You can also adopt a "If in doubt, pass it (as hard)" policy and just let the SRS do the work: if you didn't know it well, the next time you will have forgotten it. If you know it the next time, it would have been pointless to fail it.
I have had the same issues aswell, but I have come to a point where I'm not being too hard on myself when it comes to rating. I generally overrate instead of underrate, because then I can spend more time on the really hard cards.

Trust in the SRS.

Edit: And maybe you should spend some more time really learning cards.

I've tried passing them, but I usually get the same results the second time around. I dare say in the cases of like 崩れ I somehow "learned" the incorrect reading even though each time I sit there and repeat the pronunciation 10 times out loud in an attempt to beat the incorrectness out.

I understand that SRS works, but it mostly works on the forget-curve. For the rest its left up to the user and I wish sometimes there were references/tips on their site (Anki in this case) with suggestions on how to grade. Besides "Again," Hard/Good/Easy/Very Easy are completely relative adjectives. I can sort of understand that overtime as you grade things will straighten out, but before "over time" rolls around, you might be creating more work for yourself. ie: If you overrate and mark more stuff as 4's when they should be 2s. When those cards come back around you'll have to fail them, but had you good tips to go on for grading, it could have been avoided.

I don't quite follow what you mean by "spend time really learning card." It's not as though I'm going through the dictionary and adding words at random and then just hoping SRS does the rest. I pull them from what I read.

Thanks

This thread was meant to focus on how the brain processes kanji rather than the neuroscience of SRSing or passing + failing flashcards. There hasn't been nearly enough stuff on that, though admittedly I have posted a few links in this thread discussing spaced retrieval, mostly because I was growing tired of arguing that it was possible to apply SRSing to other types of information/memory besides vocabulary and the like, so those links focus on the effectiveness of spaced learning for procedural/muscle memory, etc.

However, you might find these useful: http://forum.koohii.com/showthread.php?p...7#pid38977 (various supermemo links, that site is a wealth of not unquestionable authority, but of many thoughts from people who have made these ideas their bread & butter)

Phauna posted some links in the past that we've all taken to heart in various ways:

http://www.sciencedaily.com/releases/200...110834.htm - On relearning being easier than learning...

http://www.sciencentral.com/articles/vie...=218393117 - If you can't figure something out, don't dwell on it, move on...

My own focus is ease of relearning, moving on quickly, i + x rather than i + 1, and 'failing in layers' (see previous forum posts using this keyword or 'heterophenomenological life partner', as well as my warping of what Supermemo calls 'planned redundancy').

Maybe in the future we'll see more rigorous studies of spaced repetition systems short and long term, but I'm more confident in our figuring out notions of 'retrievability' and 'stability' as Supermemo calls them and how the spacing effect works, rather than scientifically proving how to use the flashcards best and what algorithm matches everyone's brains one size fits all...

vix86 Wrote:My problem (and I may have made a post on this), is I'm never sure whether to fail a card where I come close to the proper pronunciation of a card. Ex: 登場人物-とうじょうじんぶつ I always think its とうじょう＿にん＿ぶつ. Or I get 'close' like 崩れーくずれ　I'll think "たずれ.”

Take a note from Heisig and focus on one thing at a time.

If you want to work on both meaning and pronunciation, have your SRS produce separate cards for meaning and pronunciation. If you get the meaning wrong on a meaning card, fail it. If you get the reading wrong on a reading card, fail it. You shouldn't be passing 崩れ when you read it as たずれ. That's preventing the SRS from doing its job.

vix86 Wrote:I don't quite follow what you mean by "spend time really learning card." It's not as though I'm going through the dictionary and adding words at random and then just hoping SRS does the rest. I pull them from what I read.

I meant that when you see a word for the first time, you spend some time to learn it: look at the kanji, look at other readings you learned for the kanji, look at the compound and if the meaning seems logical etc. But perhaps you already doing that.

Ow and another thing: I adjust my grading on my passing percentages. I don't know if you got the Kanji graph plugin, but it comes with graphs with passing percentages for first-seen, young and mature cards. I for instance had one deck, where I had around 95% passage for young and 98% for mature cards. I think this is too high, it means that I'm doing unnessecary work. So I decided to grade more 4's when before I would have graded 3. So far this has worked well. You could also do this the other way around (i.e. grade yourself lower when your passing percentages are low).

(Either way, it saves time to grade yourself quickly and not to think too long about what you grade)

@kame3: Ah, ya, I already do that with my words. I'll even redundantly create cards if its a compound (ie: 廃棄処分 gets 廃棄＋処分＋廃棄処分 cards). I'm not currently using that plugin but now that you mention it I'll check it out.

@JimmySeal: I'm not sure how I would go about making seperate cards like that. Unless you mean something like Recognition and Production cards. Of which, I've never liked production cards (ie: English on front, Japanese on back) and eliminated them from my decks some time ago.

@nest0r: I never thought to check out Memosyne, but now that you mention it I'll check out the links. As for the relearning stuff; I've known about that for awhile, I actually saw that SD article when it came out, interesting stuff. There was a later press release from another Uni that made me think of this article again; when they videoed and showed that nerve synapses "float in the wind" and constantly change (something we knew already, but seeing it on video was still pretty cool). I felt it really drove home the importance of long term potentiation at the synapse.

Thanks for the input, all!

I didn't recommend Mnemosyne or Supermemo, just some of the many ideas floating around the Supermemo site. ;p

vix86 Wrote:@JimmySeal: I'm not sure how I would go about making seperate cards like that. Unless you mean something like Recognition and Production cards. Of which, I've never liked production cards (ie: English on front, Japanese on back) and eliminated them from my decks some time ago.

Are you using Anki? If so, you should be able to set up a card model that has three data fields: Word, Meaning, Reading, and then create two separate card templates for Word -> Meaning and Word -> Reading.

JimmySeal Wrote:Are you using Anki? If so, you should be able to set up a card model that has three data fields: Word, Meaning, Reading, and then create two separate card templates for Word -> Meaning and Word -> Reading.

Yes, but then how would I be able to know which is which then? Unless I make a clear note on the front. Then I would have to only do Meaning and Reading seperate. I would have to suspend the Meaning card if I want to learn Reading and then unsuspend after its learned. Otherwise it makes no sense to have two of the same card front, and once I reach that point it'd make more sense to merge the two cards really. Not saying its bad idea or anything, it just seems like my current system of Kanii & hidden example sentence on front, and meaning and furigana/kanji on back is sort of the same.

Thanks

vix86 Wrote:

JimmySeal Wrote:Are you using Anki? If so, you should be able to set up a card model that has three data fields: Word, Meaning, Reading, and then create two separate card templates for Word -> Meaning and Word -> Reading.

Yes, but then how would I be able to know which is which then? Unless I make a clear note on the front. Then I would have to only do Meaning and Reading seperate. I would have to suspend the Meaning card if I want to learn Reading and then unsuspend after its learned. Otherwise it makes no sense to have two of the same card front, and once I reach that point it'd make more sense to merge the two cards really. Not saying its bad idea or anything, it just seems like my current system of Kanii & hidden example sentence on front, and meaning and furigana/kanji on back is sort of the same.

Thanks

I think it's best to do everything at once, myself, as long as the relationships, depending on your goals per repetition, are complementary. ;p Thousands of cards and a couple of years later, I feel it more strongly than ever as I refine my methods.

vix86 Wrote:Yes, but then how would I be able to know which is which then? Unless I make a clear note on the front.

.....then make a clear note on the front.

Quote:Then I would have to only do Meaning and Reading seperate.

Yes.

Quote:I would have to suspend the Meaning card if I want to learn Reading and then unsuspend after its learned.

I don't really think that's necessary. Over time, the cards would get spaced apart enough that they wouldn't interfere with each other.


Anyway, take it or leave it. All I'll say is that it makes no sense for you to mark cards as passed when you remembered the pronunciation completely wrong, and then come on here asking us why you can't remember their pronunciation.

JimmySeal Wrote:All I'll say is that it makes no sense for you to mark cards as passed when you remembered the pronunciation completely wrong, and then come on here asking us why you can't remember their pronunciation.

In the past I had heard some people suggest that if you hit close to home "pass it." But then theres the question of "how far do you pass it?" I was more interested in whether there was any suggestion from the literature for how to handle SRS and the different levels of "passing."