Incremental reading is speed-reading on steroids
This article by Dr Piotr Wozniak is part of SuperMemo Guru series on memory, learning, creativity, and problem solving.
Foreword
This is an update on an article originally written in 2014: Speed-reading on steroids (Piotr Wozniak, Nov 2014). The article explains why incremental reading makes it easy to deploy speed reading without detriment to comprehension. It also provides data on the retention outcomes and repetition costs in a speed-reading experiment in the course of 17 years.
Introduction
Most proficient readers rarely exceed 400 words per minute, even if they practice speed-reading. Furthermore, the faster one reads, the less likely they are to fully comprehend the text and retain the information over long periods. However, in the case of incremental reading, there is virtually no speed limit. Moreover, incremental reading guarantees maximum comprehension and a retention rate of 95-98% for a lifetime. This article elucidates how this seemingly impossible feat has become a reality for a growing number of SuperMemo users.
What is incremental reading?
Incremental reading is a technique that aids in transforming texts, such as those found on the web, into lifelong knowledge. When engaging in incremental reading, the learner extracts the most crucial parts of the texts and converts them into questions with a simple click. These questions are then employed in testing the student's memory using the spaced repetition system. This process ensures that the acquired knowledge can be retained for a lifetime.
What is the fastest reading method?
To an onlooker, incremental reading may appear as a combination of regular reading with additional computer operations. In other words, it may seem slow and burdensome. For a newcomer, this perception of sluggishness can be further exacerbated by grappling with the complexities of the software (such as SuperMemo), making difficult choices regarding article selection, struggling with prioritization, and lacking experience in selecting appropriate materials for extracting key information or generating relevant questions.
In this text, my aim is to establish that no other reading method can surpass the speed of incremental reading. Furthermore, I intend to demonstrate that this speed does not come at the expense of comprehension or recall. Let me be even more explicit:
Now that I have captured your attention, allow me to introduce some qualifiers that will explain that while this claim is true, it is not without its nuances. I would like to share a story in which incremental reading went head-to-head with classical speed-reading, and I will present some hard-earned data to support my arguments.
Why can incremental reading be very fast?
Incremental reading can be very fast mostly due to the fact that you do not need to worry about neither thorough comprehension nor recall during your first pass through the text. This contrasts it with speed-reading where comprehension is a measure of reading quality and maximum focus is essential for speed-reading to make sense. In other words, incremental reading can be employed as skimming without a speed limit, while comprehension and recall become a concern at later time. At the very extreme end of the reading speed, you can import a Wikipedia article, automatically split it into sections, fish for a few sections of maximum interest and set their priorities high for further processing in the incremental learning process (see: video illustration). All that should take mere seconds even if the article is a megabyte long! Nothing can beat that speed. Naturally, your initial inflow of new knowledge to your memory will be microscopic. Perhaps just a few general ideas on the subject.
For example, if you wanted to know what Biosphere 2 is, you might execute a speed-learning procedure that would take you just 2 minutes, give you a general understanding of the Biosphere project, help you illustrate it with 2-4 pictures, and prioritize the knowledge for future consumption. With hundreds of things to learn, very often, such a speedy rough outline is all that you need at the moment of first reading. In incremental learning, all well-prioritized pieces of information will compete for your attention. Their understanding and recall will depend on how you manage the learning process and how you shape your learning strategy. It is possible you will never go beyond the ABC of Biosphere 2. Perhaps you will even forget it. However, if the subject is important, you can incrementally dig into individual pieces of the imported text and ensure excellent recall for as long as you are ready to invest in repetitions with SuperMemo.
Speed-reading contest
When my friend Matt from America visited me in January 2006, he proudly shared his improving speed-reading skills, boasting a record of 2000 words per minute. He was eager to demonstrate his abilities, so I grabbed a book on American History and selected a test chapter for him to process. He breezed through the text so quickly that I was convinced his recall of facts would be rather superficial. However, his comprehension was remarkable. Although we didn't measure his actual reading speed to confirm the 2000 wpm claim, it was undoubtedly the fastest reading performance I had ever personally witnessed. However, there are a few important details to note. Matt's extensive knowledge of history made the test less reliable, as he included Alabama in the list of southern states even though it wasn't mentioned in the text. Additionally, the book was an enjoyable read that didn't contain too many difficult-to-remember or complex facts. To further evaluate Matt's skills, I asked him to tackle a more challenging text from my SuperMemo collection, where we could measure his speed under controlled conditions. We found an exemplary article on sleep that included a dozen sleep drug names. Fatigued due to the late hour, Matt admitted that the text was significantly more difficult, and his reading speed dropped to 200 wpm, only a tenth of his previous achievement. His comprehension also suffered compared to the American History text, primarily because he had less knowledge of the subject matter. Matt's experience demonstrates that speed-reading techniques can indeed help devour books at a rapid pace. However, when faced with complex and fact-rich texts, even the most proficient speed-reader will slow down to effectively comprehend and process the information.
The following day, I decided to test the same text using my own incremental reading skills in SuperMemo. As mentioned earlier, incremental reading has virtually no speed limit. I could skim and extract extensive portions of text without sacrificing long-term comprehension. The skimmed texts would reappear later in the learning process, perhaps within a day or a week. Matt required maximum focus and 7 minutes and 20 seconds to rush through the article. However, I knew that I could process it faster without concerns about focus or distractions from wandering thoughts. In other words, the first significant advantage of incremental reading is its stress-free nature (at least for experienced users). If you miss details during the initial reading, you can retrieve them later. Reading without the pressure to focus is enjoyable, and paradoxically, it improves focus! I managed to skim through the text in just 2.5 minutes. During this initial reading pass, I generated 13 extracts, hardly skipping any text. My comprehension was satisfactory, but my recall of specific details was quite poor. In other words, when in a hurry, I can easily outpace a speed-reader as long as I postpone the actual acquisition of knowledge. The best aspect of fast reading with incremental reading is the freedom, lack of stress, and the 100% guarantee that no detail will be missed unless I consciously decide to do so. Naturally, a skilled speed-reader would perform better in incremental reading than a typical SuperMemo user without speed-reading training. Years of experience with incremental reading can serve as a substitute for speed-reading training, although I'm unsure about its effectiveness as a complete replacement.
The value of speed with limited comprehension
You might be wondering, "What's the point of speed without comprehension?" During your initial pass, your focus is primarily on (1) prioritizing the material and (2) uncovering valuable pieces of knowledge. By reading at a fast pace, you can cover a hundred articles in a day instead of just a few. The following day, you can take your time to delve deeper into the most important findings from your speed-reading session. Depending on your specific needs, you may alternate between high-speed reading and slower, more contemplative reading:
Incremental reading: 8 years later
After conducting my initial speed test, I decided to repeat it on the same text on the same day. This time, I had to read 13 extracts that were generated during the first pass. The total size of the text would be 35% larger because extracts often include repeated context between individual elements. The second pass took me 5.6 minutes. This means that my two initial passes of the article only took 11% more time than Matt needed to process the article in a single pass. The best news was that my comprehension and recall increased tenfold. Despite not using a solid measure of comprehension during the 8-minute investment, my knowledge of the facts was quite solid, likely better than Matt's. However, as sleep is one of my favorite subjects, the playing field was somewhat uneven. After the two passes, we received double good news: (1) the first pass was very fast, and (2) the second pass provided solid comprehension.
Intrigued by the results, I decided to continue repeating the reading test. However, for a fair comparison with incremental reading, I needed to include an actual assessment of knowledge retention. In SuperMemo, this meant generating cloze deletions. Therefore, I chose to repeat the test at increasing intervals, allowing the active knowledge stored in SuperMemo items to solidify for long-term retention based on the principles of spaced repetition. Since the original goal of the test was to measure reading speed, I couldn't let the natural incremental learning process handle the extracts generated from the article. The objective was to read all the texts in a single pass on a single day for each review. To ensure minimal interference from my daily learning, I assigned the article and all its extracts the lowest possible priority. Over the years, I revisited the entire body of generated knowledge five times: in 2006, 2007, 2008, 2010, and 2014. During each pass, I randomized the elements to eliminate any sequencing effects. Naturally, items generated from a single article may interfere with each other and distort the retention measurements. However, in SuperMemo, one can choose their desired retention level, and retention wasn't a part of the experiment. We can safely conclude that in SuperMemo, regardless of the testing time, the retention is always excellent. My measurements also showed excellent retention (94%), which was influenced by the arbitrary choice of intervals and increased by item interference.
All incremental readers are aware that the cost of processing long texts tends to increase over time, particularly when students delve into the details of individual paragraphs and start generating cloze deletions. However, due to the usual learning overload, the complete processing cycle of a single article isn't very apparent or obvious to an average user. Over time, topics tend to disappear entirely from the process. The cost of processing topics follows a bell-shaped curve. The peak of the cost curve depends on factors such as meticulousness, text size, workload, factual richness of the text, and more. The full processing time can extend over a lifetime, but it can also be completed in just days or weeks, especially for high-priority articles, shorter articles, sparser articles, or easier articles. Once topics vanish from the process, the cost of reviewing items declines exponentially. In other words, the cost of knowledge acquisition in incremental reading, for a single text, is a combination of a bell-shaped curve and exponential curves, with a temporary increase in workload followed by a gradual dissipation of costs over time.
My little experiment aimed to plot the cost curve in terms of processing time. Naturally, the superposition of topics and items was somewhat artificial because I focused on speed-reading and topic processing, while the review of items occurred at arbitrary intervals that didn't stick to the optimum spaced repetition schedule.
Having completed the entire text and having deleted all topics, in November 2014, I decided to focus on proper item repetitions. I decided to report on the future progress in a decade (2023).
In the meantime, I asked Matt how much he remembered from the article we read together 8 years ago. I wasn't surprised when he had difficulties recalling the subject matter. However, he did remember reading about American History! This demonstrates that we can recall past events as long as they were prominent enough. Perhaps Matt also remembers how impressed we all were with his performance?
RESULTS
Size of texts
In incremental reading, the total size of texts in topics expresses the reading that still needs to be done for full comprehension. The total size of texts in items expresses the total size of knowledge acquired for life. Those two numbers keep changing in the learning process where items constantly feed on and consume the texts included in topics.
The total size of texts in topics will keep increasing for a while. The increase is caused by retaining context carried by some paragraphs. It may take 3-6 reviews for that increase to plateau. After that, there is a rapid decline in the size of topics as they get converted to items and gradually deleted (with Done in incremental reading). The size of items keeps increasing as long as new items are generated. After that, the size of items will tend to shrink due to the simplification process in re-formulating items and ditching redundant information.
Cost of reading and learning
In incremental reading, the cost of learning in time is initially dominated by reading, however, after a few reviews, the time devoted to answering questions carried by items becomes dominant. What cannot be seen in the graph, however, is the fact that this cost drops exponentially for well-formulated items. Once the article is fully processed, the cost of review will drop to negligible levels after just a few months. The total 8-years cost of review in the presented experiment amounted to 60 minutes. That's much more than Matt's 7 minutes at 200 wpm. However, if you compare recall-to-cost ratio, the superiority of incremental reading becomes obvious.
Element count
The number of topics will increase at first but will ultimately drop to zero. Items, on the other hand, are on an increase in the period of topic processing, however, their impact on the cost of review is mitigated by increasing review intervals in spaced repetition.
Incremental reading 17 years later (2023 update)
In May 2023, we met with Matt again after 17 years. I was to wait a decade before checking for memory retention. However, curiosity killed the cat. I had 58 items to check. I did not increase item priority because the original text was not as attractive as it seemed in 2006. All items have lingered at the bottom of the priority queue for 9 years since the last review. This made the test particularly interesting. We all know that spaced repetition provides for 96% retention by default.
In early June 2023, I retested my knowledge on items generated in speed-reading test. The estimated cost of the review was 5 minutes (excluding preps, sorting, calculations, etc.). The retention turned out to stand at amazing 87.9% (SuperMemo estimate 93.16%). After a 9-year break this sounds too good to be true. However, there is a lot of inter-item interference. This comes from the fact that as many sentences produced more than one cloze. The actual retrievability was estimated by SuperMemo to be 79.27% (i.e. 10 percentage points less). In addition, some of the knowledge might have leaked from the learning process in the field (sleep is central to my interests).
Even though, the high retention might seem to stand in contradiction to principles of spaced repetition, it is important to note that Algorithm SM-18 performed great showing R-Metric of 69%. This shows that for special cases like this, the new approach to computing optimum intervals is superior to the one used in Algorithm SM-15.
The statistics:
Elements: 116 Items: 58 Topics: 58 Memorized: 58 Dismissed: 58 Topic/task text size: 0 bytes Item text size: 12.84 kB Memorized items: 58 Average retrievability: 79.27% Average stability: 21 years 3 months 5 days (7768 days) Total consolidation: 978 item-years (see: How much knowledge can human brain hold) Total knowledge (in items): 45.98 Requested forgetting index: 10% Measured forgetting index: 13.36% First repetition f. index: 39.66% Last repetition f. index: 12.07% Retention: 93.16%
The limits of speed-reading
My primary focus is on reading for creative purposes, where the goal is not simply to read fiction quickly without missing the plot. Instead, I am interested in understanding the bottlenecks that arise in the creative process when individuals use reading as a source of information or creative inspiration to generate value, such as new ideas or long-term knowledge. From this perspective, the speed of reading becomes less significant compared to the true bottlenecks in the creative reading process, which are (1) processing information, and (2) retaining knowledge in the long term.
In creative reading, when you come across an insightful piece of information, it is crucial not to rush ahead and overload your mind with more information that might disrupt the newly acquired inspiration. On the contrary, allowing your mind to wander is key to fostering creativity. Multiple subconscious pathways are activated in the brain, and given the freedom to roam, these pathways can bring forth new ideas that have the potential to change the world. The speed of this subconscious creative information processing is largely beyond our rational control. Attempting to rush it could have a similar effect as imposing a tight deadline on a chess player facing a complex move. Trying to absorb more information in such a moment is like injecting more air into clouds that are just forming a meaningful structure in the sky. The inflow of information needs to be carefully managed—more speed when the mind is blank, and less speed when it is in a fertile moment. More speed when dealing with low-quality texts, and less speed when engaging with high-quality texts. For more information, refer to "The 1000-Word Dash."
The second bottleneck is the long-term retention of information, which requires spaced repetition and is best achieved by applying the minimum information principle. This means taking an attentive and meticulous approach to formulating knowledge for active recall, which significantly reduces the overall cost of retention throughout one's lifetime.
In speed-reading, there is an ongoing trade-off between speed, comprehension, and retention. In incremental reading, however, it is possible to deliberately control all aspects of the creative reading process: speed, comprehension, creativity, and long-term retention. Therefore, I must reiterate the original claim:
Conclusions
- incremental reading can be substantially faster than speed-reading, however, for an inexperienced reader, this is largely accomplished with skimming at the cost of comprehension and recall
- despite the high speed of reading, over time, the comprehension can quickly be brought up to 100% (or more precisely, to the maximum, which may depend on student's knowledge and article's quality)
- in the long run, the retention of knowledge will approach 96% by default (different levels can be programmed in SuperMemo in cost-vs-retention trade-off). This 96% level can be maintained for lifetime at low review cost
- the time cost of incremental reading increases initially due to the proliferation of topic extracts and due to the process of generating new cloze deletions. The cost curve over time is a superposition of a bell-shaped curve (topic proliferation), linear inflow of new items, and exponential decline in the cost of item review
- once all topics are fully processed, topic costs drop to zero. Depending on student's strategy, this may happen in just one review, or never. In the presented example, costs of topic processing dropped to zero after 7th pass through the text
- in the presented example, the cost of topic processing and the size of topic texts peaked in 5th review (15 min and 18 kB respectively, up from the original reading time of 2.5 min in the first review for 9 kB text)
- once topics are processed, cost of item review becomes negligible within a few months
- speed-reading training is a great introduction to fast reading with incremental reading, however, long-term incremental reading should also help develop natural speed-reading skills
- the lifetime cost of incremental reading may be ten times the cost of speed-reading, but the dismal recall record of speed-reading makes the difference in actual knowledge vast and growing with each month
- incremental reading may be difficult to master, however, it should be a dream tool for speed-readers who care about long-term retention of knowledge
Further reading
Comments
If you disagree, Vent on Facebook