• Resolved Benjamin

    (@benjaminvandenberg)


    Hi Erik!
    Hopefully you are doing okay!
    I use your plugin on multiple websites I built.

    I use a basic list that contains common english bad words. Sometimes a false positive is caused by the words I use in my bad words list. Please note that our visitors are mostly dutch.

    For example, the english bad word “offer” also occurs partly in the dutch word “offerte” (invoice). This caused a false positive.

    – Is there a way to prevent these false positives?
    – If not, do you recommend using a english list with bad words on a dutch website?
    – And also: can you recommend a good list with english bad words?

    As always, thank you for creating and maintaining this plugin!

    Kind regards,
    Benjamin

Viewing 4 replies - 1 through 4 (of 4 total)
  • Plugin Author Erik

    (@codekraft)

    Hi Benjamin (@benjaminvandenberg)!

    I hope you’re doing well! I’m sorry for any inconvenience caused by the lack of frequent updates; work has been quite demanding. However, I plan to release a major upgrade for this plugin and the SMTP one this summer. I’ll strive to be more consistent with updates in the future.

    Regarding your queries:

    1. Preventing False Positives:
      The situation you’re describing is indeed tricky because the plugin flags any instance of a “bad word” within the email content or subject. Unfortunately, the plugin does not differentiate between parts of words and full words (because i remove all the spaces and special chars from the email), so “offer” in “offerte” will be flagged, and will be flagged also “o f f e r t” or “off e rt”. To mitigate this, I recommend avoiding short words that are likely to appear in other contexts. You might consider focusing on longer or more specific terms that are less likely to cause false positives.
    2. Using an English List on a Dutch Website:
      The effectiveness of using an English bad word list depends on the nature of the spam you receive. If most of your spam is in English, it makes sense to use an English bad word list. However, if you receive significant amounts of spam in Dutch, you might need a Dutch list as well.
    3. Recommended English Bad Words List:
      Here are some strategies you might find useful, keep in mind that you can change the warning score in the main plugin page (you can find the settings under advanced):
    • High Warning Score (>0.6): Use specific phrases or names associated with spam, such as “Eric Jones” or “START EARNING NOW”.
    • Low Warning Score (<0.2): Use more generic words that are commonly associated with spam, such as “guaranteed,” “winner,” “bonus,” and “urgent.” Additionally, I sometimes create custom rules for frequent spam patterns. For example, if you notice a common spam pattern (like emails with the same subject or a sentence), you can block those more effectively.

      When the score reach 1 the email is considered spam, tune your custom settings according with this. If you want to ban with 5 bad words, you have to set a warning score of “0.2”. And remember that the score of all tests is added up so “3 bad words” and “2 fingerprints” failed will ban the IP address as well

    I found this article with some interesting ideas for the bad words list

    As always, thank you for your support and for using my plugin! 🚀

    Kind regards,
    Erik

    Thread Starter Benjamin

    (@benjaminvandenberg)

    Hi Mark,
    Great to hear you are doing well!
    Good to hear some updates are coming as well 👌

    Regarding the High and Low warning scores
    Advanced Settings shows multiple weights like “String found” and “Bot warn”.
    – Which field can I use for the calculation of each word found in a message?

    Also, the link in your message opens this forum-page, not the article you mentioned.
    – Can you send me the link of this article?

    Kind regards and many thanks!
    Benjamin

    Plugin Author Erik

    (@codekraft)

    Hello, sorry for the missing information… To be more specific the scores you are asking about:

    • ‘Bad string’ all checks where a string must match the one in the mail whether this is a string in the mail (Bad Email Strings) or a blacklisted user agent (Bad User Agent) etc. generally speaking all the fields starting for “Bad”
    • ‘Bot warn’ some automatic checks assign this score, for example the missing IP address or language check failed

    Spam words lists (sorry for the missing link was lost in some way 😥)

    https://www.pipedrive.com/en/blog/email-spam-words
    https://lix-it.com/blog/spam-trigger-words/

    Thread Starter Benjamin

    (@benjaminvandenberg)

    Hi Erik,
    Many thanks for your additional explainer about the specific scores.
    I will also check out the websites you mentioned, they look promising!

    All the best,
    Benjamin

    • This reply was modified 1 week, 3 days ago by Benjamin.
Viewing 4 replies - 1 through 4 (of 4 total)
  • You must be logged in to reply to this topic.