It’s obviously completely disallowed to read that on your own. Reading it comes with your banishment from the book clubs.
13 Likes
Thanks to the Hataraku Saibou (Cells at Work!) accident in the ABBC, now I believe judging a book based on the first three or four pages doesn’t often yield accurate results. I think character count and natively level (with enough feedback) are good indications.
16 Likes
Although I do plan to participate for once since I nominated the book, I can’t commit to running the book club. Hopefully someone else doesn’t mind taking care of it.
3 Likes
This is a really good question. I think the difficulty ratings were a good idea in the early days of the book club but I don’t think they’re needed any more.
Estimating the difficulty from the first few pages is very different to grading the difficulty of the book after finishing it. It’s not uncommon for the first few pages to be deceptive. Perhaps a main character appears who uses a difficult dialect, perhaps there are passages that use very formal speech patterns or technical terms.
Also, the difficulty system is based on the perceived difficulty to the rater so it’s skewed by which individuals rate the book. An experienced reader will mark a book as no effort, whereas a beginner may mark the same book as substantial effort. The Natively system is much better as it grades in comparison to other books the person has read.
I’m not surprised that the Natively levels and difficulty ratings don’t match up well.
Even if a book only has only been rated by one user, the chances are it will have a Natively level in the broadly correct range.
Calculating the difficulty rating also adds to the complexity of setting up a poll (and making a nomination). I think it is time for it to go.
18 Likes
See, nowadays I finish a book before recommending it in ABBC…
Theoretically I made manga analyzer to combat this, though even that’s not foolproof of course.
7 Likes
As someone relatively new (dipping in since 3 years, but active in the last year) to the ABBC and BBC, I agree with Micki’s points. And to this point in particular, several people have already said they make their decision based on a combination of their own read of the sample pages and the Natively rating. Someone else saying a book is easy or hard doesn’t help me because I guess the people voting have anywhere from N5-N3 grammar.
So I personally find the numerical rating of what other people think difficult to parse to be frank. Is a 2.2 going to be ok for me but a 2.9 not? Whereas with Natively, no matter how imperfect, I can easily tell that something in the teens will be easy/ok, something in the 20s is probably fine with more support needed at the upper end, but the 30s are out of reach right now without more work than I’m willing to put in. That’s all just to say, even if the rating is off by 5 points, which seems to be a swing some books experience before/after book clubs, that’s ok really.
Here’s an idea, maybe, if we still wanted to keep an internal rating, especially for books that have very few ratings to avoid a single voter skew, instead of a 1-5 scale, people could be polled on the range of likely Natively rating, e.g., <14, 15-19, 20-24, 25-29, 30+, and a “I don’t know” option. Instead of averaging it out, just choose the center of whatever hump emerges, e.g., “likely difficulty 20-24” or if it’s confusing/few responses, “unknown difficulty, possible 25-29”
Then whether there is a natively reading or not, we have a comparable scale. There could even be a brief explainer, e.g., for people who haven’t read a lot and don’t have a feel for natively, we could make some useful suggestions about what they experience looking at the sample pages and how to translate into the Natively reading. Natively roughly ties their ratings to N5-N1 levels, so this shouldn’t be too hard to do. I know it’s not perfect, but it’s just a bit more objective, and a small improvement is all we need.
Yeah, that was a book I was also interested in but didn’t vote for due to reasons related to this difficulty discussion. For me I categorise the nominations as books vs manga, and full vs partial furigana. So far in my limited looking into sample pages and my own reading, books are more difficult than manga (even to some extent with a similar natively level) because the number of words and unique words seems generally higher. Partial vs full furigana will change the difficulty in strange ways depending on the reader’s kanji progress and whether they read paper vs digital (ease of lookups). A book with partial furigana (like that book that just dropped out of the BBC) is on the higher end of difficulty on both metrics, and feels like something that is in between the BBC and IBC. It’s outside of my comfort zone, but looks to be too easy for the IBC which is why it got moved here.
There were other similar conversations that led to the pre-IBC trial going on right now with the slower reread of Convenience Store Woman. For me, this book would fit into the “books that are too challenging for BBC but too easy for IBC”. But maybe there doesn’t need to be a pre-IBC club. Maybe split the BBC into the BBC and BMC - Beginner Book and Manga Clubs respectively. That might naturally resolve this pre-IBC issue as there would be more “book” book club picks at the upper beginner level.
17 Likes
I quite like this idea - as others have said, I believe the numerical ratings are skewed and I even I feel unsure if I should vote at times (am I used to reading harder material by now? Then if I saw minimal effort, I could discourage a beginner from reading it, etc.)
As for the whole BBC/IBC conversation… much was said previously, and I have to say I don’t remember everything. But I believe one of the topics was the difficulty of finding enough books that fit the “upper beginner” label. There was also the questions of how much interest/added value an additional club would bring, of course. But as someone taking part in the IBC primer, I have to say it’s been quite enjoyable so far and helped me made the choice to also jump into a “normal” IBC book club. Indeed, the gap between BBC and IBC is quite large (or overwhelming?) for now.
14 Likes
I agree, if you are going to ask users to rate the difficulty of nominations this would be a good way to do it. You probably need to have read a few books before to have a feel of the Natively levels, but I think that’s fine. The users who are most likely to be able to judge the book’s difficulty are those who have read a few books previously.
I’m not 100% sure it’s needed. Even if the Natively level is based on a single user’s feedback it’s feels unlikely to be more than 5 Natively levels away from its correct position.
As a cut off perhaps you could only put this poll in the nomination post if the Natively level has a question mark against it.
8 Likes
I don’t mind taking care of it 
16 Likes
ah, good point, maybe that’s the solution, then, to read some previous IBC picks slower and build up to IBC pace. That’s probably better than splitting up clubs.
that sounds really sensible
hero!
6 Likes
I don’t think anyone addressed this, but it’s quite easy to request books on Natively and the only ones likely to get turned down are ones without covers or logistical data, which I think would be pretty unlikely to be nominated to club since previews are part of the nomination process. Even then, the only recent nominee I can think of that fits that bill 雪の女 could be added with the free pdf version since those are commonly accepted.
6 Likes
The issue with “not being on natively” is that you have absolutely no data on it then, yes getting them added to natively is relatively easy, but to be entirely honest, unless it is a hugely anticipated release, or there is an anime going, or whatever reason, there won’t be many people that will go ahead and read it on the spot, so you won’t get a lot of ratings quickly anyways.
Well, I guess Natively itself gets round this problem in a few ways. The person who requests it to be added can give an estimated level, or if they don’t then Natively staff give it an estimated level (EDIT - they only assign it a default level based on book type - see seanblue comment). It then has a question mark next to the level to show it is a tentative assignment.
Once some gradings are done, the level will be adjusted but continues to have a question mark until a certain threshold is crossed.
Given that our current difficulty rating system has poor correlation with Natively levels, this still seems a better and easier approach than what we do at present.
And we also have the option of polling within the group to give it an estimated Natively level of our own if we want.
3 Likes
But if they haven’t read the book they’re just guessing, which is arguably worse than doing nothing. Or in the best case scenario, they’re making a guess based on a short sample like is currently done here and is equally flawed.
Which to be clear, is just a default value depending on the book type. For example, all manga are level 24 by default. There is no analysis done to give it a “smart” estimated level.
For the record, I’m not defending the current poll-based difficulty rating system. Just saying that I think the default estimated levels on Natively are nearly worthless.
5 Likes
By the way, planning on making the thread for この美術部には問題がある！ (This Art Club has a Problem!) on Monday.
If you want to have a say in the icon and the schedule @seanblue let me know, otherwise for the schedule I’ll look at the chapters, number of pages and number of chars, with a short ramp up period and make a proposition in the new thread. And the start date will depend on the schedule of 舞妓さんちのまかないさん (Kiyo in Kyoto).
8 Likes
I haven’t read it, so can’t give any informed suggestions on the schedule. I see a couple good painting related emoji, so I’m sure one of those will be fine.
4 Likes
Ah, thanks for pointing that out. I assumed it was slightly more nuanced than that.
At the end of the day if no one has read the book, any estimate of difficulty is only going to be a guess. At least with Natively there’s a chance for lot of books that someone has read it and provided their feedback.
7 Likes
Another thing regarding local difficulty polls is that people generally get better at reading over time if they keep doing it. Because people rarely update their rating on older nominations (sometimes from years back!), the difficulty between nominations aren’t necessarily consistent even between readers.
I’ve seen that happen in the Advanced book club when I decided to go through the samples of older nominations. After I updated my rating, it looked like my reading skill was higher than other regulars just from looking at the poll while that is most likely not the case. We’re all getting better, but the older polls doesn’t reflect that.
Maybe that phenomenon holds true for this book club as well.
On another side, I feel like local difficulty polls can be a way to estimate whether new nominations are too easy or difficult for the intended level of the book club. But that can probably be evaluated with the help of Natively too, I guess.
11 Likes
My one concern is that natively’s rating scale doesn’t seem quite comparable between books and manga, I think because many people put off reading books until after they’ve tried some easier manga. i.e. a feel a book with natively level X feels harder than a manga with natively level X, if even just on the level of time commitment, whereas the poll in the clubs here are people comparing them all amongst the group of items up for selection, rather than other media of the same type.
The home thread for 舞妓さんちのまかないさん/Kiyo in Kyoto is up. I’ve set up a discussion for the schedule for now, so feel free to share your thoughts.
17 Likes