Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Monday October 10 2016, @01:26AM   Printer-friendly
from the fun-with-numbers dept.

Since the launch of SoylentNews in February of 2014, there have been 274,870 comment moderations made against the 412,100 comments that our community has posted to our site. Who has posted the most comments? Who garnered the most up-moderations? The most down-moderations?

Such simple questions, but they led to a fun bit of DB querying. The results surprised me, and I thought others might be interested, as well. Most surprising to me was the assessment of comments from Anonymous Cowards.

[Continues...]

Who received the most moderations?

For better or worse, to whom did Soylentils direct their greatest moderation effort?

NICK UID TOTAL DOWN UP NET
The Mighty Buzzard 18 2260 626 1634 1008
takyon 881 2315 103 2212 2109
aristarchus 2645 2494 615 1879 1264
c0lo 156 2717 183 2534 2351
Thexalon 636 3225 83 3142 3059
Ethanol-fueled 2792 3447 1238 2209 971
VLM 445 4401 346 4055 3709
Runaway1956 2926 4531 992 3539 2547
frojack 1554 5855 593 5262 4669
Anonymous Coward 1 78936 13002 65934 52932

The single greatest target of moderation was the "Anonymous Coward" with 78,936 moderations. This was followed by frojack, Runaway1956, VLM, Ethanol-fueled, and Thexalon who garnered over 3000 moderations each.

Who had the most down-moderations?

Here, only the number of down moderations was considered — it mattered not whether it was Flamebait or Troll — they all counted the same.

NICK UID TOTAL DOWN UP NET
VLM 445 4401 346 4055 3709
Hairyfeet 75 1620 387 1233 846
MichaelDavidCrawford 2339 1513 387 1126 739
frojack 1554 5855 593 5262 4669
aristarchus 2645 2494 615 1879 1264
The Mighty Buzzard 18 2260 626 1634 1008
jmorris 4844 2144 753 1391 638
Runaway1956 2926 4531 992 3539 2547
Ethanol-fueled 2792 3447 1238 2209 971
Anonymous Coward 1 78936 13002 65934 52932

Once again, our prolific AC topped the list with 13,002 down-mods. Ethanol-fueled was the only other user who topped 1000 down-mods, coming in with 1238. Runaway1956 made a valiant showing with 992 down-mods.

Who had the most up-moderations?

In the eyes of the community, who most often received an up-mod? Again, no consideration was given for the nature of the up-mod — Insightful, Interesting, or Informative — all were considered the same.

NICK UID TOTAL DOWN UP NET
aristarchus 2645 2494 615 1879 1264
Phoenix666 552 2184 80 2104 2024
Ethanol-fueled 2792 3447 1238 2209 971
takyon 881 2315 103 2212 2109
c0lo 156 2717 183 2534 2351
Thexalon 636 3225 83 3142 3059
Runaway1956 2926 4531 992 3539 2547
VLM 445 4401 346 4055 3709
frojack 1554 5855 593 5262 4669
Anonymous Coward 1 78936 13002 65934 52932

Once again AC reins supreme with 65,934 up-mods. This was followed by frojack with 5,262 and VLM with just over 4000.

Who had the highest net-moderation?

Putting it all together — subtracting the number of down-mods from the number of up-mods — who had the highest net moderation on our site?

NICK UID TOTAL DOWN UP NET
wonkey_monkey 279 1754 117 1637 1520
maxwell demon 1608 1786 55 1731 1676
Phoenix666 552 2184 80 2104 2024
takyon 881 2315 103 2212 2109
c0lo 156 2717 183 2534 2351
Runaway1956 2926 4531 992 3539 2547
Thexalon 636 3225 83 3142 3059
VLM 445 4401 346 4055 3709
frojack 1554 5855 593 5262 4669
Anonymous Coward 1 78936 13002 65934 52932

Once again, the shy but prolific AC tops the list with a net of 52,932 mod points. Only one other Soylentil was able to surpass 4000: frojack with 4,669. Two other Soylentils exceeded 3000: VLM with 3709 and Thexalon with 3059.

Who hath pointy horns?

Who managed to acquire the most down-mods as a percentage of all moderations on their comments? For a tie, number of moderated comments is the second sort field. Who is the devil in our midst?

NICK UID TOTAL #DOWN %DOWN #UP %UP NET
scarboni888 5061 1 1 100.00 0 0.00 -1
MooCow 6048 1 1 100.00 0 0.00 -1
cybergimli 436 2 2 100.00 0 0.00 -2
rancidman 769 2 2 100.00 0 0.00 -2
rmdingler 1038 2 2 100.00 0 0.00 -2
SoylentsISay 1331 2 2 100.00 0 0.00 -2
stupid 2631 2 2 100.00 0 0.00 -2
contrapunctus 3495 2 2 100.00 0 0.00 -2
killal -9 bash 2751 5 5 100.00 0 0.00 -5

Pfft, just a few minor imps around here. killal -9 bash topped (bottomed?) the list with 5 down-mods out of 5 moderations.

Who earned a Halo?

Whose comments had the best percentage of up-mods to total-mods? And in the case of ties, received the most up-mods? Who are the angels among us?

NICK UID TOTAL #DOWN %DOWN #UP %UP NET
dx3bydt3 82 69 0 0.00 69 100.00 69
romlok 1241 70 0 0.00 70 100.00 70
Hawkwind 3531 75 0 0.00 75 100.00 75
jdccdevel 1329 78 0 0.00 78 100.00 78
rleigh 4887 102 0 0.00 102 100.00 102
DrMag 1860 103 0 0.00 103 100.00 103
SrLnclt 1473 117 0 0.00 117 100.00 117
Joe 2583 126 0 0.00 126 100.00 126
Aiwendil 531 164 0 0.00 164 100.00 164

Here, it appears we've got a flock of angels, or at least people who know which way the wind blows. All folks listed here scored 100.00% meaning all of their moderations were up-mods. Aiwendil topped our list with 164, and we had 4 others — Joe, SrLnclt, DrMag, and rleigh — who each had over 100 such comment moderations... not even a single down-mod among them!

I must admit I was surprised to see the sheer number of positive moderations of AC comments, and the fact that 83.5% of those mods were positive.

[Update: Added two tables, one each for top percentage of down-mods and of up-mods. -Ed.]

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Informative) by The Mighty Buzzard on Monday October 10 2016, @02:49AM

    by The Mighty Buzzard (18) Subscriber Badge <themightybuzzard@proton.me> on Monday October 10 2016, @02:49AM (#412279) Homepage Journal

    Just for clarity's sake, we don't store ip addresses of users; anonymous or otherwise. We store a one-way hash of the address and use that. Still useful for this purpose but not capable of outing anyone's location.

    --
    My rights don't end where your fear begins.
    Starting Score:    1  point
    Moderation   +1  
       Informative=1, Total=1
    Extra 'Informative' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 2, Interesting) by toph on Monday October 10 2016, @03:23AM

    by toph (5509) on Monday October 10 2016, @03:23AM (#412290)

    Just for clarity's sake, we don't store ip addresses of users; anonymous or otherwise. We store a one-way hash of the address and use that. Still useful for this purpose but not capable of outing anyone's location.

    With only 4 billion IPV4 addresses, inverting the hash with a rainbow table is trivial. For IPV6, the hash should work well enough.

    • (Score: 2) by The Mighty Buzzard on Monday October 10 2016, @03:48AM

      by The Mighty Buzzard (18) Subscriber Badge <themightybuzzard@proton.me> on Monday October 10 2016, @03:48AM (#412297) Homepage Journal

      True enough, which is one of the major reasons why we don't make them public. We could always add a salt into the mix and make the rainbow table argument moot but there hasn't been a pressing reason to since they're not publicly shown.

      --
      My rights don't end where your fear begins.
      • (Score: 0) by Anonymous Coward on Monday October 10 2016, @04:19AM

        by Anonymous Coward on Monday October 10 2016, @04:19AM (#412304)
        Different IP address get different salts but always the same salts for the same IP address?

        There might not be that many AC IP addresses...
        • (Score: 0) by Anonymous Coward on Monday October 10 2016, @10:05AM

          by Anonymous Coward on Monday October 10 2016, @10:05AM (#412362)

          Even if you had unique salts for every IP address it would still be insecure because it wouldn't be that hard to crack an IP address if the salts and the hash list leaked.

          Perhaps to make it more secure you can use a combination of IP address and hostmask + unique salt but given the fact that hostmasks tend to follow predictable patterns I would argue that's still not that secure, especially if someone had a list of all the hostmask/IP address combinations.

          Even with IPv6 addresses it's possible that they follow patterns and various unused or not yet used addresses can be added to the list of addresses not to attempt to crack.

          IP addresses were not meant to provide cryptographic security through being an address that's hard to crack, that was never the intent and it probably won't be good for that purpose.

          I wasn't considering privacy when I wrote my above AC comment about posting those AC statistics. Obviously storing IP addresses removes privacy but storing a hash instead doesn't add that much more privacy. It makes it more inconvenient for hackers at best, which any degree of inconvenience is good but won't be good enough to thwart a determined entity with resources and access to tons of information (ie: from ISPs).

        • (Score: 0) by Anonymous Coward on Monday October 10 2016, @10:09AM

          by Anonymous Coward on Monday October 10 2016, @10:09AM (#412363)

          Also this wouldn't be possible without either

          A: Storing the IP address itself. Now you have a list of all IP addresses that post. It would be very trivial for someone with that list and with your hash to figure out who you are.

          B: Having some kind of algorithm to go from IP address to salt and from salt + IP address to hash. That algorithm is ... another hash function, easy to crack.

        • (Score: 2) by The Mighty Buzzard on Monday October 10 2016, @11:00AM

          by The Mighty Buzzard (18) Subscriber Badge <themightybuzzard@proton.me> on Monday October 10 2016, @11:00AM (#412372) Homepage Journal

          In theory, adding IPV6 addresses to the pile of IPV4 addresses before hashing easily solves this. In practice given how little IPV6 traffic there is? Someone besides me would need to do the math. I'm a crypto user not a crypto guru.

          Worrying about the salt slipping out isn't much of a worry though.

          1. Your browser is a much more porous attack surface than the rehash code, our network, or Apache.
          2. Any TLA isn't likely to be given much pause by a little TLS, and there's all your traffic wide opened for them to see.
          3. Anyone with the access necessary to read the salt where it would likely be stored could just as easily turn access logging on or even rewrite our code to email them a daily list correlating addresses to usernames.

          I'd personally be more worried about the email addresses for everyone leaking. That's really the most sensitive bit of information we store (assuming you didn't do like I usually do and give president@whitehouse.gov whenever possible). Thankfully, I have yet to hear a credible report of the slashcode we started with having any SQL injection vulnerabilities (aside from the one I found that is utterly useless for actually pulling anything off) and we've made a point to keep up those particular best practices of our coding forefathers since forking.

          --
          My rights don't end where your fear begins.