« first day (1601 days earlier)      last day (2202 days later) » 

9:00 PM
@SmokeDetector why
 
@Glorfindel thanks I'll have a look now
 
[ SmokeDetector | MS ] Potentially bad keyword in answer: Why 403 forbidden, only if i run direct on browser by user88083 on webmasters.SE
 
@GrantGarrison why do you think
 
Title - Position 27-33: Turkey
Body - Position 161-167: Turkey, Position 190-196: Turkey
tpu- by JAD
 
@JakeSymons It's likely to get a lot of down-votes, but just because it happens to be illegal in some particular jurisdiction (with some exceptions for US law compliance, see law references and the TOS) doesn't make it off-topic.
 
9:01 PM
!!/test reddit.com/r/Turkey
 
> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username
----------
Title - Position 14-20: Turkey
Body - Position 14-20: Turkey
Username - Position 14-20: Turkey
 
!!/test discord
 
> Would not be caught as a post, title or username.
 
!!/test discord.gg
 
> Would not be caught as a post, title or username.
 
9:02 PM
Discord is a social site, not spam
 
So, unless it is classified under another close reason, such as too broad or unclear then it would be on topic?
 
And blacklisting it is a horrible idea. Especially with the discord tag on SO
 
@SFTP Watch the specific one?
 
@GrantGarrison they're invite links, they're generated once
 
Not necessarily
 
JAD
@GrantGarrison why? It's already caught with another keyword
 
The specific page may be watched/blacklisted, same for the specific reddit maybe?
 
@GrantGarrison The MS goes 50/50 on discord
 
9:03 PM
You can use a permanent invite link on Discord
 
@JakeSymons MS records are biased towards spam.
 
@PrincessLuna true but they're not going to for this
@JakeSymons do not even think about blacklisting discord
 
@quartata Just a certain link
 
@quartata I was not thinking that, I was giving the 50/50 to show that it had the same number of fps if anything to give reason so that it would not be blacklisted.
 
9:04 PM
!!/test discord.gg/Hcx7dUP
 
[ SmokeDetector | MS ] One-character link in answer, potentially bad keyword in answer, blacklisted user: Does SEO value get transfered when a 301 redirect points to a page that has another canonical? by Tuncay seren on webmasters.SE
> Would not be caught as a post, title or username.
tpu- by Grant Garrison
 
!!/watch discord.gg/Hcx7dUP
 
@GrantGarrison You don't have code privileges, but I've created PR#1977 for you.
 
wut - weight 67 and already autoflagged?
 
3 reasons
 
9:05 PM
@Glorfindel Already?
 
JAD
@GrantGarrison forgot to escape the .
 
@JakeSymons Just discord isn't a good metric though. Because in addition to the chat site, there's also the english word. Most of the tp's seem to be separate from the chat site
 
Change that to /. right?
 
@JakeSymons bear in mind that if you go by Metasmoke saying discord is 50/50 spam, it also says the word "the" is 85% spam
10
 
JAD
\.
 
9:07 PM
@JAD Oh lol I why did I say the opposite?
 
Okay, so maybe MS is not the best source of reference to use so at the last count there have been 4,402 across the network with only 11 of those being tps.
 
!!/watch discord\.gg/Hcx7dUP
 
@GrantGarrison You don't have code privileges, but I've created PR#1978 for you.
tpu- by Jake Symons
 
sd - poof-force
 
it had 3 reasons
 
9:09 PM
1/3/21 is p good
catches all this garbage
 
#loopholes
 
@Glorfindel nah man
 
That's not one of my actual conditions, I should clarify. But that's a thing.
 
3 reasons has like 99.9% accuracy everywhere
you don't need to limit it to one site
 
9:11 PM
@SmokeDetector Anyone want to close this?
 
tpu by Zoe
tpu- by Jake Symons
 
@doppelgreener not that you'd be able to actually use that condition anyway
 
huh, you're right
@ArtOfCode i wouldn't? it's over 99.75%
 
@doppelgreener not over enough posts though
68 < 1000
 
oh huh okay.
 
9:12 PM
yeah that's why you can't do that on small sites
 
I really don't see "maximum poster reputation" as a parameter that should exist.
 
@GrantGarrison I think you need to escape the forward slash too.
 
IMO rep > 1 => no autoflags
 
@SFTP yeah that one's weird to me.
 
@K.Dᴀᴠɪs because managing that correctly is basically impossible
 
9:14 PM
[ SmokeDetector | MS ] Luncheon meat detected: Active menu item in node.js and express by lortschi on stackoverflow.com
fp- by Glorfindel
fp by Floern
[ SmokeDetector | MS ] Potentially bad keyword in body, potentially bad keyword in title: Hey you! Yes yes you that is watching..come visit reddit.com/r/Turkey :)) by notvladevidac on superuser.com
tpu- by Glorfindel
 
holy hell I wrote a lot into a github issue response >.<
 
Something should be upgraded from "potentially", it seems.
 
@ArtOfCode @Undo comment requested on my response to the one issue ticket that was opened for a feature request. needs sanity review for implementation "general" discussion points, as well as MetaSmoke compatibility review. Before anyone creates the changes (and if we want this, assign the issue to me as I already have the logic written in another Python window and can incorporate fairly easily in another branch.)
 
Obviously not the country as such
 
@SmokeDetector What did I miss? How is that a '2' weight with 2 autoflags?
@SmokeDetector autoflagged
 
9:16 PM
@K.Dᴀᴠɪs That post was automatically flagged, using flags from: iBug, WELZ.
 
@K.Dᴀᴠɪs iBug and WELZ have funny conditions
 
Interesting
 
So, any Travel post with Turkey in title and body gets it?
 
@K.Dᴀᴠɪs wtf
 
@SFTP no, probably they have activated that only on SuperUser.
 
9:17 PM
^
they're per site
although I was pretty sure WELZ's was disabled
 
@iBug Your flag conditions were manually disabled by @ArtOfCode: Total accuracy must be >= 99.75%.
 
kicking things
 
tpu- by K.Dᴀᴠɪs
@K.Dᴀᴠɪs Your flag conditions were manually disabled by @ArtOfCode: Total accuracy must be >= 99.75%.
 
wtf was wrong with my conditions
 
> Total accuracy must be >= 99.75%.
 
9:19 PM
According to MS, it is
 
> Total accuracy
not accuracy for each individual condition
accuracy for all of them taken together
 
really??
 
@K.Dᴀᴠɪs I don't believe so.
 
that is something we probably need visibility into
because that is not at all how i understood that
 
@doppelgreener [status-would-like-to-but-can't]
 
9:20 PM
The url reddit.com/r/Turkey is not found on either Travel or Expats (or History or Politics), so should be safe to blacklist.
 
@doppelgreener makes sense to me
 
@quartata it doesn't make sense to me because i don't even know how that would get calculated
if each of my conditions is above 99.75%, how does the total drop below 99.75%
 
magic
 
and how can i adjust for a total i'm not aware of
 
It doesn't make sense to me either, and I'm probably the only MSc in Mathematics in the room </brag>
 
9:21 PM
So......... I have to manually set it up per site?
 
lies, damn lies, and statistics
 
@Glorfindel PhD here. :)
 
@SFTP I should have suspected that.
 
Condition was disabled: 99.85%
Condition was disabled: 99.90%
Condition was disabled: 99.76%
 
0
Q: It seems like you want that DDOS against metasmoke. You know we have a JS LOIC app and enough people to bring your software to its knees

shouldwestartdestructionIt seems like you want that DDOS against metasmoke. You know we have a JS LOIC app and enough people to bring your software to its knees

 
9:22 PM
@Glorfindel because not all conditions apply to all posts
 
Ok people! There is a name for this! It is called Simpsons Paradox!
 
You can have two flag conditions that would flag two distinct sets of posts A and B, and while A and B are both >= 99.75%, A+B is not
 
tpu- by Jake Symons
 
9:23 PM
@ArtOfCode you know how to turn on magic funride right
 
tpu- by K.Dᴀᴠɪs
 
@quartata flick the setting?
 
Sounds like we're in for a fun night ^
 
Simpson's paradox, or the Yule–Simpson effect, is a phenomenon in probability and statistics, in which a trend appears in several different groups of data but disappears or reverses when these groups are combined. It is sometimes given the descriptive title reversal paradox or amalgamation paradox. This result is often encountered in social-science and medical-science statistics and is particularly problematic when frequency data is unduly given causal interpretations. The paradoxical elements disappear when causal relations are brought into consideration. It has been used to try to inform the...
 
@ArtOfCode Then what is your recommendation? Set my reason weight to 800?
 
9:26 PM
unrelated if I EVER post something to a Smokey GH issue again with that much text... someone needs to smack me.
 
@ThomasWard I want to smack you just because. I don't need a damn reason
runs away
 
@ThomasWard it's ok it happens to us all
 
@ArtOfCode I used the new MS channels signup and waited for hours but still not working.
 
@K.Dᴀᴠɪs You're going to need to describe in a little more detail which userscript you're looking for in order for me to know which one it is.
 
@Makyen nvm I found it :)
 
9:30 PM
:p
 
@quartata everyone subscribed to the Smokey issue tickets is going to hate the size of what I wrote though
 
@ArtOfCode did Undo make a setting for that one?
anyways
@ThomasWard nah
 
well i burned an hour of time at work on this either way so...
 
[ SmokeDetector | MS ] Potentially bad keyword in answer: Converting country codes in .NET by mohammad zahmatkesh kenari on stackoverflow.com
[ SmokeDetector | MS ] Username similar to website in answer, blacklisted user: Cannot install Elementary Tweaks in Loki by ysfl on elementaryos.SE
fp- by K.Dᴀᴠɪs on Converting country codes in .NET [MS]
tpu- by Grant Garrison
naa- by doppelgreener
 
9:39 PM
@SmokeDetector huh? that one was already reported and tpu'd 5 hours ago? what happened?
 
@SmokeDetector doesn't look like actual spam, just naa.
@Floern the question was edited.
 
@doppelgreener No such command 'ignore'.
 
uh
what's the ignore post command?
 
tpu by Jake Symons
 
@doppelgreener Undisclosed affiliation.
 
9:40 PM
@doppelgreener ignore is a type of feedback.
 
@Makyen they're just linking to a forum post they made
that's not advertising
 
@doppelgreener yes, but why was it not red flagged 5 h ago?
 
@Floern what's red flagged mean?
 
spam nuked
 
because it's not spam
@SmokeDetector ignore
 
9:42 PM
@doppelgreener Post ignored; alerts about it will no longer be posted.
 
ah, so we're tpu'ing but not flagging?
 
@Floern i didn't tpu it at all
4 mins ago, by SmokeDetector
naa- by doppelgreener
i don't think anyone should tpu that, but we can make up our own minds, and we're each responsible for what we choose to flag
 
@doppelgreener Which by SE's standards, they need to disclose that it's something on a site they control/that they wrote. Our Feedback Guidance is that such links must be disclosed, or they are tpu. That doesn't mean that they should necessarily be flagged as spam. In the past, the option was to leave a comment to the user informing them of the SE policies. Now, it's custom mod-flag and let the mod do it, if desired.
3
 
I count 7 tpus.. shrug
 
@Makyen considering they're not advertising or promoting a product or service, just linking to a forum post (that is not, itself, advertising anything either) i don't see it as falling within the bounds of self-promotion and thus needing to engage in disclosure of affiliation. if it were their blog, i'd feel differently, but it's just their post on some random forum.
 
9:47 PM
@Floern I assume that people are seeing that it's not deleted and keep giving feedback. That's only appropriate if the post should be deleted, which this one probably shouldn't be. IMO, the user should be educated, not significantly chastised.
 
it's a crap answer, but it contains zero promotion and thus our guidelines about how to do promotion don't apply
 
11 mins ago, by SmokeDetector
[ SmokeDetector | MS ] Username similar to website in answer, blacklisted user: Cannot install Elementary Tweaks in Loki by ysfl on elementaryos.SE
@SmokeDetector why
 
Post - Username similar to website
Blacklisted user - blacklisted for //elementaryos.stackexchange.com/a/15297 (https://m.erwaysoftware.com/posts/by-url?url=//elementaryos.stackexchange.com/a/15297) by the metasmoke API
 
oh ~_~
 
@doppelgreener The domain, e.ysfl.cn matches their username. I think its a bit more than just some random post they wrote.
 
9:48 PM
i see, their domain name = their name = promoting their post on their forum
@SmokeDetector tpu-
that's that then
 
!!/watch demono
 
They have 3 posts all with the same website in, all have been caught by SD and have been feedbacked as 2 tpu and 1 naa
 
@YvetteColomb Added demono to watchlist
 
so... now I just retracted my flag after you said the post was fine, and then change your mind again to it's really spam? grrr...
 
!!/test ysfl
 
9:50 PM
> Would not be caught as a post, title or username.
 
!!/watch ysfl
 
@ByteCommander sorry, i didn't see the display name.
 
Restart: API quota is 9840.
 
i've raised my own spam flag on their other spam post though, and a custom mod flag on that one.
 
9:51 PM
!!/watch ysfl
 
@GrantGarrison You don't have code privileges, but I've created PR#1979 for you.
 
@GrantGarrison While that combination of letters is rare, that feels a little broad. Perhaps you meant ysfl\.cn.
 
Its quite rare, there's only one StackOverflow post so its possible
 
@Makyen I think that it will work.
!!/help
 
@GrantGarrison I'm SmokeDetector a bot that detects spam and offensive posts on the network and posts alerts to chat. A command list is available here.
 
9:57 PM
How do you scan without reporting if detected?
Oh well I will just delete if it catches it.
 
do !!/test <paste-body-here>
 
oh
!!/test This is my first time asking a question about R, so forgive me if I leave out details that are necessary. I'll try to be as thorough but concise as possible.

My data has 4 columns: score (values of -1, 0, or 1); species (at least 10 different ones); habitat (three different ones); and site (at least 8 different ones). So each row is a unique combination of score, species, habitat, and site. See below for top rows:

score species habitat site
0 NOCA NAG NAG1
0 BRTH NAG NAG1
!!/tea
 
@GrantGarrison brews a cup of earl grey tea for @GrantGarrison
 
Did that get skipped?
!!/test structure(list(species = structure(c(27L, 5L, 7L, 2L, 18L, 12L,
22L, 28L, 6L, 13L, 15L, 23L, 4L, 10L, 14L, 11L, 20L, 30L, 9L,
19L, 26L, 8L, 16L, 25L, 3L, 17L, 21L, 29L, 24L, 1L, 27L, 5L,
7L, 2L, 18L, 12L, 22L, 28L, 6L, 13L, 15L, 23L, 4L, 10L, 14L,
11L, 20L, 30L, 9L, 19L, 26L, 8L, 16L, 25L, 3L, 17L, 21L, 29L,
24L, 1L, 27L, 5L, 7L, 2L, 18L, 12L, 22L, 28L, 6L, 13L, 15L, 23L,
4L, 10L, 14L, 11L, 20L, 30L, 9L, 19L, 26L, 8L, 16L, 25L, 3L,
17L, 21L, 29L, 24L, 1L, 27L, 5L, 7L, 2L, 18L, 12L, 22L, 28L,
@ByteCommander Um... that's not working very well.
 
10:05 PM
/shrug
 
@GrantGarrison Post Error when attempting pairwise comparisons following lme does not look like spam.
 
maybe it breaks on multilines...
 
@SmokeDetector No meat found?
 
It's broken
 
Ah. OK. Hmm...
 
10:07 PM
[ SmokeDetector | MS ] Few unique characters in answer, no whitespace in answer, repeating characters in answer: blender game: spawn object by Gamerzangjack 21 on blender.SE
tpu- by Floern
 
4
Q: Is anything spam anymore?

Paul CrovellaI recently flagged this answer as spam It's an ad for iPistis and (since that's not enough) doesn't disclose the author's obvious affiliation: This flag was declined stating: a moderator reviewed your flag, but found no evidence to support it Is evidence looked for? If this isn't spam...

 
[ SmokeDetector | MS ] Potentially bad keyword in answer: What’s the origin of the Weird Sisters in Macbeth? by Thunder on mythology.SE
fp- by ByteCommander
 
10:23 PM
@SmokeDetector k
Nicola? Apparently they ran out of female names
> In the UK and Germany, the male form Nicola has been used as a female name, which originally came from incorrectly presuming the name's vowel ending "a" was the female form, as has also occurred with the male given name Andrea.
TIL
 
As in "Nicola Tessla".
 
10:52 PM
!!/watch themasters-golf\.ca
 
@doppelgreener That pattern looks like it's already caught by Bad ns for domain in answer and Bad ns for domain in body; append -force if you really want to do that.
 
!!/watch livemasters\.ca
 
@doppelgreener Added livemasters\.ca to watchlist
 
!!/watch frozen-four\.ca
 
@doppelgreener That pattern looks like it's already caught by Bad ns for domain in answer and Bad ns for domain in body; append -force if you really want to do that.
 
10:53 PM
!!/watch frozenfour2018\.ca
 
@doppelgreener Added frozenfour2018\.ca to watchlist
 
!!/watch-force frozen-four\.ca
 
@doppelgreener Added frozen-four\.ca to watchlist
 
!!/watch-force themasters-golf\.ca
 
@doppelgreener Added themasters-golf\.ca to watchlist
 
10:54 PM
!!/watch themasterss\.org
 
@doppelgreener Added themasterss\.org to watchlist
 
!!/watch wrestle-mania34\.ca
 
@doppelgreener Added wrestle-mania34\.ca to watchlist
 
thanks for the tall order, smokey.
 
CI on 7577dfd succeeded. Message contains 'autopull', pulling...
Restart: API quota is 9377.
 

« first day (1601 days earlier)      last day (2202 days later) »