Skip to content

Do Leeches downgrade SM efficiency?

May 16, 2009

Nattan (aka LittleFish) comments got me thinking and got me into search mode. Hope my whole brain had a search while you type mechanism; ok, maybe this is just to geeky.

Well, if you are current user of SM, or been trying to get help from the SM site, you now is not exactly easy to find a clear answer. Even if we do find the answer, could this be the only answer. This is one of the primary reasons why this blog got started, searching from SM true lovers. Those who want to go a step further in their SuperMemoing process.

“If you have many “wrong” answers on the same item, does this affect the overall algorithm in a bad way?” (LittleFish Comment)

So, does repeatedly answering bad (leaches), affect the SM algorithm making it less efficient?

At least 5 webpages on the SM site give clues to this question, although it repeats the same text and not surprisingly is an answer to another question. Because Nattan says he’ll experiment doing a separate set for leaches and I think that’s a good idea. I’ll try to analyze SM site response.

(David Mckenzie, New Zealand, Apr 8, 1998)
Is there any point in keeping collections separate?

No. Once you master  categories, templates, and subset operations, there is no point. You gain global search, global registries, global repetitions, global optimization, etc. Presently, the item difficulty measure (A-Factor, or absolute difficulty factor) is absolute and does not depend on the context in which an item is placed. Only the length of the first interval will significantly be affected by the average difficulty of items in the collection. However, this shall not bear dramatically on the speed of learning. Especially that variable forgetting index for individual items makes it possible to set different first intervals for whole contents categories or branches of the knowledge tree (

This answer does not conform to the simplicity principle, specially in the answer side of the item. Lets decompose the answer:

“No. Once you master  categories, templates, and subset operations, there is no point.” (SM site)

What about before your mastered? (if you ever really do). Personally I think that if you are just starting, a good measure would be to make separate collections. For how long? well, it depends on your proficiency with SuperMemo, but maybe not two long (3 months or so). After your collection grows big, you’ll want to use just one collectioin. (Except for testing purposes: themes, layouts, organizing, even data).

“Presently, the item difficulty measure (A-Factor, or absolute difficulty factor) is absolute and does not depend on the context in which an item is placed. Only the length of the first interval will significantly be affected by the average difficulty of items in the collection.” (SM site)

Presently? well, not other information says nothing different, so this affirmation should still be true. The second part, about first interval length significantly affected, shows that indeed it is affected, as me and other found out. For those of us who don’t want this to happen using a second collection until the elements on it or a branch on it has no further leaches would be a good alternative.

Finally, I do think leaches affect us all, either in a bad or good way, last post on unconscious memorization wanted to help us all acknowledge that no matter how many times we have to repeat something it would finally stick.

But perhaps for many perfectionist, who don’t want to trust the haphazard construction of our mind,  this is not enough. Those would like to give a try at Nattan’s recomendation of using two separate collections.

8 Comments leave one →
  1. LittleFish permalink
    May 17, 2009 22:01

    I’ve thought and thought about this (I went on a trip to see some family and was disconnected for about two days to think about this problem; Don’t worry, I kept up with Supermemo), and I’m convinced that having a “training grounds” database will be beneficial.

    I think one of the main principles of Supermemo is relevant to this problem: Supermemo is used to REMEMBER, not to LEARN. Information should already be “sticky” before it is put into Supermemo; Supermemo’s job is to make sure the “sticky” information stays stuck in your head. If information is not inherently “sticky” through violation one or more of the “20 rules of formulating knowledge,” it can thus easily be corrected, and the item becomes “sticky.” What though, do we do when certain information is, in it’s most basic form NOT “sticky?” How does one MAKE knowledge become “sticky?” Yes, mnemonics can help out a great deal, but as your first post pointed out, PURE REPETITION can cause certain information to “click” after some time, made “sticky” by sheer familiarity.

    I like metaphors. Since I’ve played lots of video games in my life, they are frequently used as metaphors in my head:

    If you have ever played a fighting game, it usually features a “training mode.” It allows you to fight any character (Which can be set at hard difficulties) for the purpose of honing your skills without the consequence of possible failure. If you or your enemy is hit, they quickly recover their lost health. Thus you feel free to experiment with characters, fighting styles, moves, etc. that you would not normally experiment with due to the risk involved. Here’s an example of the training mode in Street Fighter IV:

    The point? I think that Supermemo ASSUMES that you have gone through the “training mode” with ANY AND ALL information entered into it. It assumes you already know it and can recall it to a certain extent. Granted, some information is easier and more “sticky” than other information, but ultimately it has to have a degree of “stickiness” in order to be efficiently used in the cogs of Supermemo, otherwise it bogs everything down and falls by the wayside (Leeches!).

    So if one uses Supermemo merely to “make an item sticky,” I believe that this is using Supermemo for a purpose that it wasn’t initially designed to fulfill. ANY flashcard software (Or physical flashcards themselves) could be used to simply drill difficult information into your head, but what sets Supermemo apart is the ability to review with the LONG-TERM in mind.

    THEREFORE, if we want to take difficult information and make it easier to remember, creating a second Supermemo database (Or using other flashcard software) to “make it sticky” is a viable option. Again, I say this not having any real tangible results, but the more I think about it the more sense it makes.

    (Also, I realize I put quotes around “sticky,” which probably weren’t necessary after once or twice. But that’s how I roll, I guess)

  2. LittleFish permalink
    May 17, 2009 22:06

    (One other point)
    Having a separate database will allow for an experimentation with different mnemonic devices. If one is found that can be implemented across-the-board on many items, then you’ve found your answer. The separate database is merely a means to an end, and such a solution would be the ideal “end” we’re looking for.

  3. BlueMountain permalink
    May 17, 2009 23:24

    I like the amussment in wich you guys chat about using Super Memo, for me it has been a hit and miss until know, maybe I have been trying to hard in competition mode. Must get into training mode.

    • gersapa permalink*
      May 20, 2009 18:55

      Learning should be fun, I must, It is. If you enjoy learning then any thing that goes into helping you is going to be part of your life to. I advice to take SM by steps, an even better, as Nattan points out try out any idea in another database.

      • LittleFish permalink
        May 22, 2009 15:20

        Yeah, that is one thing I have learned after using Supermemo for so long: Learning is very enjoyable, it doesn’t have to be painful and boring. I SO WISH I had Supermemo when I was in school, maybe I would have cared more about learning rather than playing video games.

  4. July 31, 2009 14:46

    Frankly, one of the reasons I consolidate collections is to make more room for others I start (since the “File” section on the Menu for Supermemo only allows you to “quickly” open 10 collections). I frequently find myself starting a new collection to learn a new language, new keyboard shortcuts, etc.

    • gersapa permalink*
      May 31, 2010 14:08

      Talking about long delayed reply’s, well better then never right?. I like the idea of using more then one collection know. I didn’t think so in the past but I finding it useful now. The reason is that I have started a new collection with audio files on some preparation course, I want to incrementally process 1 min audio clips as items, the only criteria to score this item as good or not, is that I already now its contents, in this context I doubt my optimal factor matrix would be similar to other normal type of items. I also found this could be a flawed criteria when learning much more different languages, in any case its my appreciation only. But if you got a lot of different learning categories using your approach might be very flexible and effective.


  1. Using Leeches Collection to enhance learning. « The SuperMemo Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: