RECENT TOPICS » View all

Spidey/How do Check Yr Progress reviews affect SRS-learning

Index » RtK Volume 1

« Previous
1
Next »

From: Perth Australia Registered: 2009-11-09 Posts: 52

1) Hi. First, I’ve just finished Lesson 36 and I want to say a big thank you to everyone who contributed Spider-Man stories. When I began the lesson I wasn’t feeling hopeful about being able to create a whole lot of stories about ‘thread’. I knew very little about Spidey, but I chose him for my primitive in order to give myself plenty of stories to be inspired by. I haven’t laughed or totally enjoyed any other lesson quite so much. (And can’t get Homer’s rendition of Spidey-Pig out of my mind either!…) Thank you all.

2) I’ve found that giving personalities to the primitives has really helped (e.g. Spider-Man). I’m finding it much easier to visualise the stories and I think it’s taken me this long to get the hang of what Heisig was trying to teach us about using imaginative memory. Perhaps like some other people interested in language-learning, I’m primarily a “words” person and actually find imagination exercises very difficult. (I usually miss the first scene in a movie because I’m reading the opening credits and can’t follow the action at the same time – the words jump out at me – whereas I suspect a lot of people don’t notice the credits….)

In the earlier lessons – before, say, 人 was introduced, my stories aren’t nearly as easy for me to visualise and, as time goes by and the earlier kanji appear less frequently in the reviews, I’m failing ones that I really thought I knew. It’s frustrating to think they were only in my short-term memory, after all. So I’ve been thinking about using the Check Your Progress tests to go over those earlier lessons – and try to create better stories for ones I fail. But I don’t understand what the implication of that will be for my SRS learning. Will reviewing them outside the SRS effectively put them back into my short-term memory? So when they come up next time, I’ll have just “re-learnt” them and won’t know if I would have remembered them without the progress test reviews. Sorry if that doesn't make any sense. Can anyone help me understand how it works? Thanks.

dtcamero Member

From: new york Registered: 2010-05-15 Posts: 653

This is a common myth/misunderstanding about how the SRS actually works. There is not a magic time that the SRS divines somehow in order to make you not forget (as you found out yourself) that you can't disrupt or it will make your memory lose its balance and fall off the anki bicycle...
It simply calculates when you are likely lose the material from your active and passive memory, to just your passive memory, and schedules the review for that time. In this way you will eventually know 95% of everything with the least amount of work possible.
It is an efficiency shortcut, it's not about memorizing things quickly. If you want to memorize something faster by all means review it yourself outside the srs. When people say you should read outside of the srs, every word you read is like a little review... doesn't impede your progress when you open up anki.

Last edited by dtcamero (2012 August 26, 11:21 am)

scarby dancer Member

From: Perth Australia Registered: 2009-11-09 Posts: 52

"When people say you should read outside of the srs, every word you read is like a little review... doesn't impede your progress when you open up anki."
Thanks, dtcamero, that makes sense.

"It simply calculates when you are likely lose the material from your active and passive memory, to just your passive memory, and schedules the review for that time."

May I ask, then, what the factors are that affect that calculation? In the last 12 days, for instance, the number of kanji that have come up for review has ranged from 38 to 76, with the average being 60. That doesn't seem like very many to review each day, given that I've added about 1400 in total cards so far. I've read about people saying the number of reviews gets really intense - in the hundreds - whereas I feel that I'd like more cards to come up for review more often. Is it to do with how quickly/slowly one is working thru RTK? [I started in April and will finish mid-November.] I'm also wondering if I choose "remembered with some effort" versus "remembered easily" it affects that calculation? If I keep choosing "remembered with some effort" does that bring those ones back more often and therefore displace others so that it takes longer for them to return?

[Oh, btw, I used the term "SRS" meaning this site, not anki. Sorry if that wasn't accurate.]

JapanesePod101 Sponsor

undead_saif Member

From: Mother Earth Registered: 2009-01-28 Posts: 635

@dtcamero But I read somewhere that the shorter the interval between reviews, the less the fact goes into the long term memory, is that a myth too?

dtcamero Member

From: new york Registered: 2010-05-15 Posts: 653

scarby dancer wrote:
May I ask, then, what the factors are that affect that calculation? In the last 12 days, for instance, the number of kanji that have come up for review has ranged from 38 to 76, with the average being 60. That doesn't seem like very many to review each day, given that I've added about 1400 in total cards so far. I've read about people saying the number of reviews gets really intense - in the hundreds - whereas I feel that I'd like more cards to come up for review more often. Is it to do with how quickly/slowly one is working thru RTK? [I started in April and will finish mid-November.] I'm also wondering if I choose "remembered with some effort" versus "remembered easily" it affects that calculation? If I keep choosing "remembered with some effort" does that bring those ones back more often and therefore displace others so that it takes longer for them to return?

I only know anki, and I'm not an expert on the algorithym, but I believe it takes your current interval, and multiplies that by a fixed number depending on your answer (1.7ish for easy, 1.1ish for hard, etc.) and then that is all affected by an ease factor, which is dependent on how many "easy" and "fail" answers you've given that card... like a momentum multiplier.
you can read more about it here:http://ankisrs.net/docs/FrequentlyAskedQuestions.html#_what_spaced_repetition_algorithm_does_anki_use

yes your reviews are dependent on how many cards you are adding... so if you are doing RTK over an 8-month period you won't have heavy reviews. Someone here did RTK in 15days and had hundreds of reviews/day thereafter. Alternatively if you want your new cards to stay at the same pace, you can artificially grade yourself poorly (all hards for example) to get more reviews.

undead_saif wrote:
I read somewhere that the shorter the interval between reviews, the less the fact goes into the long term memory, is that a myth too?

I know what you're trying to say here... I don't think it works that way... The interval is determined by how well you know that thing, it's not the other way around. Anyways if you're really immersing yourself, then there are no long reviews because you will keep seeing that word over and over again in your immersion, which just becomes your normal existence eventually.

I'm no expert but I have been using anki for years. Things get into your head in weird ways that can be very different and difficult to assign causality to... but in the end anki is no different from encountering a word in a book or radio or streetsign... if you see them regularly you know their meaning. Lets say hypothetically that there is a memory bonus to a long-interval review... but surely having had 3 short-interval reviews in that same period seems more beneficial. What if it were a streetsign...wouldn't you know that frequently seen streetsign better??

for the record, when I was doing RTK I used the iknow website and it's SRS schedules crazy amounts of reviews. Adding 25 cards/day and I was doing on average 10hrs of reviews, and had a very good grasp on things. So I can be your guinea pig for reviewing early... no harm done. Worst case is you waste a lot of time drilling stuff you would know already.

Last edited by dtcamero (2012 August 27, 1:53 am)

undead_saif Member

From: Mother Earth Registered: 2009-01-28 Posts: 635

dtcamero wrote:
undead_saif wrote:
I read somewhere that the shorter the interval between reviews, the less the fact goes into the long term memory, is that a myth too?
I know what you're trying to say here... I don't think it works that way... The interval is determined by how well you know that thing, it's not the other way around. Anyways if you're really immersing yourself, then there are no long reviews because you will keep seeing that word over and over again in your immersion, which just becomes your normal existence eventually.

I'm no expert but I have been using anki for years. Things get into your head in weird ways that can be very different and difficult to assign causality to... but in the end anki is no different from encountering a word in a book or radio or streetsign... if you see them regularly you know their meaning. Lets say hypothetically that there is a memory bonus to a long-interval review... but surely having had 3 short-interval reviews in that same period seems more beneficial. What if it were a streetsign...wouldn't you know that frequently seen streetsign better??

for the record, when I was doing RTK I used the iknow website and it's SRS schedules crazy amounts of reviews. Adding 25 cards/day and I was doing on average 10hrs of reviews, and had a very good grasp on things. So I can be your guinea pig for reviewing early... no harm done. Worst case is you waste a lot of time drilling stuff you would know already.

Thanks for the reply. What I had in mind too, was the following situation:
If I reviewed something using Anki and I had a good memory of it, then in the middle or toward the end of the interval until the next review I encountered the fact which refreshed my memory of it. When the review time was due, I found it "easy" to remember because it wasn't so long before I saw it (of course without consciously remembering that), so the interval became much longer, but I didn't encounter the fact until the next review, so this time it's actually way longer than the one before, remember that the previous interval was cut! I would most likely fail it this time!
I hope I made it clear. SRS can never be ideal I guess!

Last edited by undead_saif (2012 August 27, 7:10 am)

partner55083777 Member

From: Tokyo Registered: 2008-04-23 Posts: 397

undead_saif wrote:
If I reviewed something using Anki and I had a good memory of it, then in the middle or toward the end of the interval until the next review I encountered the fact which refreshed my memory of it. When the review time was due, I found it "easy" to remember because it wasn't so long before I saw it (of course without consciously remembering that), so the interval became much longer, but I didn't encounter the fact until the next review, so this time it's actually way longer than the one before, remember that the previous interval was cut! I would most likely fail it this time!
I hope I made it clear. SRS can never be ideal I guess!

Yup, this is pretty much how it is supposed to work. There is no way around this unless the ONLY expose you get to Japanese is through Anki.

Don't let Anki control you. And remember, failing a card is a good thing. It gives you time to build stronger memories.

dtcamero Member

From: new york Registered: 2010-05-15 Posts: 653

partner55083777 wrote:
undead_saif wrote:
If I reviewed something using Anki and I had a good memory of it, then in the middle or toward the end of the interval until the next review I encountered the fact which refreshed my memory of it. When the review time was due, I found it "easy" to remember because it wasn't so long before I saw it (of course without consciously remembering that), so the interval became much longer, but I didn't encounter the fact until the next review, so this time it's actually way longer than the one before, remember that the previous interval was cut! I would most likely fail it this time!
I hope I made it clear. SRS can never be ideal I guess!
Yup, this is pretty much how it is supposed to work. There is no way around this unless the ONLY expose you get to Japanese is through Anki.

Don't let Anki control you. And remember, failing a card is a good thing. It gives you time to build stronger memories.

Ya I'd say my first reaction is that is probably a large part of the 5% we can't get a handle on...
but then partner's response is legit too, and it's good to remember that that 5% is always changing. As you miss that review, you'll be slapping yourself in the face feeling silly, and then as the intervals become short again you'll end up giving it relatively more attention than other cards, really learning it... Sometimes these hard-won cards become the most firmly cemented in one's memory.

in the end, although SRS is an imperfect system, it's kind of like an article of faith that you just keep doing it, keep immersing, and you will steadily get better. you don't need to understand how it works or be 'good' at using it. it's a very forgiving process.

undead_saif Member

From: Mother Earth Registered: 2009-01-28 Posts: 635

@partner55083777
@dtcamero
I've never thought about failed cards that way.
What you said sounds logical. Thanks guys!

scarby dancer Member

From: Perth Australia Registered: 2009-11-09 Posts: 52

partner55083777 wrote:
And remember, failing a card is a good thing.

Wow, is there a "Quote of the Week" thread on this site? That deserves to be there! It's such a shocking concept - which shows just how ingrained I am by traditional teaching methods that are all about passing exams and getting a high score - while not necessarily learning anything longterm.

@partner55083777
@undead_saif
@dtcamero
Thanks for your discussion. I'm not using Anki but I think it all applies.

partner55083777 Member

From: Tokyo Registered: 2008-04-23 Posts: 397

scarby dancer wrote:
partner55083777 wrote:
And remember, failing a card is a good thing.
Wow, is there a "Quote of the Week" thread on this site? That deserves to be there! It's such a shocking concept - which shows just how ingrained I am by traditional teaching methods that are all about passing exams and getting a high score - while not necessarily learning anything longterm.

I actually stole the idea from AJATT.

undead_saif Member

From: Mother Earth Registered: 2009-01-28 Posts: 635

partner55083777 wrote:
scarby dancer wrote:
partner55083777 wrote:
And remember, failing a card is a good thing.
Wow, is there a "Quote of the Week" thread on this site? That deserves to be there! It's such a shocking concept - which shows just how ingrained I am by traditional teaching methods that are all about passing exams and getting a high score - while not necessarily learning anything longterm.
I actually stole the idea from AJATT.

Haha, I like your honesty!

« Previous
1
Next »