Keichobot's more advanced techniques

It is preferable to be able to use stationery without having to explain a lot, but on the other hand, it is also common for people who are familiar with the mechanics of the stationery to use it in sophisticated ways to bring out its value, and I thought that it might be better to verbalize such uses more and more.

Specific example from Is the next development a network connection or a solution to the in-drag full-size display problem?

• Branch bending (topic guidance)

  • What “fragment” is that “fragment”?

  • (This is digging too deep into something that is not the main topic.)

  • You can experiment with the arrangement of the icons by dragging them.

  • I want you to hear about “drugs.”
  • But they asked about “fragments.”
  • So I asked myself, “What is the relationship between drugs and fragments?” and then he answers the question, “What is the relationship between drugs and fragments?
  • This is a technique that Kei came up with because he’s a guy who knows that keywords that appear a lot are important, so he digs in.
    • It looks like someone who only reads the regular commentary can’t get to it.
    • Even if you realize that you’re not going to answer the question “I want you to ask about drugs, so I’m going to answer by including drugs as a keyword,” you’re going to answer something like “fragments that can be arranged by trial and error by dragging them.”
    • This would cause the word “fragment” to reappear, which would raise the score for “fragment” and raise the question to dig into the “fragment” again.
• bushkiller (species of herbaceous plant, Cayratia japonica)
  • In this chat example, the human wanted to compare two objects
  • Kei doesn’t know that there are two objects, so he digs into only one of them.
  • So I’m digging down enough on one side first to reach a realization on that branch, then I’m letting it die with a series of 🙁 ng hits.
  • It would be more immediate to use 👎NGKW, but that sounds like cutting off a branch, so I avoided it.
    • That you avoided truncating it because you might realize after the fact that “you think they’re two unrelated branches, but they’re actually related.”
    • In this case, the two branches were ultimately unrelated, so they could have been cut off, but this is hindsight bias?
    • To disclose the internal situation regarding this decision as well, NG will reduce the keyword score by 20%, and NGKW will be removed from the keyword candidate list.
      • Removing it from the candidate list means that it has not appeared at all in the past, which is a more severe treatment than having a score of zero.
        • The assumed use case is the treatment of something like the “W” in “seriously w” has been chosen as a keyword.
      • This was abused more than expected, and many people were cut off and wiped out before the branch developed, so we added a mildly effective NG command that can be abused.

---

This page is auto-translated from /nishio/Keichobotのさらに高度なテクニック using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.