Card models are hardly unique to the deck. Statistics are mixed, but that's only because calculation of statistics of subsets of cards isn't currently implemented.

Edit: also, what sort of functional usefulness is there to be gained from the statistics?

(Please don't read my posts as being too argumentative; I do believe there is a "correct" answer here, but also in the event that there are advantages to separate-decks that I'm missing I want to know them, so I'm likely to keep up the debate until I'm convinced one way or the other)

~J

I use several different decks. Firstly because that's how they came. Also it keeps them apart if I want to concentrate on a certain aspects. Also keeps stats separate. If you decide to stop a deck it's hard if they are mixed into one deck.

Is there something more to it than select-by-tag then suspend or delete?

~J

If you merge the decks, it's easy to export a subset using tags in order to look at statistics if you need them.

As to the purpose of stats, everyone has their reasons. For myself, some are more useful than others.

I think the Kanji stats are over-rated simply because you don't know the context of each Kanji being counted. If you use a RTK deck, you're getting 100% stats which is know is meaningless. Plus, we don't know what on-yomi and kun-yomi are getting covered, which holds higher value when you begin reading native material and come across new combination of Kanji vocabulary. Still, it's nice to see more and more kanji get crossed off as time goes on.

Granted, my main purpose for stats is to see how much I've done over the last xx months. It's nice to see I did 50 hours of studying and 6000 items reviewed in 30 days. If I slack, it's right there in my face. Plus, that little function was added recently.

The stats that track how much work you've done are not exported, because they are stored in a cached format to save large graph load times. So if you want to track work per category, you need to keep separate decks.

woodwojr Wrote:Like hknamida says, all together. I've got RtK, sentences, topology, abstract algebra, MIPS architecture, chemistry, and assorted other miscellany mixed together.

Can you elaborate a little bit on the math part? I mean, how do you take advantage of SRS to study topology and abstract algebra, for example? I have no idea how to use SRS for this.


Thanks in advance.

resolve Wrote:Note you should use a single deck for similar material, and a different deck for widely different material (eg, two different languages).

This is an issue I've been thinking about a lot recently. My primary concern is creating conditions for optimal learning.

Robert Bjork - the benefits of interleaving practice - https://youtu.be/l-1K61BalIA

Of course I separate all of them. I separate new cards from reviews too. Making life tougher for yourself is dumb and will destroy your motivation and discipline in the long run.

I have separate decks and sub decks. I don't only use Anki for Japanese, but for my academic courses too. It would be a huge mind F**k if I was simultaneously reviewing my Core 6k and my genetics material. @_@

(2016-03-23, 1:25 am)risu_ Wrote: Of course I separate all of them. I separate new cards from reviews too. Making life tougher for yourself is dumb and will destroy your motivation and discipline in the long run.

I must say that when I was extensively using filtered decks that pulled in all kinds of cards, the variety was refreshing. Since I don't use filtered decks so much these days, the monotony of 500 kanji reviews yesterday was a bit motivation destroying.

Of course I probably wouldn't mix Japanese cards with geography or scuba jargon, but kanji mixed with vocab and grammar, sure.

I mixed my grammar recognition cards with my vocab recognition cards as separate sub-decks, but the grammar production cards (made by someone on this forum) are separate, because they take much longer to review (and I haven't bothered to make recognition versions yet).

I don't know if this is a setting or not (haven't bothered to look), but my cards in sub-decks are shown in groups; vocab reviews first, then grammar reviews afterwords, for me; it doesn't actually mix them together like: vocab, grammar, vocab, kanji, grammar, anatomy, etc.

I personally made subdecks for my sub2srs decks, anime/manga vocab decks and academic material like so:

Code:

Sub2srs
-Rich Man Poor Woman
-BORDER

Anime/manga
-Shokugeki no Souma
-Shirokuma Cafe

Genetics (or any other academic deck)
-ch 1
-ch 2
-etc...

Premade decks that tend to have no relation to the ones I make myself such as RTK and core 6k are their own decks. It just seems cleaner that way to me.

RawrPk Wrote:It would be a huge mind F**k if I was simultaneously reviewing my Core 6k and my genetics material.

I agree. I'm still trying to figure out of this is what Robert Bjork is suggesting when he talks about 'interleaving'. Would the mental gymnastics that we'd have to go through to review a variety of material in one deck aid our long term retention? Looking forward to reading that study

My current system is extremely disorganized and composed mostly of premade decks that I make slight changes to (usually adding kanji for words I know the kanji of (if they are semi-common), changing misused kanji, or removing kanji I see no use for). The current count of decks is 13, with 13k mature cards (a large portion of them redundancies) and 2k learning. If I get a new deck I don't suspend things I already know, I just go through them and hit easy (with easy for initial cards temporarily set to 99 days).

I think that if you reviewed all decks at once (or combined them into one deck) you would make your reviews take longer than the time it takes to switch between them, because each new card will have a microsecond of you determining what kind of answer it wants.

For cards of the same format though (like kanji front, reading+meaning back), I think everything should be in one deck of course.

I also sort new vocabulary from different books and song lyrics into different subdecks (I'll admit this is kind of useless as I use the same format for all of them).

(2016-03-24, 2:29 am)synewave Wrote: I agree. I'm still trying to figure out of this is what Robert Bjork is suggesting when he talks about 'interleaving'. Would the mental gymnastics that we'd have to go through to review a variety of material in one deck aid our long term retention? Looking forward to reading that study

I got curious about this so I found a Wired article about interleaving study and here is wha the man says:

Robert Bjork Wrote:“People tend to try to learn in blocks,” Bjork said. “Mastering one thing before moving on to the next.”

Instead of doing that Bjork recommends interleaving. The strategy suggest that instead of spending an hour working on your tennis serve, you mix in a range of skills like backhands, volleys, overhead smashes, and footwork.

“This creates a sense of difficulty,” Bjork said. “And people tend not to notice the immediate effects of learning.”

Instead of making an appreciable leap forward with your serving ability after a session of focused practice, interleaving forces you to make nearly imperceptible steps forward with many skills. But over time, the sum of these small steps is much greater than the sum of the leaps you would have taken if you’d spent the same amount of time mastering each skill in its turn.
....
There’s one caveat: Make sure the mini skills you interleave are related in some higher-order way.

So it doesn't appear he is suggesting interleaving disparate material, but different aspects within the same subject. The article is well worth a read as if mentions a few other memory tricks.

[RawToast]

After keeping decks separate for a long time, I ended up mixing decks (and tagging the cards so I could go back).

I don't know if this is still a problem, but in Anki 2 when you attempted a full review it would go deck by deck. e.g. Deck A's cards, then Deck B's, and so on. I wanted my cards to be mixed together, rather than following the same pattern every day (grammar, then vocacb, etc.), so using the one deck was the only solution for me.

I don't use Anki anymore, so that may no longer be an issue.

Interleaving has always interested me. The only problem is that way, cards will be harder to customize and edit and all those cards will have a single options group.
It'd be good if Anki allowed interleaving behavior even just by the use of a master deck with its subdecks. Is there a way? Maybe I simply don't know of it?

Interleaving might make each card more effective, but costs the time spent thinking about what type each card is. The time loss is most serious for a simple vocab deck where each card takes a couple of seconds, so I definitely want to keep that one separate.

Also, reviewing writing cards or cards with audio separately can allow them to be done in different situations.

(2016-03-27, 10:11 am)polyt urn Wrote: Interleaving has always interested me. The only problem is that way, cards will be harder to customize and edit and all those cards will have a single options group.
It'd be good if Anki allowed interleaving behavior even just by the use of a master deck with its subdecks. Is there a way? Maybe I simply don't know of it?

There's no problem with customizing or editing the cards: putting different card templates in the same deck works fine.

However, I have had pairs of decks where the only reason I didn't combine them was because of wanting different options.

(2016-03-29, 9:02 am)HelenF Wrote: Interleaving might make each card more effective, but costs the time spent thinking about what type each card is. The time loss is most serious for a simple vocab deck where each card takes a couple of seconds, so I definitely want to keep that one separate.

Perhaps this is the case for some types of cards, but I have vocab mixed in with sentences and there is no problem deciding what type of card each is because it's not important - the answer is always the english equivalent. I have some other cards where it is important to differentiate card types, but it's still not much of a problem, as I've styled the cards differently, giving a hint as to what type of card it is.

The only thing I haven't mixed is rtk cards with vocabulary as I don't want to get confused between keywords or translations. I know they are kind of the same thing usually, but rtk keywords have a 1-1 relationship with kanji and I'd like to keep it that way in my mind until I've matured most of my rtk deck.