yogert909 Wrote:I'm not sure what you mean by not realizing you are done with a deck. Theoretically you are never done with a deck because reviews will keep coming forever.

Say I'm doing a deck with 1000 cards and at some point I've reviewed every card, naturally I'd want to move on and do another deck when it ends like that. Is that the right way?

You'll know the day after you finish when it reports "New: 0". To see how many cards are remaining, you could temporarily set the option to a large number (the reporting caps at "1000+", though).

All,

What is the best way to revise grammar on Anki?

Vocab lists are obvious, grammar points... less so.

Thanks.

A lot of people here recommend Tae Kim's Grammar Guide for which there is an Anki deck. I just followed along with the lessons and then tested myself with anki. And I didn't bother with the cloze deletion part - just japanese sentence on the front, english on the back. It's also helpful to put the tag field on the back so that it's makes it easier to find the section in the grammar guide.

I have about 30 leeches in my RTK deck and I want to relearn them in their own deck. I tried to make a filtered deck with tag:leech and isuspended but it was saying it found no card matches. The browser, however, is able to filter them by both.

How do I make such a deck? Maybe there's another way I can go about this?

You can't review suspended cards even if you select them. You have to manually "unsuspend" them.

Codexus Wrote:You can't review suspended cards even if you select them. You have to manually "unsuspend" them.

Thank you. That did it. I tagged them all leech just in case, unsuspended them and then filtered by tag:leech.

I don't know if this question fits well this thread but... there is a way to calculate how many reviews I'll have in the future if I add 'x' cards and if I have a failure rate of 'y%' for my reviews (supposing I press always "good" for non-failed cards)? It's a rethorical question because obviously there is a formula to calculate this, so math guys don't be shy :3

I'm not particularly good at math, but the figure I've heard thrown around is expect 10*[new cards per day] based on an expected fail rate of roughly 10%.

I'm consindering suspending all my production cards in my Core6K to speed up my progress. But if it doesn't suit me I would like to return to reviewing prodcution card (i.e. unsuspending the cards) without being drenched in review. If I suspend a card that is to appear in say 10 days and then unsuspend it after 30 days will it then be due for review the day after I have unsuspended it?

RandomQuotes Wrote:I'm not particularly good at math, but the figure I've heard thrown around is expect 10*[new cards per day] based on an expected fail rate of roughly 10%.

So if I add 100 new cards (only a purely theoretical number) I'll eventually reach a maximum of 1000 reviews? Or if I keep adding the same number of cards every day, at some point I'll pass that number? Or, in other words, if one keep adding the same number of cards every day, he will eventually reach a ceiling in the number of reviews? Or does they'll continue to increase endlessy?

cophnia61 Wrote:I don't know if this question fits well this thread but... there is a way to calculate how many reviews I'll have in the future if I add 'x' cards and if I have a failure rate of 'y%' for my reviews (supposing I press always "good" for non-failed cards)? It's a rethorical question because obviously there is a formula to calculate this, so math guys don't be shy :3

Depends on other variables as well, including how the failures are distributed (do more fails occur on the first review, or on the third/fourth/fifth/etc. reviews?), when the cards are added, etc.

And it's a pretty complex formula. I'm sure someone who just finished high school and was good at math could figure it out, but it would take a bit of work. I bet the word "logarithmic" would come up.

cophnia61 Wrote:So if I add 100 new cards (only a purely theoretical number) I'll eventually reach a maximum of 1000 reviews? Or if I keep adding the same number of cards every day, at some point I'll pass that number? Or, in other words, if one keep adding the same number of cards every day, he will eventually reach a ceiling in the number of reviews? Or does they'll continue to increase endlessy?

As long as you keep adding 10 every day, the number of daily reviews will of course keep going up slowly. There is no ceiling. If you add 10 each day for an infinite number of days, you will have an infinite number of daily reviews.

But it's a logarithmic growth, like this: http://en.wikipedia.org/wiki/Logarithm , not a linear or exponential growth. In other words, it's nothing to be scared of, the growth will be tiny after a while. It would take a very long time (way longer than a year, or even 10 years) for you to reach 1000 reviews/day, just by adding 10 each day. But you would reach it eventually. You would also reach 1 million daily reviews eventually, just by adding 10 a day. The Earth would be long gone by then, but you would reach it.

Hinsudesu Wrote:I'm consindering suspending all my production cards in my Core6K to speed up my progress. But if it doesn't suit me I would like to return to reviewing prodcution card (i.e. unsuspending the cards) without being drenched in review. If I suspend a card that is to appear in say 10 days and then unsuspend it after 30 days will it then be due for review the day after I have unsuspended it?

I'm not entirely sure how it works but I don't think suspending a card affects it's due date, so if you pass a cards due date while it is suspended and then unsuspend it you will be asked to review it immediately. You can always just unsuspend them a few at a time though.

Splatted Wrote:

Hinsudesu Wrote:I'm consindering suspending all my production cards in my Core6K to speed up my progress. But if it doesn't suit me I would like to return to reviewing prodcution card (i.e. unsuspending the cards) without being drenched in review. If I suspend a card that is to appear in say 10 days and then unsuspend it after 30 days will it then be due for review the day after I have unsuspended it?

I'm not entirely sure how it works but I don't think suspending a card affects it's due date, so if you pass a cards due date while it is suspended and then unsuspend it you will be asked to review it immediately. You can always just unsuspend them a few at a time though.

Thank you. It appeared to me that it would be quite easy to just test out how unsuspending works with a recent card.... I'll tell how it works in a couple of days

Hinsudesu Wrote:I'm consindering suspending all my production cards in my Core6K to speed up my progress. But if it doesn't suit me I would like to return to reviewing prodcution card (i.e. unsuspending the cards) without being drenched in review. If I suspend a card that is to appear in say 10 days and then unsuspend it after 30 days will it then be due for review the day after I have unsuspended it?

One good thing about going past the review date is that anki uses your actual review date as the basis for it's rescheduling. So for sake of illustration let's say you have a card with an interval of 1 week, but you don't review it for a month. If you pass the card, it's new interval will be like 2.5 months. This would help a bit with reviews should you decide to start them back up.

cophnia61 Wrote:

RandomQuotes Wrote:I'm not particularly good at math, but the figure I've heard thrown around is expect 10*[new cards per day] based on an expected fail rate of roughly 10%.

So if I add 100 new cards (only a purely theoretical number) I'll eventually reach a maximum of 1000 reviews? Or if I keep adding the same number of cards every day, at some point I'll pass that number? Or, in other words, if one keep adding the same number of cards every day, he will eventually reach a ceiling in the number of reviews? Or does they'll continue to increase endlessy?

Stansfield is correct, but at timescales closer what most people care about, I believe the rate is more like 20x according to supermemo's theory page. So in your example of adding 100/day after a year or so, you'll get up around 2,000 reviews/day instead of 1,000.

supermemo Wrote:In a long-term process, for the forgetting index equal to 10%, and for a fixed daily working time, the average time spent on memorizing new items is only 5% of the total time spent on repetitions. This value is almost independent of the size of the learning material

Hey guys I'm new here, just have a question about the Anki App on my iphone.

(I've searched high and low for info, can't get ahold on anything).

I've been using it on my phone for few months now (been great). But just recently daylight savings hit my town.... And it's f****ed the reschedule time (everything would reset at 2pm...which was perfect). Now resets at 5pm

I can't find any options to change it on the app. I know you can do it on PC... But I haven't found any information on how to do it on your mobile.

Any help or links would be great.

Thanks

http://anki.tenderapp.com/discussions/an...day-starts

Damien Elmes Wrote:We'll look into adding a setting directly in AnkiMobile in a future release.

Stansfield123 Wrote:

cophnia61 Wrote:I don't know if this question fits well this thread but... there is a way to calculate how many reviews I'll have in the future if I add 'x' cards and if I have a failure rate of 'y%' for my reviews (supposing I press always "good" for non-failed cards)? It's a rethorical question because obviously there is a formula to calculate this, so math guys don't be shy :3

Depends on other variables as well, including how the failures are distributed (do more fails occur on the first review, or on the third/fourth/fifth/etc. reviews?), when the cards are added, etc.

And it's a pretty complex formula. I'm sure someone who just finished high school and was good at math could figure it out, but it would take a bit of work. I bet the word "logarithmic" would come up.

I think this is a little more complicated than that because the answer will need to model the randomness of the failures. I don't have an answer. I coded up a little simulator to visualize possible sequences of review intervals for a single card, given a 90% success rate and your other assumptions ("Good" for those 90% success): http://fasiha.github.io/anki-random-simulator/ click on "Rerun" to see various possible sequences of reviews over five years.

(The source code can be used as a tutorial on how Anki calculates reviews, i.e., factors, modifiers, etc. Something new I learned was that the factor doesn't change if you rate a card "Good", only again (-200), hard (-150) or easy (+150), with a minimum of 1300. See https://github.com/fasiha/anki-random-si...nkirand.js)

As you add X cards a day, soon you'll get many reviews on a single day. How to model this probabilistically isn't clear to me right now: the underlying model simulated above is in terms of reviews, not days. How to convert the model to a day-indexed one, to answer your original question? No idea yet.

You could certainly run Monte Carlo simulations (possibly by enhancing this simulator) to see on average how many cards turn up per day, and answer your question numerically. But before doing that, you'd want to think about whether your failure model is good: the simulator above assumes any given review has a 10% chance of failure. So after 13 reviews (however long that takes---this is my average # of reviews), there's about 25% chance of having no failures (binomial(13, 13, 0.9) = 0.254187). But I have never failed a review for about 55% of my cards (see the second plot at http://forum.koohii.com/showthread.php?p...#pid211869, and a much more hardcore Anki user reported having 70% "perfect" cards). This isn't a perfect comparison but it makes me think that the 10% failure rate is concentrated in "hard" cards, and that that reviews aren't independent, identically-distributed 9:1 coin flips.

So even an answer from a Monte Carlo simulation might not match your experience well unless you spent some time thinking about a better failure model...

Corner a Real Statistician and ask them

cophnia61 Wrote:I don't know if this question fits well this thread but... there is a way to calculate how many reviews I'll have in the future if I add 'x' cards and if I have a failure rate of 'y%' for my reviews (supposing I press always "good" for non-failed cards)? It's a rethorical question because obviously there is a formula to calculate this, so math guys don't be shy :3

This is a maddeningly difficult problem with many dependent variables to consider. I've worked on it from a few different angles and I'm not able to come up with anything close to my actual results. I even took a set of real results from here and tried to curve-fit a model but the variables I used didn't seem to capture everything going on. The Supermemo theory page has the following equation, but even that doesn't seem to match my stats very well.

Supermemo Wrote:NewItems=aar*(3*e^- 0.3*year+1)+1)
where:
NewItems - items memorized in consecutive years when working one minute per day,
year - ordinal number of the year,
aar - asymptotic acquisition rate, i.e. the minimum learning rate reached after many years of repetitions (usually about 200 items/year/min)

If anybody comes up with something I'd love to see it.

yogert909 Wrote:I've worked on it from a few different angles and I'm not able to come up with anything close to my actual results.

Could you elaborate on what you mean by 'actual results'? You mean you can't predict your reviews per day accurately using just "X cards added per day" and "Y% pass rate"? Since the source code is available, you can tell exactly why you got how many reviews in a given day, so I guess you're trying to find a generalized prediction using less data than your entire review history?

aldebrn Wrote:

yogert909 Wrote:I've worked on it from a few different angles and I'm not able to come up with anything close to my actual results.

Could you elaborate on what you mean by 'actual results'? You mean you can't predict your reviews per day accurately using just "X cards added per day" and "Y% pass rate"? Since the source code is available, you can tell exactly why you got how many reviews in a given day, so I guess you're trying to find a generalized prediction using less data than your entire review history?

That's correct. I was talking about a generalized formula to calculate an approximate workload based on failure rate, deck age, daily add rate,... It would also be useful to predict how a workload would decrease over time when you stop adding new cards.

I have <br> tags in a field that I added to the back of a card and it's not giving me line breaks. Is this standard behavior, or is there something I'm missing? Other html tags are working fine.

yogert909 Wrote:I have <br> tags in a field that I added to the back of a card and it's not giving me line breaks. Is this standard behavior, or is there something I'm missing? Other html tags are working fine.

If the text in your field (after you press Ctrl-Shift-X) looks like:
line 1<br />
line 2

then this should work - it's working for me.