I can change the Card Selection System to Repetition-spaced, but I would also like to limit the total number of cards I view per test to 50 total cards (instead of the nearly 5000 HSK vocab). I seem to recall that I was able to do this in the past but now I can't figure it out. Help appreciated.
Flashcard test question
- Thread starter zang
- Start date
There's no way to limit the number in a specific test - doing that with SRS would mean forgetting a lot of words (since you wouldn't be reviewing them often enough to keep them in memory) - but you can use the "limit new cards" option in Card Selection to ensure that it only doles out a small number of new cards at a time rather than all 5000 at once.
If you've already turned that on and you're still getting too many new cards, try resetting all of your cards' statistics - Organize Cards / Edit / Batch / scroll down to the bottom and you'll see that reset option.
If you've already turned that on and you're still getting too many new cards, try resetting all of your cards' statistics - Organize Cards / Edit / Batch / scroll down to the bottom and you'll see that reset option.
Hi Mike, I think I have a related question.
I've been using Pleco for around 6 months and love it, particurly for looking up words quickly on the move. I have used quite a bit, and really tried to love, the flashcards system. I won't rehearse in detail the advantages of Pleco over other programs (integrated with dictionary unlike Anki, mobile unlike Wenlin), because you know how brilliant it is, but do want to ask your help on how to use it so that I can properly focus and not spend hours tinkering with settings and worrying I'm missing things.
Essentially I love the way Wenlin does its flashcards, but since it mainly works for characters and also is not mobile, I would like to use Pleco for words to complement this. I've been trying to force Pleco to behave as much like Wenlin as possible by tweaking the settings, but have also tried to impose some order by using categories.
I think what I like best about Wenlin is its simplicity. I've flirted with Anki because it shares this feature - the reason I want to be happy with Pleco is the ease of adding cards straight from the dictionary (I also love in Pleco being able to flick back to the dictionary to find example sentences when I come across a word I haven't remembered properly or just can't remember how to use).
A couple of specific things I like about Wenlin's method is that (i) each 'round' of testing gives you a limited number of new cards to learn then makes really sure you know them, but crucially still uses what seems to be a very successful spaced-repetition algorithm to ensure you occasionally get shown things you know better, (ii) each round tests a particular set of characters in every direction - recognizing it for meaning, remembering its pronunciation (including tone separately), and then reproducing it (ie here's the definition, write it before moving on).
To explain how I'm using Pleco flashcards at the moment, I now have a category for each chapter of the textbook I'm doing, categories for things I find useful in day to day life, for other books I'm reading, and so on. The result is that I have 3200 cards in dozens of categories, relating to old textbooks and vocab lists, as well as the current one (some cards are in more than one category too), and an ever-expanding 'Uncategorised' section. I then test myself in a variety of ways, and often find myself confused and worried I'm not being systematic. For example, I may wish to spend time properly learning the current day's vocabulary (because I'll need it tomorrow) so run a test just on that chapter's category. But then I also want to simulate the long-term spaced repetition approach I would get from Wenlin or Anki so I will select all categories and run through those for a while. But there I'm not allowed to limit the number of cards (I don't understand this point that spaced repetition doesn't work if you limit cards per test session). Other times I will worry that going through the 3000 will mean too many less useful ones coming up and overloading my brain (I accept that this may point to the need for some brutal editing!), and therefore do spaced-repetition rounds for all chapters of the current textbook only, or of my Uncategorised folder (where I tend to dump things that come in daily life) for example. But since I can't limit the bumber of cards in a session I have to break the sessions artifically - the number of 'due' cards then seems to go down but somewhat but it's still nearly 2000 (and I think goes up again each day?). This all feels a bit messy and unsatisfactory and I'm sure there's a better way!
I guess my general question is
- whether you can advise on how to get a more systematic process that will allow me to take advantage of the power of spaced repetition to ensure I get to review everything at the appropriate intervals, preferably via a much simpler workstream than I have now.
And the more specific, related questions are:
- how is the spaced-repetition mechanism affected by my confusing workstream with multiple categories?
- how should I use the 'sort cards randomly' and 'sort not-due cards randomly' settings under spaced repetition? I've read the user guide but still feel confused about whether this means that effectively spaced repetition isn't going on any more? Which is more like Anki / Wenlin? I had this off at first and it was showing me stuff in essentially the order I added them (so really old stuff first), but when I switch it on I get some very new stuff - which is a true spaced repetition process? And apologies if I've missed this in the manual but what does the date in the top right refer to - the last time that character came up, when it will next come up, or something else?
- what am I missing on 'due' vs 'not due' cards? Right now as I said I just have to break the session if I do an all-categories session. But likewise when I do smaller subsets I am constantly having to tell it to show me cards that aren't due (and my muddled brain can't work out which of the two options I am then given is more sensible - to reset the clock or not).
Also a final related question:
- The second thing I like about Wenlin: (ii) above is perhaps less relevant to my main concerns here, but it's related to the point on categories - I would love to see Pleco's system somehow allow me to test things in both ways rather than the 'set' being tested in each round if I limit the numbers being essentially emphemeral (that's one reason I need categories - so I can run them both ways to ensure I properly understand them). I suspect this might not be possible, but if I were able to get 'today's' menu' down to a manageable number, I'd love it to be kind of 'frozen' for the day so that I can test myself on that set of cards in several of the rich variety of ways Pleco provides (another big plus of Pleco). If I'm not mistaken what happens is that once I've got it, say, from character first, the core is amended and it goes to the back of the pack. Perhaps fixing this would need separate scores for the different 'directions' - in Anki you would just put two flashcards in, but I don't think Pleco will allow that because the 'direction' isn't in the card, it's in the test settings.
I did wonder if part of my problem here (specifically too many showing up when I ask for the full set) is that I recently transferred to a new device, but it does look like my scores have transferred with my cards - I have 1800 in the 0-100 range, 180 in 100-200, 320 in 200-300, ... , 214 in 3201-6400, 213 in 6401-12800, 53 in 12801-25600.... Perhaps there is a way of tweaking something so that when I do the full set roughly 50 or 100 a day come up? 'Limit new cards' is already set to 20 - I think this isn't the issue here. Should I be setting a filter to limit the number of cards below 100 that come up? Is the issue that that profile (a huge pile at <100, then a thin spread above) is unusual and reflects my odd usage to date?
Apologies for the long query but I hope you have a good idea of what my problem is now and will be able to advise - as a full-time student right now on a very intensive course to get to as high a level as possible by the end of the year to start a job in China, I am spending several hours a day bashing through flashcards, so ensuring I am using them as effectively as possible is absolutely crucial!
Huge thanks in advance for your thoughts.
Best wishes,
Sam
I've been using Pleco for around 6 months and love it, particurly for looking up words quickly on the move. I have used quite a bit, and really tried to love, the flashcards system. I won't rehearse in detail the advantages of Pleco over other programs (integrated with dictionary unlike Anki, mobile unlike Wenlin), because you know how brilliant it is, but do want to ask your help on how to use it so that I can properly focus and not spend hours tinkering with settings and worrying I'm missing things.
Essentially I love the way Wenlin does its flashcards, but since it mainly works for characters and also is not mobile, I would like to use Pleco for words to complement this. I've been trying to force Pleco to behave as much like Wenlin as possible by tweaking the settings, but have also tried to impose some order by using categories.
I think what I like best about Wenlin is its simplicity. I've flirted with Anki because it shares this feature - the reason I want to be happy with Pleco is the ease of adding cards straight from the dictionary (I also love in Pleco being able to flick back to the dictionary to find example sentences when I come across a word I haven't remembered properly or just can't remember how to use).
A couple of specific things I like about Wenlin's method is that (i) each 'round' of testing gives you a limited number of new cards to learn then makes really sure you know them, but crucially still uses what seems to be a very successful spaced-repetition algorithm to ensure you occasionally get shown things you know better, (ii) each round tests a particular set of characters in every direction - recognizing it for meaning, remembering its pronunciation (including tone separately), and then reproducing it (ie here's the definition, write it before moving on).
To explain how I'm using Pleco flashcards at the moment, I now have a category for each chapter of the textbook I'm doing, categories for things I find useful in day to day life, for other books I'm reading, and so on. The result is that I have 3200 cards in dozens of categories, relating to old textbooks and vocab lists, as well as the current one (some cards are in more than one category too), and an ever-expanding 'Uncategorised' section. I then test myself in a variety of ways, and often find myself confused and worried I'm not being systematic. For example, I may wish to spend time properly learning the current day's vocabulary (because I'll need it tomorrow) so run a test just on that chapter's category. But then I also want to simulate the long-term spaced repetition approach I would get from Wenlin or Anki so I will select all categories and run through those for a while. But there I'm not allowed to limit the number of cards (I don't understand this point that spaced repetition doesn't work if you limit cards per test session). Other times I will worry that going through the 3000 will mean too many less useful ones coming up and overloading my brain (I accept that this may point to the need for some brutal editing!), and therefore do spaced-repetition rounds for all chapters of the current textbook only, or of my Uncategorised folder (where I tend to dump things that come in daily life) for example. But since I can't limit the bumber of cards in a session I have to break the sessions artifically - the number of 'due' cards then seems to go down but somewhat but it's still nearly 2000 (and I think goes up again each day?). This all feels a bit messy and unsatisfactory and I'm sure there's a better way!
I guess my general question is
- whether you can advise on how to get a more systematic process that will allow me to take advantage of the power of spaced repetition to ensure I get to review everything at the appropriate intervals, preferably via a much simpler workstream than I have now.
And the more specific, related questions are:
- how is the spaced-repetition mechanism affected by my confusing workstream with multiple categories?
- how should I use the 'sort cards randomly' and 'sort not-due cards randomly' settings under spaced repetition? I've read the user guide but still feel confused about whether this means that effectively spaced repetition isn't going on any more? Which is more like Anki / Wenlin? I had this off at first and it was showing me stuff in essentially the order I added them (so really old stuff first), but when I switch it on I get some very new stuff - which is a true spaced repetition process? And apologies if I've missed this in the manual but what does the date in the top right refer to - the last time that character came up, when it will next come up, or something else?
- what am I missing on 'due' vs 'not due' cards? Right now as I said I just have to break the session if I do an all-categories session. But likewise when I do smaller subsets I am constantly having to tell it to show me cards that aren't due (and my muddled brain can't work out which of the two options I am then given is more sensible - to reset the clock or not).
Also a final related question:
- The second thing I like about Wenlin: (ii) above is perhaps less relevant to my main concerns here, but it's related to the point on categories - I would love to see Pleco's system somehow allow me to test things in both ways rather than the 'set' being tested in each round if I limit the numbers being essentially emphemeral (that's one reason I need categories - so I can run them both ways to ensure I properly understand them). I suspect this might not be possible, but if I were able to get 'today's' menu' down to a manageable number, I'd love it to be kind of 'frozen' for the day so that I can test myself on that set of cards in several of the rich variety of ways Pleco provides (another big plus of Pleco). If I'm not mistaken what happens is that once I've got it, say, from character first, the core is amended and it goes to the back of the pack. Perhaps fixing this would need separate scores for the different 'directions' - in Anki you would just put two flashcards in, but I don't think Pleco will allow that because the 'direction' isn't in the card, it's in the test settings.
I did wonder if part of my problem here (specifically too many showing up when I ask for the full set) is that I recently transferred to a new device, but it does look like my scores have transferred with my cards - I have 1800 in the 0-100 range, 180 in 100-200, 320 in 200-300, ... , 214 in 3201-6400, 213 in 6401-12800, 53 in 12801-25600.... Perhaps there is a way of tweaking something so that when I do the full set roughly 50 or 100 a day come up? 'Limit new cards' is already set to 20 - I think this isn't the issue here. Should I be setting a filter to limit the number of cards below 100 that come up? Is the issue that that profile (a huge pile at <100, then a thin spread above) is unusual and reflects my odd usage to date?
Apologies for the long query but I hope you have a good idea of what my problem is now and will be able to advise - as a full-time student right now on a very intensive course to get to as high a level as possible by the end of the year to start a job in China, I am spending several hours a day bashing through flashcards, so ensuring I am using them as effectively as possible is absolutely crucial!
Huge thanks in advance for your thoughts.
Best wishes,
Sam
Limiting new cards is something we actually do support, via the "limit new cards" option in Card Selection - you can specify a set # of new cards per day or a set # in total (so that it'll introduce new cards anytime some existing cards pass the threshold at which they're no longer 'new').
Testing the same card in a sequence of fields is something we also support, though it's a bit hidden - if you set "Subject selection" in Test Settings to "Score-based," you can have cards go through a progression of different configurations based on their score. This doesn't yet support switching between different test types, but you can practice writing in a self-graded test via the sketch box.
As far as SRS not working if you limit the number of cards: the whole idea of SRS is that you *aren't* going back and reviewing all of your cards randomly, you're reviewing the specific cards that the system thinks you need to review today. So if you don't review all of those cards then you're liable to forget them, and if we let someone simply set SRS to review no more than (say) 100 cards a day when they actually need to review around 200 to keep current with their vocabulary then they'll find themselves forgetting lots and lot of words because they won't be reviewing them often enough.
We are working on a new SRS system that will accommodate set numbers of cards per day, but it'll do it through prioritization - essentially, if you only want to review 100 cards in a day, it will try its best to keep as many cards in your memory as possible, but will remove some cards from active study / accept the fact that you're probably going to forget them in order to ensure that the remaining ones are reviewed often enough to stay in memory.
But for your specific scenario with the current system I think I'd suggest a multiple profile approach. Set up one profile for straight-up SRS - add only the categories you really care about, turn on "limit new cards" with a reasonable threshold, and do that profile first on any given day to make sure that you review the cards you need to and keep them in memory. Possibly with score-based subject selection so that you'll keep being tested on them in different ways.
Then, set up another profile for additional review of the cards that you've recently studied - you can use a 'card filter' to limit the test to only cards last reviewed in the last few days. I'd suggest that you turn scoring off (Scoring System = None in scoring) so that these reviews won't affect the cards' SRS scores. Card Selection system here should be Random rather than SRS, and Limit New Cards off.
Finally, set up a third profile for incidental review of whatever you feel like - this should probably also have scoring off, or maybe use a different scorefile entirely. Card Selection likewise random, Limit New Cards (probably) likewise off. This will let you play around with categories that you want additional review on without mucking things up for your SRS study.
You can look at this as something a bit akin to a physical fitness regimen. The SRS profile is your cardiovascular exercise, basic workout you do every day to stay in shape, and the other two profiles are stretching (yoga etc) and strength training (weights etc).
Re your specific questions:
SRS essentially considers whatever categories you've currently selected to be the pool of cards that it's managing - it'll determine how many new cards you have / how many it ought to introduce / etc based on that. So mixing up categories in an SRS test stands a good chance of leaving you with a sub-optimal stream of new cards.
The two sort options don't override SRS, they simply determine what happens within an SRS test - sorting due cards randomly means that instead of showing you the cards that are most past due first, it'll show all of the due/past-due cards in a random order. So you'll still see the same cards, but a card that's a week overdue might nonetheless show up late in the test, so if you're canceling the test in the middle or you're not sure you're going to have time to do all of your due cards today then you might end up missing out on some cards that you really need to study. Sorting not-yet-due cards randomly does the same thing for cards that appear after the "you've studied all of your due cards for the day" alert.
Your last question I think is addressed by my profile suggestion above - you can indeed review those cards that way if you set "scoring system" to "none" and add a card filter for recently reviewed cards.
Testing the same card in a sequence of fields is something we also support, though it's a bit hidden - if you set "Subject selection" in Test Settings to "Score-based," you can have cards go through a progression of different configurations based on their score. This doesn't yet support switching between different test types, but you can practice writing in a self-graded test via the sketch box.
As far as SRS not working if you limit the number of cards: the whole idea of SRS is that you *aren't* going back and reviewing all of your cards randomly, you're reviewing the specific cards that the system thinks you need to review today. So if you don't review all of those cards then you're liable to forget them, and if we let someone simply set SRS to review no more than (say) 100 cards a day when they actually need to review around 200 to keep current with their vocabulary then they'll find themselves forgetting lots and lot of words because they won't be reviewing them often enough.
We are working on a new SRS system that will accommodate set numbers of cards per day, but it'll do it through prioritization - essentially, if you only want to review 100 cards in a day, it will try its best to keep as many cards in your memory as possible, but will remove some cards from active study / accept the fact that you're probably going to forget them in order to ensure that the remaining ones are reviewed often enough to stay in memory.
But for your specific scenario with the current system I think I'd suggest a multiple profile approach. Set up one profile for straight-up SRS - add only the categories you really care about, turn on "limit new cards" with a reasonable threshold, and do that profile first on any given day to make sure that you review the cards you need to and keep them in memory. Possibly with score-based subject selection so that you'll keep being tested on them in different ways.
Then, set up another profile for additional review of the cards that you've recently studied - you can use a 'card filter' to limit the test to only cards last reviewed in the last few days. I'd suggest that you turn scoring off (Scoring System = None in scoring) so that these reviews won't affect the cards' SRS scores. Card Selection system here should be Random rather than SRS, and Limit New Cards off.
Finally, set up a third profile for incidental review of whatever you feel like - this should probably also have scoring off, or maybe use a different scorefile entirely. Card Selection likewise random, Limit New Cards (probably) likewise off. This will let you play around with categories that you want additional review on without mucking things up for your SRS study.
You can look at this as something a bit akin to a physical fitness regimen. The SRS profile is your cardiovascular exercise, basic workout you do every day to stay in shape, and the other two profiles are stretching (yoga etc) and strength training (weights etc).
Re your specific questions:
SRS essentially considers whatever categories you've currently selected to be the pool of cards that it's managing - it'll determine how many new cards you have / how many it ought to introduce / etc based on that. So mixing up categories in an SRS test stands a good chance of leaving you with a sub-optimal stream of new cards.
The two sort options don't override SRS, they simply determine what happens within an SRS test - sorting due cards randomly means that instead of showing you the cards that are most past due first, it'll show all of the due/past-due cards in a random order. So you'll still see the same cards, but a card that's a week overdue might nonetheless show up late in the test, so if you're canceling the test in the middle or you're not sure you're going to have time to do all of your due cards today then you might end up missing out on some cards that you really need to study. Sorting not-yet-due cards randomly does the same thing for cards that appear after the "you've studied all of your due cards for the day" alert.
Your last question I think is addressed by my profile suggestion above - you can indeed review those cards that way if you set "scoring system" to "none" and add a card filter for recently reviewed cards.