SRS: best 'environment' to remember?

[realmayo]

This question doesn't just apply to SRS I suppose: obviously SRS is a method of testing yourself and helping yourself remember vocabulary.

 

Reading a book about brain plasticity and the excerpts quoted below in particular made me think if my current SRS-ing is getting less and less successful because my brain is less and less interested in making me remember long-term the things I learn (or relearn) during SRS.

 

I normally work though my SRS deck at work, in between doing other things. So I'm not really concentrating.

Also SRS is such a routine and I've got such a big deck that it's neither new and exciting, nor is there any big obvious sense of progress or achievement. This of course wasn't the case when the deck, and my overall vocab knowledge, was much smaller. 

 

I'm also thinking about the way Imron and others periodically delete most of their deck to just have a more focused, relevant one -- that "freshening" might mean that the brain is more interested and therefore more ready to remember stuff (that is, make changes permanent)?

 

Any thoughts? Any ways to improve the environment or make the brain think the process of vocab review is more important and exciting than it currently is for me, when I'm sat doing my daily SRS chore?

 

Coming across words "in the wild" naturally is the best solution, assuming you're engaged and interested in whatever's going on when you come across them. Just wish there was a way I could replicate that for the grind of SRS or regular written vocab lists.

 

Change is mostly limited to those situations in which the brain is in the mood for it. If I am alert, on the ball, engaged, motivated, ready for action -- the brain releases the chemical modulatory neurotransmitters that enable brain change. Again, it's helpful to think of them as on/off switches. When I'm in a learning mode -- alert, concentrated, and focused -- the brain's plasticity switches are turned "on" and ready to facilitate change. If I'm disengaged, inattentive, distracted, sleeping, twiddling my thumbs, doing something without thinking about it, or performing an action that requires no real effort to succeed on my part, my switches are mostly turned off.

 

Also: "The harder we try, the more we are motivated, the more alert we are, and the better (or worse) the outcome, the bigger the brain change.

 

And: "Initial changed are just temporary. They only become permanent if the brain judges the experience to be inherently fascinating or novel, or if the behavioural outcome is a good (or bad) one.

 

 -- Book is called Soft-Wired: http://www.soft-wired.com/

[OneEye]

Back when I was more hardcore, there were times that I just got completely burned out on SRS stuff. I deleted my entire deck, gave SRS a rest for a month or two and focused instead on reading or TV shows or something. Then, once I felt like I was ready to push forward again, I did. Never regretted it. Words that I had previously studied and then forgotten about in the interim either showed up again and were easy to re-learn, or they didn't and so it didn't matter.

[realmayo]

Right, that's certainly one way to recapture that earlier "importance" or "freshness" to the whole process, but obviously a more efficient way would be to have a deck that is not only less riddled with useless words, but also represents an important and fresh task -- without deleting.

[OneEye]

I'm not convinced that's more efficient. But I don't really like SRS in the first place. It's kind of like taking out the trash for me. I don't like doing it, but it's better than the alternative. So those breaks were very welcome, refreshing, and exciting. And they allowed me to solidify my knowledge at a new level, because every time I took one of those breaks, I had learned a lot since the previous break. I learn a lot during the break, too, of course, it's just that I learn different things.

 

Anyway, I'm not convinced that what might seem most efficient on paper will necessarily be so in real life. Our brains don't really operate according to set, predictable patterns. Variables such as sleep, diet, weather, stress from work, relationships, finances, and much more will affect how your brain processes and retains new information. Sometimes I feel motivated to do lots of SRS. Sometimes I don't. These are phases, and I tend to think that it's better to go with what my brain is trying to tell me than to force myself to do something that I really don't want to be doing. Like I said, I learn a lot in those non-SRS phases, too. A change of focus can do wonders. For me, an SRS phase is usually an intensive vocab-learning phase. Non-SRS phases might be more focused on listening, pronunciation, intonation, fluency, usage, reading ability, writing ability, or a combination of those and other things. Then when I'm ready, a change of focus back to intensive vocabulary learning is refreshing, and starting off with a fresh, new deck helps to ensure that I'll actually do the reps.

 

Then again, when I talk about those vocab-learning, SRS phases, I'm talking about the times that I'd be doing 3 different textbooks at a time, and also adding any useful vocabulary I came across in real life. I was adding hundreds of new words per week, and I was spending hours per day studying, adding words, and then reviewing them. I think the first time I did a big delete, I had gotten to around 10,000 words in my deck, and was doing well over 500 reviews per day regularly. That was a long break from SRS (2 months, I think). After that, the deletions were more frequent, maybe every 2000-3000 new words. Now that I'm in grad school though, I don't do any vocab reviews. I do use Anki some for other things, though, like studying for my classes and learning Japanese. If I move to Japan next year, I'll probably go back to heavy SRS use again. For a while.

[realmayo]

Yes I can see scenarios where I'd stop SRS, definitely. But for now I'm using it and am interested in the circumstances where I'm likely to remember things better, versus the circumstances where I'm not likely to remember well.

Our brains don't really operate according to set, predictable patterns. Variables such as sleep, diet, weather, stress from work, relationships, finances, and much more will affect how your brain processes and retains new information.

I certainly don't know enough about the topic to say that the book is right and you're wrong, but the guy is well-respected and if he has researched how the brain works, how the brain remembers things, then for me at least it's interesting to see if that can help me with remembering Chinese words.

[imron]

Never regretted it. Words that I had previously studied and then forgotten about in the interim either showed up again and were easy to re-learn, or they didn't and so it didn't matter.

 

This was a key epiphany for me that came when I first started doing a lot of reading.  Consumption of native content is a natural repetition system that shows you words in exactly the order that is relevant to you at a given moment in time, with words that are not relevant to you not showing up with any frequency.

 

Once I realised that, it completely removed all the pressure and worry about forgetting words (a fear that SRS plays up to), because as long as you are regularly consuming native content and making sure to learn a small amount of the words you don't know, then over time you'll learn all the words that are relevant to you, and if you forget something, it will come up again if it is useful.

 

The other good thing about the consumption of native content is that it really highlights the things that you don't know well enough yet (because you won't understand) and gives you both a real and immediate point to focus on, and immediate and practical feedback showing improvement.

 

 

I'm also thinking about the way Imron and others periodically delete most of their deck to just have a more focused, relevant one -- that "freshening" might mean that the brain is more interested and therefore more ready to remember stuff (that is, make changes permanent)?

 

For me, it's not the freshness of the deck that makes things easy to remember and makes the learning permanent.  The deck is just something to use to get vocab in to my short term memory, with regular reading and consumption of native content used to make the changes permanent.

 

What the small deck does allow me to do is short-term drilling with much more focus and intensity, which is important, because unless you are being very strict, SRS doesn't really train you for skills that you need to read.

 

For example say you want to be able to read Chinese and so you SRS a whole bunch of vocabulary in order that you'll know enough words to allow you to do that.  The problem is that the requirements for 'knowing' a word in a flashcard revision and 'knowing' a word for reading are not the same - although there is some amount of overlap.

 

To read comfortably you basically need instant recognition - preferably of both the meaning and the pronunciation, although with context, sometimes you can miss one or the other of those without it being too disruptive.  If you have to think or pause even slightly between seeing the card and knowing the meaning then that means you don't know the word well enough to read it at a comfortable pace.

 

If you are not drilling for that, then you are only training half of the skill that you need to make reading easier (i.e. only training 'recognition', and not training 'instant').

With a large deck it becomes unwieldy to do such focused revision because it increases the amount of time for revisions (due to a higher rate of failed cards), so what you end up with is a deck with a whole bunch of half learnt cards that you 'know', but don't know well enough to be useful in comfortably consuming and understanding native content.

 

Deleting your deck also gives you a quick easy way to discard dead weight, because either you'll already know the words well enough in which case deleting them is no problem, or you won't, in which case the important ones will show up again in your reading and you can learn them then instead.

 

What this leads to is a freshness and immediacy to your learning, which definitely helps to make it seem less like a chore.

 

 

But for now I'm using it and am interested in the circumstances where I'm likely to remember things better, versus the circumstances where I'm not likely to remember well.

 

Just be mindful that language learning isn't just about being able to remember things.  SRS might give you the comfort that you are 'learning', but that comfort can sometimes be an illusion because you might not be focusing your learning on the things that are relevant.

 

Ask yourself what is your purpose for using SRS and what is your end goal.

 

Do you ever evaluate how far away you are from your end goal? (this is something you should be doing regularly).

 

When you do, what do you find is the most immediate problem holding you back?

 

Now look at your learning habits (including use of SRS) and assess whether or not they are addressing that most immediate issue.

 

If vocabulary acquisition is not the main issue (and it probably won't be if you have a large deck) then you should be asking yourself what changes you need to make to your learning process in order to address that.

[imron]

but obviously a more efficient way would be to have a deck that is not only less riddled with useless words, but also represents an important and fresh task -- without deleting.

 

This is a contradiction.  By definition, if you are not deleting things from your deck, then over time it will stop being fresh.

 

You can't both be fresh and have a whole lot of vocab hanging around from long, long ago.

 

So the question then becomes how can you efficiently prune your deck so that it only contains fresh and relevant things?

 

I've found the most efficient way is simply to delete everything at regular intervals and trust in regular consumption of content to catch any useful words that I hadn't fully learnt.  That way my deck will only contain things that are recently relevant and fresh.

SRS really plays to the fear that you need to keep everything around forever otherwise you're going to forget it.  The reality is that with language learning it's ok to forget things because as long as you are regularly consuming native content and regularly learning a small amount of words you don't know, then you'll never forget anything important.

 

In the words of Doctor Strangelove, you should learn to stop worrying and love forgetting

[Ruben von Zwack]

 

SRS really plays to the fear that you need to keep everything around forever otherwise you're going to forget it.

That's it, your essay there is like a talk with a psychoanalyst*. I'm so gonna delete my decks.

It's a pity because it's satisfying to marvel at how your statistics and pie charts grow. But why then did Anki become a dread to me? The answers are already given by the others, up there.

 

*PS - you don't think you could you possibly put it into a Chengyu?

[imron]

I've discussed my reasoning for deleting decks in a couple of other threads too.  You might also find those discussions interesting:

 

http://www.chinese-forums.com/index.php?/topic/39946-manage-anki-reviews-double-productivity-like-a-boss/

http://www.chinese-forums.com/index.php?/topic/33464-how-big-is-your-anki-deck-and-how-many-leeches/

http://www.chinese-forums.com/index.php?/topic/23561-who-else-goes-both-ways-with-srs/?p=282531

[realmayo]

This has become a thread about the virtues of deleting SRS decks. I recognise there's a time and a place for that. But:

 

SRS is simply an artificial way to help remember and (re)learn words. People use lots of artificial ways to improve their language learning. Lots of drills that are more than just 'real life' acquisition.

 

Many people will try to make a specific effort to study vocabulary as they come across new words they want to remember.

 

That can be without SRS. That can be with SRS. That can be with an SRS deck that is regularly deleted. Whatever.

 

There are some "connections" the brain chooses to make strong, and later make permanent. Others it doesn't.

 

According to the book I'm reading, when it becomes a chore, or if there is background noise, or if it's dark, etc etc, and for other reasons too, then the adult brain won't bother remember this stuff. It wouldn't make sense for an adult to remember everything he sees, so his brain has to be selective, choose what it thinks is important.

 

I want my brain to think that the process of learning a few words -- whether it's SRS or 10 underlined words on a page of newspaper -- is important enough that I remember them.

 

I want to trick my brain into releasing the chemicals, the modulatory neurotransmitters (!), that will make the change (the connection between the word and its meaning) permanent rather than discarded.

 

It struck me that when I do my normal vocab work, I'm ticking lots of the boxes that the book mentions for when an adult won't remember stuff. It's a routine and boring part of my life. There's no obvious great reward that makes the task worthwhile. There's often background noise of people talking. And so on. That's why I wrote the initial post.

 

 

 

 

Just be mindful that language learning isn't just about being able to remember things.

 

Don't worry, I know it. SRS can be such a powerful thing for the first few thousand words that it's tempting to assume it's just as powerful when you go beyond that. In my experience, it isn't. My eyes are perfectly open to that. It's just that it suits my current circumstances, to keep things ticking over. I'm far more interested in how to better remember things in general -- whether it's SRS or a vocab list or a book or whatever it happens to be.

Edit: SRS these days only gets to me base camp: when I had a much smaller vocabularly it felt like it was hauling me three quarters of the way up the mountain.

[JustinJJ]

I've recently changed the way I use SRS. Previously I would learn a new word by finding a simple example sentence and SRSing it. If I read and understood the sentence I would pass the card, similar to the 'AJATT method'. I found that doing this there were a lots of words which I would only really be half-learning (e.g. would get stuck into my passive vocab) as it's easy to gloss over a card and pass.

 

To increase my focus, recently I have been doing a 'look, cover, write, check' process where I first read a sentence card, write/type it out and then check it. This way forces me to focus on all the words used together, and hopefully recall some of the words / sentence patterns more quickly in real-life.

 

The process takes me about 30 mins a day and I add 10 cards each day.   

 

I would like to be doing more real-life reading however I've found the novel I have been reading (撒哈拉的故事）to be too boring, so my reading consistency has declined a lot. I previously read 或者 and 许三观卖血记 and found them much more interesting (and so my reading speed and motivation was far better).

[Pingfa]

The way I used SRS, at my peak I was adding hundreds of words a day, but I always maintained the immersive environment. I would aim to delete more than I added.

If the SRS became too much of a burden I would delete even cards I was reluctant to part with - though honestly when my SRS was most active I mostly enjoyed doing it; I didn't commit myself to a schedule, I just always had the SRS active and would flick through it at my leisure throughout the day (it's not particularly hard to do hundreds of reps when you spread it out throughout a whole day).

 

I understand why many people are not fans of the SRS system, though, as it became quite a burden for me in the early stages, until I stopped scheduling reps.

For me the SRS just makes things so much simpler. Simply by being exposed to a word every now and then throughout the day I can assure higher frequency and higher recollection than I believe would be possible otherwise.

Furthermore, if I don't understand something, it doesn't go into the SRS, which gives one a more concrete sense of progress.

 

In any case, priority number 1 is maintaining momentum. The SRS is not necessary and there's no harm in discarding it if it is holding one back in some way.

[Ruben von Zwack]

This has become a thread about the virtues of deleting SRS decks.

But actually with the new input here, I intend to go back to Anki and use it more.

 

I'm far more interested in how to better remember things in general -- whether it's SRS or a vocab list or a book or whatever it happens to be.

Your book does not give suggestions there?

 

I agree that our brain is selecting all the time. And at the same time, it is producing a lot. When I was studying, I had to read up on theories of perceiving and creativity, and there seems to be the opinion that only about 10% of what we perceive is "real" and the other 90% are fluff (emotions, judgements, memories, etc.).

[icebear]

On pruning - I add excessively, and regularly search for and delete cards which have more than a 50% failure rate in the first 8 reviews. I find that ensures I'm only SRSing cards I've got a decent understanding of, which then makes the SRS system more like a reminder, rather than test. Most cards that pass that criteria are pushed out very far, very quickly.

[realmayo]

I agree that our brain is selecting all the time.

Specifically, apparently, it decides whether to release chemicals that will strengthen a new connection in your brain, or let it fade away. Apparently the prospect of a reward at the end of the task is one way to achieve this, if the material itself is not suprising or interesting. Not sure how to reward myself for 20 mins vocab work. Hearty self-pat on the back?

Delete cards which have more than a 50% failure rate in the first 8 reviews 

I like this idea. I'm also toying with the idea of taking out each week's worth of failed cards and relearning them away from Anki, away from the computer. That might add the necessary variety.

[roddy]

That might be worth its own topic - would any of you like to start one, or shall I split out the above few posts?

[JustinJJ]

I find that preserving my deck but adding cards at a reduced rate if it becomes too time consuming is a good way for me to measure progress e.g. how many words I have learnt. For those who delete your decks, what other ways do you use to benchmark your progress?

 

Clearly the 'natural SRS' is great and many people recommend extensive reading, which I was doing until I read a boring book. Often I prefer listening to the radio or tv programs for input as I can receive 听力 practise at the same time (e.g. exposure to different voices/accents which you can't get from a book) and select words to learn from the transcript. Are there additional benefits to reading over listening? e.g. If you had an hour to devote to input would you find an hour reading or an hour listening to be most useful?

[imron]

is a good way for me to measure progress e.g. how many words I have learnt.

 

The question then though is whether 'words learnt' is a good measure of progress for one's language ability.  It's progress of a sort, and obviously you need to have a good vocab base to work with, but the more advanced you get, the less meaningful it becomes due to the laws of diminishing returns.

 

For example, say person A has a deck with 10,000 known words and person B has a deck with 20,000 known words.  Going by words learnt, it looks like a huge difference, and it definitely would have been a huge difference in effort to build each of those decks.

 

If you look at it from a real-world perspective however, all that extra effort will only represent an increase in understanding of a couple of percentage points for most general purpose native level content, say 98% comprehension instead of 97% comprehension.

 

That extra 10,000 words of progress in SRS land, only translates to a tiny amount of progress in actual consumption of native Chinese content.  So yes, it's a way to measure progress, and it's very easy to measure compared to other things such as percentage understood, but after a certain point it's not the sort of progress you want to be focusing on if you have other skills that need improvement (e.g. instant recognition of words you already know when used in context).

 

It comes back to what I mentioned earlier about thinking about your end goals, and regularly trying to measure progress based on that.  It's not going to be easy to measure the progress of such other skills but so long as you are putting in the work, you will improve.

 

Do you want to be able to read books?  Then try reading a book and find where your problems are.  What do you struggle with?  Find a way to fix that.  If the book is boring or too difficult, swap it out for something easier and/or more interesting.

 

Do you want to be able to watch TV?  Then try watching a TV show and find where your problems are.  What do you struggle with?  Find a way to fix that.  If the TV show is boring or too difficult, swap it out for something easier and/or more interesting.

 

 

For those who delete your decks, what other ways do you use to benchmark your progress?

 

Total books read (I have bookshelf full of them now), TV shows watched, days in a row continually spent working on a given skill, going back and revisiting stuff I've read/watched previously and noting if I find it easier etc etc.  It's not as clear or precise as what you get from SRS, but it's still easy enough to measure and notice.

 

 

Are there additional benefits to reading over listening? e.g. If you had an hour to devote to input would you find an hour reading or an hour listening to be most useful?

 

You need to clarify whether this is an hour to devote to input including learning and revising new words you encounter, or an hour to devote to input, with more time available after that for learning and revising the new words you encountered.

 

If it's the former, you will really only have 1/2 an hour to devote to input because you'll need to spend at least that amount on learning a small amount of the vocab you encounter.

With regards to listening over reading, they are both important skills, and there will be some overlap between them.  Personally I'd mix it up and do listening some days and reading on others, or maybe if you have the time, to do a bit of both.

[WestTexas]

I want to trick my brain into releasing the chemicals, the modulatory neurotransmitters (!), that will make the change (the connection between the word and its meaning) permanent rather than discarded.

 

It struck me that when I do my normal vocab work, I'm ticking lots of the boxes that the book mentions for when an adult won't remember stuff. It's a routine and boring part of my life. There's no obvious great reward that makes the task worthwhile. There's often background noise of people talking. And so on. That's why I wrote the initial post.

 

 

The thing is, SRS is designed to work around the brain's typical system of remembering things. The guy who designed the Supermemo algorithm has his own PhD and is also an expert in memory. If you are trying to concentrate more on SRS, I would suggest either (1)Doing it in the morning or (2)Having a specific window of time to do it each day.

 

For me, I do SRS each day while I microwave my oatmeal and wait for it to cool. I have around 16,000 cards in my deck but only do like 90 reviews per day. This means I can get all the reviews done in less than 10 minutes. I have not been adding cards recently, but if I add 10 cards a day, I will still only have around 120 cards per day.

 

If the reviews are getting to you, I think there are better steps to take than deleting your whole deck. One thing I did was to go into the browser (hit ctrl+f) and sort by the number of reps. Then delete everything with above a certain number of reps. Or sort by difficulty and delete everything with below a certain difficulty. Statistically, you will find that a low percent of your cards (5-10%) account for a large number of your repetitions. This means that removing the most troublesome cards can drastically reduce the number of reviews you have each day. This is how I have 16,000 cards but only 80-90 reviews per day.

 

A large deck size does not necessarily lead to a large number of reviews. Why? As the card intervals grow longer, it is reviewed less and less frequently. One can see that, for example, if the average interval in your deck is one year, and you review 50 cards a day, you could potentially have a deck of over 15,000 cards. In fact, most reviews come from either (A)Newer cards or (B)Cards that are repeatedly missed. By simply deleting (B) and by lowering (A), you can have a big deck but not spend so much time reviewing.

[WestTexas]

If you look at it from a real-world perspective however, all that extra effort will only represent an increase in understanding of a couple of percentage points for most general purpose native level content, say 98% comprehension instead of 97% comprehension.

 

That's true, but here's another way to look at it: The second person in your example misses 50% more words than the first.

 

Having said that, I'd agree that SRS becomes less useful beyond a certain point. Personally, I kind of like SRS so I do it anyway, but I think it becomes less useful as you learn more and more words. It's still useful, just not as useful. 

 

Like for me, I have stopped adding new cards to my SRS deck regularly. I add particularly useful terms or words I feel like I should know, but it's no longer a matter of "I need to do X new words per day."