SuperMemo: The brain cell trainer

Andrzej Horodenski, Computer World, Poland, June 1992

This article was published in Computer World Poland in June 1992, and was one of the first independent press reviews of the SuperMemo method and software. Comments of the translator - Monika Morawska - are placed in square brackets

There is no shadow of a doubt that SuperMemo is one of Polish software sensations of recent months. SuperMemo is a prizewinner of national finals and a Polish representative for the CeBIT competition Software for Europe 1992. It is frequently present in the columns of Polish press, on the radio, and it also became a subject of private conversations.

And yet, despite this uproar, few people know how SuperMemo works, and even fewer have seen it with their own eyes [1992].

SuperMemo, beyond any doubt, is outstanding in one sense (which at the same time makes its major weakness): it is a product of the intellectual effort of someone who lives in this country [Poland], is known by a typically Polish name [Wozniak] and ... apart from the invention of SuperMemo, does not differ from an average man in the street. It is a cardinal sin of marketing! People like to have a feeling that they play with a toy that aroused a huge excitement abroad, esp. beyond the Atlantic! This variety of internationalism, cleverly fueled by advertising, results in the fact that Polish girls love the nightmarishly kitschy Barbie doll, and are less inclined to play with dolls coming from the regional cultural circle. So there is a significant chance that SuperMemo will become success in Poland only then when it becomes truly famous abroad!

Most of all, SuperMemo is a method of fast learning whose principles were developed on the basis of the knowledge of dynamic processes of learning and forgetting.

In order to effectively remember knowledge, it is necessary to repeat the material in specific intervals of time. Determining the optimum sequence of repetitions is the essence of the SuperMemo method. A computer program, with the same name, SuperMemo, has been designed just for the purpose of optimally scheduling repetitions.

The work with SuperMemo looks as follows: the program asks successive questions to which the student replies on his or her own. After replying, it is necessary to compare the given answer with the correct one (presented by SuperMemo after pressing Enter). After the comparison of the answers, the student must grade himself or herself. The grade is fed back to SuperMemo via keyboard. When we end a session with Esc, the program goes through the final drill. In the final drill, SuperMemo drills the questions that have received a lower grade until they are all mastered perfectly.

At first sight, this approach might be questionable, esp. the part with grading yourself! The practice shows, however, that self-grading is not a SuperMemo's weakness! The author of the method defines a scale of grades very precisely. The temptation to inflate one's grades is not significant. After all, it’s obvious that the user would not cheat anybody else apart from himself or herself. An important benefit is that the user is free from typing the answer on the keyboard, which is usually required in other programs for learning.

In the beginning, the program assumes a student's average learning ability, but in time, it will know more and more about the user and schedule repetitions with increasing precision. So the effectiveness of learning increases with time.

SuperMemo provides a tool for creating one’s own databases (if I could only convince my son to write down some dates in history and review them with SuperMemo from time to time!). It also makes it possible to trace one's own progress in a precise language of the parameters of the mathematical model of the SuperMemo method, and even helps one to study his or her own memory [see: Statistics and Analysis]!

The menu layout and interface are planned well and do not cause any problems at work. But it will not be easy to explain to Polish customers why all menu options are displayed in English, esp. that the leading application of SuperMemo is ... learning English!

The first contact with SuperMemo usually arouses enthusiasm: “This is something for me, I must get down to it at once!” - this is quite a typical reaction. I don’t want to say that the enthusiasm is not justified; it is. I only want to warn a new user that beyond the enthusiastic phase, there is a phase of disappointment, which builds a barrier that is difficult overcome for many people (a law in psychology stands behind it, if I am not mistaken). The beginning of work with SuperMemo should be preceded with a serious consideration, because this is not a program with which you may play from time to time only. I checked it myself using a database for learning English.

SUPERMEMO REALLY WORKS! After a month (half an hour per day) my head was packed with all kinds of words and phrases, which I would never before suspect I could remember! How can an ordinary city dweller remember words such as tree trunk? I have also long lost hope that I will ever be able to remember the English word scissors (there are a couple of words that by no means want to stick to my memory!). And here, voila: tree trunk and scissors. No problems at all!

And here we are coming to a quite dangerous moment – the euphoria coming from the effectiveness of SuperMemo makes the user memorize more and more new items. As a result, the amount of outstanding repetitions grows quite quickly, and sessions get longer and longer. Tiredness and lack of time come into play, and we begin to skip some sessions. Unfortunately the computer does not show any mercy – lesson that is not done today must be executed tomorrow as an outstanding lesson. The outstanding repetitions will double if you skip two sessions. After a two-week break you are cooked! A longer trip and ... result is the same. Naturally, this is not an accusation of the method or the program. This is just a warning to beginners – if they come to a conclusion that the program falls short of theirs expectations ... this is in all likelihood their own fault.

There is no escape from the truth: if we want the method to work, we must meet its conditions. Regular repetitions are necessary for success. In the same way like turning on the gas is necessary to boil water to prepare a cup of tea.

If you have already fallen into the trap of excessive repetitions, you got only one choice: clear the learning process and start from scratch. I think it should be possible to automate such situations. The program should be enhanced by an option that would help reduce the amount of repetitions by removing or postponing best-remembered items. This option could be called Mercy [see footnote]

[few weeks after publishing the presented article, SuperMemo 6 for DOS was enhanced by an option for rescheduling repetition. It has retained its originally proposed name: Mercy. Mercy has since become a frequently abused tool by all students of SuperMemo! We have not heard from anybody who would have gone with SuperMemo for a year with over 100 repetitions per day without using Mercy at least once!]