Although I don’t like much the concept of “knowing a Kanji” as other users already mentioned, the results were quite accurate, indeed.
I answered about 90 questions, but after a while it was just hovering around 2000, what is pretty accurate, considering I took 漢検2級 but definitely forgot a bunch of them already (while knowing other non-2kyu kanji from everyday life)
I took the quiz a second time, still being pretty strict.
I thought the range was supposed to get narrower the longer you went. Seems like my kanji knowledge is so all over the place that my range didn’t really change even after answering 90 more questions. 
There is still something weird about what this website considers easy. I got 彦 which had a number of ~1100 next to it, implying it’s not hard. But it’s an N1 kanji taught in 9th grade, not taught on WaniKani, and I have never once encountered the kanji or any word written with that kanji (even in kana) in any book I have read. Looks like it might be used in a bunch of names, but still. Getting it wrong knocked off like 30 points.
Also, case in point about not recognizing kanji in isolation: I saw 屑 and was like “I feel like I know this, but I can’t think of the reading or meaning or a word it’s in, so I guess I don’t know it”. As soon as I flipped the card and saw the meaning it hit me… I am literally reading a manga called 星屑テレパス right now. How embarrassing!
Very interesting. I actually know 彦 because of Danganronpa, the one character’s name is 冬彦 I don’t think frequency implies simplicity either. 藤 is a very common kanji, but it’s mainly used in names, so it’s probably not one of the first 500 kanji one typically learns.
I could give more assumptions, but again, just make of whatever you feel like of the test.
Yeah, if you look at it on Jisho it’s listed as 1117th most used in newspapers, presumably because of names. A brief search of NHK seems to show that as well. Perhaps they are even using that specific ranking.
So, yeah, it’s dumb of them to just use that raw ranking and knock people down for not knowing it, but it seems like a pretty simple website.
If they’re just using the newspaper ranking with no other thought, then it can definitely be improved. If only because newspapers are just one, very specific, collection of kanji usage.
EDIT:
Also, it’s not taught in school at all. It’s a jinmeiyou kanji. So, not 9th grade. And I doubt that the classification of N1 is meaningful either. We’ve talked about the issues with N1 kanji lists many times, but I really doubt you would ever see this kanji in a non-name sense on the JLPT. And even then, it would just have furigana or something.
The website certainly has its flaws, but it was still interesting to do. Thanks for sharing!
I’m not a native speaker, but it says I know 2883.
It said I know 1745 after 50 attempts, but 2 things
I’m not good with kanji in isolation and I didn’t really know how to score “knowing” a kanji
recently I’ve been using a more hand-written font for all text so seeing kanji zoomed in like this threw me off
On the up side, I am supposedly ready for N2 kanji wise .
I’m not too good with kanji in isolation either. I was a bit more lenient this time, and it said I knew 1100 instead of 750 in the OP. I also know at least a hundred hyogaiji, and since the algorithm mainly works by frequency, I doubt any of those characters are taken into account.
Yeah, this test has a lot of that Anki smell, where you grade yourself on each kanji. This is why I hate Anki and use WK instead, because on WK there is no self-grading. You either give a correct answer or you give a wrong answer.
Eh, I just reviewed 男 and got it wrong for answering male instead of man. I clearly know the kanji. Anki and wanikani each have their respective uses.
No one is claiming that WK is perfect. In this case, they should add “male” as synonym and then the problem is solved. Anki will forever suck balls.
Because kanji meanings and translations of words are known for being black or white when it comes to answers!
Black meanings matter
I actually zoomed out when taking the quiz because the default size made it hard to “see” the kanji for some reason.
hmm interesting
first guess was something about 1400
then it went down to I think about 6-700
after 30 kanji it’s at 1060 which is I think about 20 off of what WaniKani suggests so…
I gotta say it’s quite impressive! 
Did about 50 tries.
Did the test again with 50 tries and scored even lower (around 1500 kanji). I noticed this time around there was 1-2 kanji that I didn’t know so well and which somehow kept the overall score low. I guess I might know “only” that many kanji, but without proper word context it’s hard to guess some of them for me 
.
EDIT:
Some more testing with a 100 sample. Looks like a good or bad answer very early on has a massive impact on the end result. I knew some slightly rarer kanji and that bumped my score to over 2300 to slowly drop to around 1950 at the very end. I think the 1700+ I got on my first try might be closer to how things are.
That strangely made me realise that if I don’t see it in a word within a sentence I don’t really recognize a Kanji.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
