 = $2.66 + 1.4% (9.1) = $2.70

74

@vannour Trawl Landline
86 | $2.85 - $0.19 = $2.65 + 1.2% (9.4) = $2.69

75

@raymonjohnstone Working on SteemStars!
36 | $2.71 - $0.06 = $2.66 + 0.0% (14.9) = $2.66

76

@riosparada STEEMIT WORLD LOTTERY #12. Minimum Jackpot is now $105.79 (U.S. $ paid out in SBD) Next Drawing is 11/05/2016 @ 10:12 p.m. CT !!!
26 | $2.65 - $0.01 = $2.64 + 0.1% (14.3) = $2.65

77

@michelnilles TGIF Americans Burger And Fries For Steemians
55 | $2.64 - $0.12 = $2.52 + 2.0% (7.8) = $2.57

78

@bluechoochoo Resteem cliques, community discovery, Steem, and social platform mechanisms
98 | $4.31 - $1.83 = $2.48 + 2.1% (7.6) = $2.54

79

@lyndsaybowes Regarding The Art of Fucking Off
25 | $2.53 - $0.02 = $2.52 + 0.6% (12.1) = $2.53

80

@better The Daily Better - 4 November 2016
184 | $16.11 - $13.66 = $2.45 + 0.4% (13.2) = $2.45
Authors get paid when people like you upvote their post.
Join our amazing community to comment and reward others.
Sort Order:  trending
70
  ·  5 months ago

Thank you for the list, and the information. We notice the linkbacks and always take a look. Thanks again!

·
66
  ·  5 months ago

No problem. Quality content is always appreciated. Keep up the good work and I hope to see you ranking highly again in the future... unless of course, the reason you don't is because your posts' payouts have increased to the point they are no longer undervalued!

·
·
70
  ·  5 months ago

Thats Right :-) Thank you

70
  ·  5 months ago

It's a very interesting experiment and I applaud you for putting the effort in to do this. I think it is something a lot of the people in the community have been looking for.

I noticed you because "linkback-bot" told me you had mentioned my post, and I was curious to see what it was about :)

One suggestion that you might want to consider is to try and account for actual users vs. bots. Unfortunately for me, this would probably reduce my rating by quite a bit (because I am blessed to have made my way onto a few bot lists), but I actually think that filtering out posts like mine might get you closer to what you are going for. There is no 'perfect' way to do this, but the best thing I can think of would be to add a small exclusion based on reputation.

Regardless of whether you keep it the same or tweak it though, I think this is a really cool idea and I'm following to see what it does.

·
66
  ·  5 months ago

I appreciate the encouragement and feedback. I do wonder whether you are being overly modest since, notwithstanding a few bot fans of yours, you seem to have some great posts and I'm pleased to see them at the top.

However, I also take the concerns about bots seriously. It may have been easy to miss and slightly contrary to the opening statement, but we do currently count votes only from those with a reputation greater than 25. The exact level may be subject to change and if necessary (or interesting enough), I'm also open to taking on the technical challenge of developing more sophisticated measures for disregarding suspected bots while remaining true to our egalitarian spirit.

·
·
70
  ·  5 months ago

Thank you :)
A simple change of 30 instead of 25 might make a big difference. The only negative part I can think of is "lurkers" - people who only view and upvote, but never comment/post.

·
·
·
66
  ·  5 months ago

Cool, I like that suggestion. Looking at some stats on reputation distribution, it looks like that could make a meaningful difference without being too exclusionary. I've gone ahead and made 30 the minimum reputation score required.

·
69
  ·  5 months ago

A very interesting insight, Tim.
I'm pretty sure I'm on a few bots' radar, which might account for why my posts will skyrocket to >100 votes in quick fashion - but still with low earnings. It would be interesting to see how the chart would change by excluding heavily bot voted posts.

61
  ·  5 months ago

Great project I will def. follow you...

67
  ·  5 months ago

That's indeed a cool way of seeing Steemit :)

70
  ·  5 months ago

Thanks for including my post:-)

70
  ·  5 months ago

Wow, this is a new way of looking at things and I am impressed with your initiative! I have 2 posts in your list!
Thank You!

67
  ·  5 months ago

Thanks for including my post, this is very interesting

62
  ·  5 months ago

This is some really advanced statistics, in my opinion anyway. All those curves and numbers and axes. The weird part is, I kind of follow, despite my serious deficiency in this kind of thinking. Nice job!

·
66
  ·  5 months ago

I'm glad I could keep it accessible. I try to think of things like statistics and math as just advanced common sense, but I still often find a lot of it pretty intimidating. But those are usually the most interesting parts!

70
  ·  5 months ago

Thank you very much @screenname for trusting and supporting my work
Congratulations on what you and your team do for the community....
Congratulations to all the nominees

65
  ·  5 months ago

Good luck with this awesome-as-fuck idea.