I am at Kanji 98 [lesson 5]
Haven't skipped a day of reviews although I took yesterday off from doing a new chapter.
First three chapters were actually pretty easy. I am finding things more difficult in 4 and even more so in 5.

I finished the learning phase for 5 in reasonably good shape it it is getting harder and hard so I'd though I'd ask some questions. More and more don't feel set very solid compared to 4 and 4 compared to the first 3.

How much of a failure rate is acceptable. I am starting to fail cards at the review stage. Will setting failed cards to the 10 minute stage and reviewing via ANKI eventually take care of it?
Also I find myself having to have a serious think to remember the story.
I am having trouble at times with character order. The craft of getting clams is Tribute, wait the I beam goes on top?
Craft a page using paragraphs, goes on side. Also more words seem to be coming up that are hard to keep apart. I vs Ourself, for example.

I guess as more and more Kanji come up, I fear that both remembering all those stories and keeping them straight and translating them to how to write characters will eventually be undoable.
Second languages and myself don't have the best of relationships and I am in my mid 40's.

I am running my deck with a starting ease of 130 which as I understand it is the lowest and thus the best setting for more difficult material. I run learning at 1 2 3 4 5 6 7 8. I am thinking of adding maybe another 4 reviews for 12 hours ahead, 720 721 722 723 which I assume will in 12 hours let me run though 4 more times with a card plus failures before getting into review stage.

I set max interval time to 30 days although it hasn’t come into play yet as I got other plans between finishing the deck and starting serious Japanese study and I figure 70 reviews a day won't take long and 30 day intervals will keep things in the skull.

Any other settings I should consider changing.

Dudeist Wrote:I am at Kanji 98 [lesson 5]
Any other settings I should consider changing.

No, and you actually should consider undoing most of those changes again.
The idea of Anki (and spaced repetition in general) is that it minimizes the amount of reviews, by only making you review things when it thinks you're about to forget them. That allows you to learn more material while still retaining *most* of it.

Failing reviews is normal. If you fail a card, Anki makes you learn that card (and only that one) again and lowers its ease so you'll see it more often in the future. If you don't fail the card, you don't need more reviews because you remembered it. Lowering the ease of all cards to the minimum means you'll waste a lot of time reviewing cards you know well, when you could review cards you don't know well (or learn more new cards) instead.

The learning intervals also don't work how you think they work. Each step is the time it'll wait after a successful review. So those learning intervals mean you would need a minimum of
1+2+3+4+5+6+7+8+720+721+722+723 minutes (= ~49 Hours) for each card to move from learning mode to review mode. If you fail a card at any point in the learning process, it goes back to the start, so you'll need another 49 hours. Seeing the same card 20+ times in a few days will just make you burn out and hate Anki.

Also, your age has nothing to do with this. If "your memory is bad" that just means that Anki will make you review the things you're forgetting more often. Forcing reviews of the things you ARE able to remember instead won't do you any good.

Tamba wrote a pretty good explanation for why your anki settings are probably sub-optimal. I'll do my best to suggest some better settings.

First off, depending on your goals, you could reverse your cards and do kanji recognition which is MUCH easier and faster. Of course you won't get any writing practice, but if you don't paln on doing a lot of writing immediately, that might not be much of an immediate problem. It is my suspicion that it is much easier to learn to write kanji if you are already very good at recognizing them. Either way you do is fine, I just wanted to make sure you are aware that there is another option.

As far as anki goes, I fee that it is appropriate to target a certain percentage accuracy and adjust anki until you reach your target accuracy. According to research from supermemo(anki uses the supermemo algorithm) an accuracy between 70-90% is the most efficient. For me personally, 70% ruins my day and makes me want to quit no matter how efficient it it, so I've been targeting 90% although it's not as efficient, it's more enjoyable to me. That's the trade off.

For learning steps, currently your settings add one minute each time you answer correctly until you get to 10 minutes, then you won't see the card again until the next day. This doesn't make sense to me because you are adding one minute, one minute, one minute....then something like 18 hours. It should be more even, like multiplying each interval by 3 to get the next interval. So something more like 1 3 9 18 54 162 486 1440 makes a lot more sense to me and is the same number of steps. Incidentally 1440 is the number of minutes in a day, so you could omit that interval if you want.

For starting ease, I would look at your anki stats and adjust so that your young and mature accuracy is within your target range.

I hope this helps

I prefer a higher success rate. It's a personal thing and I am in no rush to get through the deck.

I went back to 1 through 8 because I hit the decks once a day so there seems little point to setting study times 6 hours later unless I want to still be in learning stage which it seems from this thread I probably don't.

I like the 8 reviews, I write things out 5 times and mull over the Kanji by hand but often the stories are complete crap and even the deck stories which are better, sometimes it takes a few kicks at the can to come up with something better or notice flaws in a story I am using. Also soon there will be no RTK stories and I need the deck stories to either use or at least help me craft something similar that I can work with. I can't do it cold. I also find the recalling the Kanji from a key word a lot more productive than try just writing them out over and over.

I did ditch the next day learning intervals. I read a lot about how 1 and 10 just isn't enough. I also read that Anki is meant to memorize things you already know and as I mentioned above, I often am lost and despondent after RTK.

I was considering doing them both ways but it is recommended by buddy to do keyword to Kanji so meh I will obey Dalek style and exterminate those Kanji. It takes a bit of time but when I do see them in the real world I seem to be able to pull out the ones I've done. Not that I am reading native material but every so often I come across them and when I do I see what I can spot. He predicts that recognition will take care of itself and it seems to be... slowly as I am often story dependent but still.

I find writing it much slower, sure, but it really packs it in the brain. Again I got the time to really solidify it.

I have reset all the settings back to default except for that 1 to 8, I put that 130 setting to 200 which as I understand it doubles the interval for the good setting which I am happy with and seems to be the standard progression as opposed to 250.
I've kept the 30 day max interval because there may be a bit of time between ending the RTK and starting other Japanese stuff and I want it fresh and what's 15 or 20 minutes a day. I hope to get started on French pretty soon and All French and no play make homer something something. I figure Kanji vs full language study won't run interference with each other as it is completely different. Also it will give plenty of post lesson time to really hammer things in the skull before I start on the texts/vocab decks and whatever.
I've also set leech to 99 times and note not suspend. I don't seem to be alone in this. I understand the arguments on both sides.

Dudeist Wrote:I was considering doing them both ways but it is recommended by buddy to do keyword to Kanji so meh I will obey Dalek style and exterminate those Kanji. It takes a bit of time but when I do see them in the real world I seem to be able to pull out the ones I've done. Not that I am reading native material but every so often I come across them and when I do I see what I can spot. He predicts that recognition will take care of itself and it seems to be... slowly as I am often story dependent but still.

Just to be clear, I wasn't suggesting to do both ways, production only or recognition only, not both. Reason being that you can get through them quicker doing recognition and I believe it may be more efficient in the long run.

Everything else makes sense except maybe setting your starting ease back to 200. You'll know if that's working out or not depending on how your young and mature accuracy reacts after a month or so.

yogert909 Wrote:Just to be clear, I wasn't suggesting to do both ways, production only or recognition only, not both. Reason being that you can get through them quicker doing recognition and I believe it may be more efficient in the long run.

Everything else makes sense except maybe setting your starting ease back to 200. You'll know if that's working out or not depending on how your young and mature accuracy reacts after a month or so.

I understood your point regarding time saving. I was just mentioning where my head was at regarding that topic.

If I understand properly how starting ease works, I don't think 200 vs 250 will make that much of a difference especially compared to 130 to 250 or even 130 to 200. As I mentioned above, I have the time and am willing to sacrifice it for accuracy. I really hate failing cards on multiple levels which I won't bother explaining. Maybe when I get started on Vocab decks I'll tolerate a drop in accuracy to get more words in, 85% of 1000 is better than 95% of 800 after all and there are a lot of vocab to learn [18K for N1, even more if you want to breeze though material]
With RTK there are only 2200 in RTK 6th edition and they are building block material and I suspect it will be a while before I will move on to the next stage.

I am very glad I asked, 130% would have been a mistake, I see that now.

Just to clarify again, a lower starting ease number correlates to higher accuracy. It is the percentage that your current interval is multiplied after you press good. So you get more reviews with a lower number. From reading what you wrote it seems you may be thinking a higher ease number means easier, but it's the opposite.

yogert909 Wrote:Just to clarify again, a lower starting ease number correlates to higher accuracy. It is the percentage that your current interval is multiplied after you press good. So you get more reviews with a lower number. From reading what you wrote it seems you may be thinking a higher ease number means easier, but it's the opposite.

Thanks for your assistance.

I understood that lower means more and thus easier. As I understand it, 200% under good would mean a 10 day card gets reviewed in 20 days, 250% in 25 days and 130% in 13 days.
A 130% would mean a glacial move up in interval lengths so I am glad I changed it to 200. 250 seems a bit too high though.

Thanks to this thread I am trying to trust Anki and for that matter myself more. Let the hooks to their word like in that Simpsons episode when Flanders took over the world and everyone in reNedification camp had hooks in their mouths [there is a Kanji there I hope] and were told to just let the hooks do their job in creating fake smiles.

As I understand it, if I fail a card it restarts the interval count process anyways so I shouldn't worry too much about accuracy. If something causes me problems I'll see it a lot more.

Ah, ok. You got it.