How to read a book in an hour

Jump to navigation Jump to search

This article by Dr Piotr Wozniak is part of SuperMemo Guru series on memory, learning, creativity, and problem solving.

Massive reading brings massive forgetting

Some people claim they read a book a day. However, without a good system of notes, their recall may be pretty poor. Forgetting wipes most of the effort in just a few weeks. The more they read, the more they forget. Reading books is a wasteful process.

In this fun video, Matt D'Avella presents the problem as well as the solution. His solution uses paper. This text presents an analogous solution that uses software. The greatest advantage of going digital is (1) speed, and (2) lifelong recall with the help of spaced repetition.

In the presented technology called incremental reading, "notes" are generated naturally in the reading process. The entire reading investment provides a capital for the rest of one's life. Portions of the book can be read or reviewed again and again. Most of all, quintessential portions can be etched in memory with nearly costless cloze deletion, and with very cheap review based on spaced repetition.

Massive reading of books makes little sense without deploying some weapons against forgetting

I do not read books

I love to claim provocatively: "I do not read books". In the era of declining readership, it is a good starter for a conversation about the horrible state of education.

My claim that I do not read books is actually correct. I use incremental reading, which is not linear (like reading). It requires electronic texts (paper not welcome). It rarely brings me to the point of saying: "I have read the book. Entirely". However, increading is a superior form of reading. I can read a book in an hour, and come out with a good understanding of the core message. Moreover, every minute of that hour is an excellent lifelong investment in learning.

Paper books are dying. It is time to go digital

Books still matter

Recently, I incread more and more books. With factual knowledge becoming easier and easier to get. With assistance of ChatGPT or Bing, I get most answers fast enough to take knowledge access for granted. This brings a strengthened realization of the unique value of human models of reality. Consistent models based on scientific consensus are easy to get. Inspiration that comes from a lifelong effort of great minds became the best source of new inspiration in the quest for new truths. In creatively jumping out of the mental box of old models, artificial intelligence is still awfully inadequate. Almost parochial. It will stubbornly refuse any departure from the consensus, esp. in the areas of controversy. In my case, the harms of coercive schooling are the area where chatbots are the least helpful and the least open-minded. In a human-like fashion, they can interrupt a conversation as if losing patience with my persistent questioning of their reasoning.

As we get older, we tend to build new layers of our own models of reality in the brain. The more we think, the more we learn, the more elaborate those structures. A great deal of those models goes awry. They often turn out wrong. This is what makes them inspirational (see: Value of wrong models). This is why a book written by a thinker in his 70ies or even 90ies is such a treasure.

Great people are still superior to AI in providing unique inspiration

Internet archives

The Internet Archive provides more and more books in electronic form. This is why I tend to employ a "book in an hour" more and more often. The algorithm is useful for reading classic pieces that provide insight into the thinking of philosophers, dictators, criminals, inventors, madmen, etc. It helps build models of human mind and how the conceptualization process drives the adaptation to this world. I like to know why we become who we are?

Below I present the outline of the methodology with quick hints about why incremental reading is the best tool for the job.

Importing books

The whole process begins with importing the books to SuperMemo (see the manual to find optimum ways depending on the version of the program and Windows). The initial steps involve cleaning up the book (e.g. filtering scripts and advertising), formatting, adding references and, usually, a mnemonic picture illustrating the book (e.g. just the book cover).

Splitting large books

Very large books may slow down SuperMemo on slower computers. This can be remedied fast with Split the article. Splitting the article is also a good opportunity to get a general grip of what the book is about. In the first rush through the contents of the book, which may take 3-10 min, I mark chapters and important sections with horizontal lines, which are used as split points. This is equivalent to previewing a book in a bookshop. The end result is a set of chapters and/or subchapters which are all assumed to be of similar priority. In my case, the first split usually produces roughly 10-100 pieces. However, more meticulous review may produce more.

Splits meet the old speed-reading principle: skim before you read

Reading, extracting and splitting

In incremental reading, the process of reading never ends. The pieces of the book kept being reviewed and analyzed at different times at different sequences using a set of algorithms that suit the needs and goals.

The reading of a single piece of texts follows the standard approach known from incremental reading:

  1. read the text from the top
  2. extract important pieces for further processing at high priority (Alt+X)
  3. extract larger skimmed pieces of text that can be read later with lower priority
  4. delete or dismiss portions that will never need to be visited again
  5. memorize important facts with cloze deletion that will help keep knowledge in long-term memory

Independently, spaced repetition is used to review key pieces of knowledge to provide a lifetime recall of the essence of the book.

See also: Incremental reading step by step

The strategies for prioritization and scheduling of multiple pieces will depend on your need:

  • you may rely on the incremental reading process and have your pieces served naturally in competition with the rest of your learning material
  • when you think you need to make a more comprehensive review, use subset review, in which you may sort pieces as they occur in the book, by size, by priority, by review intervals, at random, etc.
  • if monothematic reading is tiring, you can intersperse your subset with your standard learning process by using Add to outstanding in the browser

For more details on strategies see the toolset described in sections that follow.

Cloze deletion

Cloze deletion replaces a keyword with [...] and forms a question that you can use in testing your memory in spaced repetition. For example:

 Question: Nobody reads much faster than [...] words per minute
 Answer: 400

There are two functions of cloze deletion:

  • consolidating memories associated with the book for lifelong recall
  • protecting high priority material from being lost in the crowd of books, articles, book fragments, etc.

The priority protection is important when you import a great deal of articles and naturally set high priority for them all. In incremental reading, cloze review is cheaper and protection of items is higher than the protection of topics. This is why a simple, cheap and productive trick for protecting vital books is to generate at least a single cloze. In addition to its usual memory function, the cloze may remind you what books you tend to neglect (if you happen to read dozens or hundreds at the same time).

Cloze is a tool that moves memories from books to your head with a click

Spaced repetition

Spaced repetition is the key technology that makes the whole "speed-reading" concept sustainable and productive. The speed of reading, omissions, and the seeming chaos are well compensated by the fact that whatever we learn from reading a book will be stored in human memory for life (given sufficient micro-investment in time).

Readers who boast reading a book a day must devote a significant chunk of the day to reading or need to engage in brutal skimming. Whatever the strategy, a month after reading, the ability to actively recall content of the book is negligible. There is still a great deal of passive value, but in well-schooled terms, the reader won't pass the test on the content of the book, even if the test was to be taken just a few days after reading!

In contrast, the procedure proposed in this text may turn out to be pretty spotty at first. There will be a great deal of skimming, but this is lossless skimming. Whatever we skim today, will safely return tomorrow. However, the main difference is that incremental reading will leave tangible traces in memory. Those won't be accidental titbits that attracted attention. These will be the golden nuggets that change lives.

Spaced repetition is what makes speed-reading books highly productive


Skimming is a natural component of speed-reading. Fluent readers recognize patterns in the text. This allows them to grasp the meaning by reading just a beginning of a section, filling up the missing parts with guesswork, and moving on to further sections with negligible harm to comprehension.

In "Speed-reading on steroid", I explain that incremental reading adds one powerful ingredient to skimming. It eliminates the FOMO factor. The mind is relaxed due to the fact that there is no risk of missing anything. The reader may miss something today, but all he skims is retained in the learning process. All pieces will keep returning until they are fully processed, deleted, dismissed, etc.

Incremental reading supports a lossless form of skimming

Sorting by text size

Having the read book in an hour, you are likely to have (1) some general understanding of the message, (2) couple of golden nuggets of knowledge to remember for life, and (3) hundreds of semi-processed splinters of the text.

The understanding (1) and the nuggets (2) are your stable productivity gains. The splinters of the book (3) are the material for future optional work at your choosing. Each time you engage in futher reading you will effectively increase the gains from the book. Your understanding of the message will crystallize, while the volume of precious finds will keep building up. Naturally, all such processes are subject to diminishing returns. After a few months, you are likely to abandon active decisions to return to the book. The material will integrate with the rest of your knowledge and compete for your attention on equal far basis.

Before that happen, you can use an informative tool for sorting by size. You can review the book starting from the longest pieces of text. Those are usually pieces that were most affected by skimming. If the length of the text is the result of low priority, you can mark its priority accordingly, and process speedily using techniques using in reading the book in the first place. That will knock the lengthy pieces from the top of the list and relegate it to lower echelons of priority. If you engage in reading hundreds of books in parallel, chances are, those low priorities will never enter the learning process in equal competition. However, you may see them in semantic review (e.g. due to the occurrence of specific keywords) or in neural creativity if the book happens to be link to other topics of high importance.

Sorting by text size helps the review of pieces most affected by skimming

Search and Review

Once the book is processed, it becomes integrated with the rest of the learning process, which involves reading and reviewing other books, reading articles, review of questions that are to be remembered, etc.

Depending on book's priority, individual pieces that are yet to be processed or analyzed may show up rarely in the crowd of other portions of knowledge. One of the likeliest ways of return is not regular learning, but subset review, when you need to review specific portions of your knowledge. If relevant pieces show up in the book, the book itself will show up in the learning process.

Let me illustrate the process on an example:

I heard that Frederick the Great used an alarm clock as a child. It was in the early 1700s. His dad ordered a wakeup call with a canon. This made me want to review my knowledge of "Frederic II". I used search and review. I found lots of forgotten stuff such as Frederick II - Holy Roman Emperor or Frederick II of Denmark.

After a while, I hit on a recently read book by Tom Durrie: School and the end of intelligence. The fragment I found spoke of the old dilemma: do rules want the masses to be uneducated? Does it apply to ruling elites in 2023?

As early as 1717 Frederick William I, thinking that some education might keep the peasantry under control and might produce better, more obedient soldiers, made a half-hearted attempt at setting up primary schools [...]. And the peasants were not easy to corral. At that time, and well into the nineteenth century, there was a fear that education might make the peasants uppity. This fear was well-expressed by Frederick II in 1779: “In country places, a little reading and writing should be enough, for if they know too much . . . [they] will run off to the cities and become clerks or some such thing.”

I recalled how much I get from Durrie's book. I gave up on my Frederick II review and went on to subset review of the book. This makes such explorations incredible fun. Unlike exploring with Google, I am searching my private knowledge space, and this allows for a significantly more enriching experience. Moreover, in incremental reading, no paragraph is ever lost to forgetting. It is either (1) lost to deleting (as irrelevant), or (2) set for future review, or (3) instantly converted into durable knowledge.

Each time you need to research anything, you generate a fantastic context for new learning

Adding to outstanding

Add to outstanding is a procedure that injects pieces of the book in your learning process in a set proportion. For example, you may wish to review other books, or memories from other sources, but get back to portions of the book every 6th review. You will then see five other pieces to read, review or recall, and then you will get another piece of text from the book.

Add to outstanding is the best way to combine your standard learning with reading a specific book without detriment to either.

You never need to read at the cost of everything else. There is a golden mean

Subset Review

You can store a book in a subset file and use it as the main starting point for reading. Each time you complete processing a portion, you can execute Done, Dismiss or Delete. All descendant elements will not be part of the subset. They may be re-included with Subset : Add in the Commander. However, if you stick to a single subset file, you can be sure you process the entire book piece by piece before you move on to the second wave of review. An alternative way to achieving the same is to sort the book branch by last review and proceed from the oldest pieces.

Random Review

Randomness in learning has a powerful cueing impact. This fosters creativity. For most people, reading books in random order of chapters or paragraphs is chaos. For an incremental reader, it is an occasion to learn new things, find new angles and new association. You cannot read a novel at random, but a science book is usually a perfect material.

Random review is a kind of subset review that you can do at anytime to assess the status of your incremental reading process. Random review can help you assess how much has been done, how much of the book is unprocessed, and how much you can get from further investment of time in the book. Due to the new algorithm of spaced repetition used in newer SuperMemos, random review can include items without the risk of spacing effect or other forms of interference in the learning process.

One less appreciated aspect of random review is that it can act as a motivator on a bad day. If you are in no mood for learning, you may fancy a specific book. If you jump into its content at random, chances are that you will encounter a piece that will revive your appetite for learning. It needs to be stressed that on a bad day, learning might be one of the best remedies to recover a productive mindset. If you ever experience insufficient motivation for anything productive, click Learn in SuperMemo. If it does not work (e.g. you added too much material required for school), recall your favorite book and run a random review on that book. See: Learning and depression

The don'ts

I googled, and I found quite a number of texts with an identical title and identical intent. However, not all advice makes sense. Here are some do's and don'ts garnered in a quick preview:

  • software that implements speed-reading by dishing words or characters at a set "optimum" speed (RSVP) violates many rules of incremental reading. If pattern recognition and skimming cannot be employed, comprehension and creativity will suffer
  • focusing on speed is a bad idea. Reading must be optimized for value per minute. This usually means reading big volumes in short time, however, when a valuable piece is found, slowdown is unavoidable and should not affect the optimization. My advice is the opposite to the usual, slow reading is actually an indication of value (assuming you are a fluent reader)
  • shutting off your inner monologue is a bad idea. It may serve speed, but it takes away the most important part of reading: creative expansion of the content. The purpose of reading is to improve thinking. So turning off thinking is the opposite of what we need
  • reviewing the table of contents should be optional and depend on the quality of the titles as well as the speed at which the meaning of chapters can be extracted in incremental reading
  • adding more attention to the beginning and the end of the book is natural, however, there are no optimum percentages or numbers of pages to stick to. Great value may be found equally well right in the middle. The greatest benefits of the start and the end are the guidelines and conclusions, which may or not be located there
  • I do use a timer in running my daily schedule (Plan), but I would never use it to time the reading itself. Plan tells me when I run out of time and need to focus on other jobs. If need be, I will come back to reading on the next day. There should be no limit on time. Time invested depends on the yield
  • I do not do mind-mapping, which does not mean it is a bad idea. I believe that some sort of mind-mapping happens in my head automatically and subconsciously. It is so subtle that I struggle to describe it. It might be the effect of many years of reading incrementally
  • a great deal of speed-reading theory refers to the saccadic movement (the way eyes trace the text). I never engage in any form of training or programming of the saccades. I bet the best training is a lot of reading, and the main difficulty is that a good reader automatically builds a sort of 2D semantic map of the read portion of the text (usually a sentence, or a paragraph). In that some more attention is paid to keywords, nouns and verbs, and there is more backtracking in passages that are hard to decode. Any form of intentionality may interfere with this process that is best optimized by the involved neural networks on their own. Speed-readers often see reading as if it was watching marathon. I rather see it as watching a football match. You never know where the action happens next
  • reading reviews and summaries, esp. in the era of ChatGPT is a great strategy, but it is not part of reading the actual book, so I did not include it in the described algorithm
  • similarly, I do syntopical reading all the time. I just never use the name. Syntopical reading is a great idea to read books in the same subject at the same time and thus build better generalizations. When I research the subject, I often import a great deal of material to incremental reading and then rely on planned interleaving to maximize inspiration (e.g. using neural creativity or subset review as described above). I reached the pinnacle of syntopicality, when I researched my dislike of fiction (acquired at school). I imported to SuperMemo the book "Lord of the Flies", two movies based on the book, and several summaries and complementary material from Wikipedia. This was fun, but it did not last long. Each imported piece had its faults, but in combination it all made some sense. I can say I know the plot of the book (roughly). However, I still do not read fiction


I do indeed usually read a book in an hour or less. I rarely spend more time on reading in a single session. On one hand, Plan rarely provides more room for reading. On the other, my attention tends to flag after 30-45 minutes. Even the best book can heat up the processor. I consider limiting the time to an hour as a good brain hygiene practice, however, if passion drives you to read for 5 hours straight, I do not think you should stop the joy.

Depending on the nature of the book, one hour may be enough for you to say "I read the book" with conviction, or just stay with the claim that you previewed the content. This largely depends on how data rich the book is. For a medical textbook, the density of important knowledge may be high enough for you to make a claim of reading only after a month of reading at 30 hours total. For books in psychology or philosophy, one hour is just enough to accomplish the mission. In addition, solid grounding in the subject makes a world of difference. A great deal of books reiterates well known facts or theories. In such a case, reading has a form of pleasurable nodding in agreement with the author.

Critics can argue for years on how much of a book you can truly grasp in an hour. My key message is then this: the presented algorithm is the best approach I know to maximize the pleasure and the outcomes of an hour invested in a book.

Maximization of outcomes is measured accurately with the pleasure of an hour of reading


You can see incremental reading in action in the accompanying video. However, before you dive into boring nitty-gritty of incremental reading, see this video by Matt D'Avella who provides a livelier illustration of the problem.

Once you know the goals and the outline of paper tools, see how this is done with incremental reading (video: How to read a book in an hour?). This process may seem boring for the observer; however, all the fun comes from the new knowledge you get when you do it on your own.

Further reading


If you disagree, Vent on Facebook

For more texts on memory, learning, sleep, creativity, and problem solving, see Super Memory Guru