September 16, 2018 at 7:42 pm #30304JesusalvaParticipant
All the messages are basically under the same pattern.
If @msg-commander wasn’t busy, he could (rather easily, even hack-ish if he so wished) add a filter to reject such posts.
A more appropriate (and secure, website reputation speaking) approach would be submit every post from new accounts (< 1 day) which contains an hyperlink to moderation.
That would make SEO happy (unreachable pages and broken links won’t count for SEO), and the website would look cleaner. Only a few bots would wait some time before spamming again. (The bot could also not use the hiperlink feature, and that’ll be annoying, but it shouldn’t have an impact on SEO).
But then, if you only cared about SEO, you could just add the no-follow mark to every hiperlink. We actually do*, but as you see, it doesn’t prevents spam, but it doesn’t affects normal users nor SEO.
Honeypots help, yeah, just like a reCaptcha field (you can add one in 15 minutes tops). There are also some websites which keep lists of compromised IPs/emails for checking and etc. (can’t remember atm). In the end spam messages produced by bots have a very distinct pattern. Google reads emails, find these patterns, and decide what’s spam and what’s not. (Not good on privacy side, as you see, but does the trick.)
Antispam filters & patternsThe very first antispam filter I did was to find the number “100” on messages that did not came from my contacts. That time, most spam messages had that pattern, of using “100% discount” or whatever. It’s not common receiving a message from someone not on my contacts, let alone with the number “100” on it.
Of course, smart bots could circumvent these patterns, as an example, This questionable post. Seems legit, but yeah, that counts as spam by the book**. You can refer to This steam discussion and This one other link, about the linking reputation (which is not really bad, actually). Well, here is no rule which forbids them from doing so, unless you say that the link in specific is of a computer game and therefore, unrelated. And the poster could be a human just as well. That’s why I’m using this one as an example: It only takes so much to make a human moderator to doubt. A more advanced bot could easily make “passable” spam.
No offense intended to the poster nor the post on itself, I’m not accusing that it is spam. I am using it as an example. It was the most questionable post I could find.
But then, that’s a single questionable post in a myriad of recent blatant spam posts. That’s why a simple filter can work, and while it won’t have a 100% efficiency and need to be updated once in a while, it deals with the main bulk which harms the website.
*: It is currently used on the message body, but doesn’t applies to MSG Commander signature on rules post.
**: Technically, by the book, almost if not every off-site post from new users is treated as spam.
In the end, spam will always exist. That’s why we have moderators, there’s only so much an automated solution can do ¬.¬
Jesusalva / Jesusaves
September 16, 2018 at 9:03 pm #30310MastertrekModerator
- This reply was modified 1 month ago by Jesusalva. Reason: This is not a report, this is an analysis in hopes to be useful
@jesusalva,you’re right about that post (it can be considered as a spam so I deleted it), and thanks for that info. I also hope that MSG Commander will return and check what solution will be the best for the site to get rid of spam.
And for a very strange reason the site loads up very slowly. And I see a person who looks very suspicious yet is not considered by the site as Online at all.
You must be logged in to reply to this topic.
Members: 1,653 Forums: 121 Topics: 1,272 Replies: 8,504