The best way to combat this is to add more sentences using the vocabulary that you keep failing.  It really makes them stick for both long and short term.

I usually get tripped up on the kunyomi readings with 3-4 kana syllables, but if it's onyomi then referring to rtk2 is a sure shot.

Hehe.. I think I'm using the same method. And I also use the Sai of Hikaru no Go... Of the kanji that I keep forgetting the readings of after a few tries, I just write down the reading, and then the Heisig keyword. As I add more vocabulary to Anki (and meet more kanji+readings), the consequence is that for many readings, already two or more keywords have been attached. For example, with reading ちゅう, I have now keyword 'pour' (#267) and 'mid-air' (#1109). (The main storyline is the Chewie from Starwars).

I'm not sure, but I think this way of doing it is similar to the kanji chain method, only here the story 'evolves' as more keywords get added for the reading. On the opposite, in the kanji chain method -as I understand it- you create one story for a single reading and add all relevant keywords (which can be a huge amount)...

Lots. And lately I started watching everything without subs. Major difference. I dont understand a lot, but this is not the point. It kinda gets some hits combo with the sentences.

Last edited by mentat_kgs (2008 September 14, 4:00 pm)

If you're doing KO, then there is a lot of repetition later on, so just keep failing words, and eventually you will see them in other sentences.  If you're collecting them randomly, then you should add more context or words with the same readings for each of the readings you're having problems with.  Also, I've begun to notice some phonetic markers in kanji, they aren't very consistent, but when I see them they get noticed.  For example, kanji with the 市 radical in them often have a reading of せい, and ones with the  父 particle in them often use こう.  They:re not very useful as a complete system, but kind of assist in guessing a reading sometimes.  The more readings you know, the more likely you can guess part of a word, which makes it a little easier.

I think the making out in japanese stuff is quite idiomatic, so the words don't actually match the real meaning in many cases.  It's just rote learning and there's not much to be done about it I think.