I Was Wrong About Speed Reading (2015)

I took a speed reading course one summer when I was in college. My mother thought that it would give me an edge. First they test you and my reading speed was already on the high end of normal.

In the course I did theoretically more than double my reading speed. But my comprehension was based on the test they gave me. Plus I noticed that I tired quickly doing speed reading. I was a kid who loved to read and had no problem doing it for hours on end without tiring.

After the course I quickly discovered as I suspected that my real world comprehension suffered greatly by speed reading.

So what value did I receive from the course? After college there have been times I needed to quickly get the gist of a document and was under extreme time pressure. I found speed reading worked fine though in my business career I doubt that I used it more than once a year.

Bottom line my personal ROI would have been much higher if my Mother have hired a tutor to help me over the summer learn to paint or play blues harmonica instead.

I'm what I've come to call a 'slow thinker'. My perception speed is normal, but the gray matter ticks over at a very slow rate of knots and my comprehension of what I'm seeing/hearing/feeling unpacks over the course of several seconds. I've noticed I often have an initial reaction to things, then a 'slow reaction' that is sometimes quite different and factors in more things I know.

It's like it takes some time to 'dial in' on what my response to a particular stimuli is going to be, but often the situation demands a response faster than I can provide one that I feel is properly considered. This leads to interesting social stuff sometimes, and sometimes a kind of short-duration l'esprit de l'escalier social anxiety.

As I've got older, I've learned that in a lot of cases the initial reaction is fine, and I can suppress the desire to issue a follow-up correction once the brain has caught up. Younger me was less able to do that and would often blurt out a second response to nobody's particular benefit, or hold off on a response to the point the moment is lost.

So perhaps for this reason, speed reading has never really clicked for me to the point that I've always been 'suspicious' of it. It's as the author says, comprehension is often the bottleneck. Even with fiction, I find I'm relating the current occurrences I'm reading with whatever world the author has built up, and considering ramifications.

My SO reads very quickly, and I often wonder how much of it is being deeply considered as she makes her way through something. It makes me hesitant to recommend any text I've enjoyed for the richness of it, as I fear a superficial read somehow does it less justice than just not reading it at all.

Reading speed strikes me similarly to typing speed when it comes to programming. It's fun to read/type fast, but it's not really the bottleneck.

On “subvocalization” ...

> It’s simply not possible to comprehend what you’re reading and avoid using that inner voice. So reading faster means being able to use this inner voice faster, not eliminating it. To further that, expert speed readers who were studied also subvocalized, they just did it faster.

I’ve seen this argued elsewhere, and am not convinced they’ve found readers who do not subvocalize to study. Most readers subvocalize. Even if you don’t, if you ask about it, then you do (like asking someone to not think about their breathing).

If you do not normally subvocalize, you become painfully aware of it when you drop into it, feels like slamming on the brakes, and very difficult to let go of again.

I suspect the studies referenced[1] have simply not studied those who don’t subvocalize, akin to studies missing the phenomenon of aphantasia until 2015. Or, they’ve established experimental conditions where the reader cannot help but think about it, causing it to occur.

1. https://skeptoid.com/episodes/4229

Testing the notion that subvocalization slows you down should be reasonably easy. Just test people deaf since birth. Do they read faster? Visit some deaf communities.

It's not immediately obvious, but many deaf people don't know how to read at all. Sign language (hand gestures) is a completely different language, different from English, Russian, Italian etc. To complicate stuff, there's not a single global sign language.

I practice what could be called "speed listening" -- listening to audiobooks and podcasts (and watching videos) at 2x to 3x speed.

It saves a lot of time, and I only wish I'd started sooner.

I was obsessed with speed reading at one point and still practice some of the techniques. I can attest to the author's recommendation to skim before reading. Especially with challenging material, I find skimming first reduces anxiety-induced procrastination. The material becomes less intimidating once you've "gotten to the last page," even if fraudulently. Then you can go back and read, perhaps selectively and non-linearly, until you understand it.

Speed reading attracted me because I easily get bored and drift off. If the pace is faster, I can stay engaged better. It's not really about speed or saving time (the irony of that while posting on HN).

I can understand a lot of Japanese words written in kanjis without knowing how to pronounce them. So I am basically forced to read without any subvocalization and it’s pretty tough to understand any text longer than 5 lines.

It feels like trying to understand a mathematical proof that has been written only with mathematical symbols.

So I'm scrolling through this article, reading more carefully at first to model his thought processes, then picking up a few words per paragraph, formulating a hypothesis as to what he's saying in that paragraph, making a spot check to confirm my hypothesis, and moving on. I was oblivious to the irony here, as I was focused on reading. When I realized I was a counterexample to his thesis, I cracked up and had to stop.

The recommendations for more effective reading largely match those in Mortimer Adler's How to Read a Book, particularly chapters 2 (The Second Level of Reading: Inspectional Reading) and 7 (How to X-Ray a Book). Both address how to rapidly determine a book's structure. More generally, Adler suggests approaching reading with intent, that is, what are your goals in reading a book, often corresponding to the type of book. Fiction is different from poetry, practical books, history, science and mathematics, philosophy, and social science (each addressed in the book).

I'm reminded of an HN perennial, "Why GNU grep is Fast" (https://lists.freebsd.org/pipermail/freebsd-current/2010-Aug..., discussion https://news.ycombinator.com/item?id=1626305), which boils down to two rules:

#1 trick: GNU grep is fast because it AVOIDS LOOKING AT EVERY INPUT BYTE.

#2 trick: GNU grep is fast because it EXECUTES VERY FEW INSTRUCTIONS FOR EACH BYTE that it does look at.

If you want to read faster, read less, and read more intelligently.

By "read less", I mean simply don't waste your time on irrelevant material. For pleasure reading, this doesn't mean skipping large parts of books, in general (there ... are exceptions), but it could very well mean skipping entire books, or deciding that you really don't want to read them. For nonfiction, it may very well mean sucking the marrow out of their bones without getting bogged into the detail. There are few writers who are highly synoptic and for whom all words matter (this is more common in philosophy than elsewhere, though I can think of other examples).

By "more intelligently", I mean understanding what you hope to gain (amusement/entertainment, general information, specific answers to questions, understanding or rebutting an author's arguments, etc.). Skip what doesn't address that goal, and stop once you've attained it.

There is a tremendous amount of published material. There have been about 300,000 traditionally published books annually in English since the 1950s, if not before, and over the past decade or so, "non-traditional" ("vanity", self-published, print-on-demand, ebooks, etc.) raise that number well over 1 million. Given roughly 300 million to 1 billion native English speakers, this means one book per 300 to 1,000 people.

Quality literature is far less common. Many authors have commented on the folly of reading other-than-excellent books. I'd suggest some sampling of the common genres, but ... don't get stuck there, if you can help it.

Shopenhaer addresses this in "On Authorship and Style":

Writing for money and preservation of copyright are, at bottom, the ruin of literature. It is only the man who writes absolutely for the sake of the subject that writes anything worth writing. What an inestimable advantage it would be, if, in every branch of literature, there existed only a few but excellent books! This can never come to pass so long as money is to be made by writing.

https://www.gutenberg.org/files/11945/11945-h/11945-h.htm#li...

Diderot the broader problem of a superfluity of books and information overload:

As long as the centuries continue to unfold, the number of books will grow continually, and one can predict that a time will come when it will be almost as difficult to learn anything from books as from the direct study of the whole universe. It will be almost as convenient to search for some bit of truth concealed in nature as it will be to find it hidden away in an immense multitude of bound volumes.

https://www.historyofinformation.com/detail.php?entryid=2877

(This was, of course, a solutions-oriented pitch for a technical IT solution Diderot was launching, the Encyclpaedia.)

A depressing tendency of most speed-reading programmes, and discussions, I've encountered, is that few show any wareness of these extant models and guides. Which makes one wonder if their authors actually read intelligently....

Reading with intent gets me more of what I want than any of the speed reading techniques I've tried or practiced. Business non-fiction books in particular tend to do a good summary in the first chapter, and spend the rest of the book expanding on the main concepts. I'll read the first chapter and call it good if I've got it, or listen on audible at 1.8x if I want more reinforcement.

Several times people have told me I’m an exceptionally fast reader. I notice if I’m in a group and we’re asked to read a sheet of paper (eg some training seminar) I’ll usually be done way ahead of everyone else.

I vocalise everything. It’s usually a lot faster than you’d say it out loud though, probably if I tried to read out loud that fast I’d trip over words.

The secret is rather boring: I’ve just read a lot. I rarely have problems recognising words, so that’s key. For many types of reading I also infer content a lot, and it seems like my brain completes sentences, then I’ll continue on and “jump back” if one of the words was not as expected. I very easily for those tricks where people write “gondolas in Vienna”, so my brain is obviously doing a lot of inference.

I’ll often read the economist cover to cover, and I used to do the same with newspapers where I was young. Maybe reading every single word on such a variety of topics fixes the “surprised by words” category.

My tip is: read thoroughly and carefully and vocalise, read enough and the speed will take care of itself.

Do people using languages like Chinese, Japanese or Korean read faster? In Chinese and Japanese a single character often means a word. In Korean, a single character is 3 letters. It's very simple: words use much less space in languages like that, so you need to move your eyes much less. Would a Chinese or Korean person read the same work of literature faster than an European - assuming they know all the words?

If "visually condensed" languages like that do facilitate faster reading, converting to a different language is the way to go! What a lovely research topic for neuroscientists!

I think my reading efficacy can be improved significantly by not spending time reading junk and spending more time reading good stuff.

That is, I have a prioritization problem not a throughput problem. Most people I know are like this.

I sped-read this article so I'm _pretty_ sure it just confirmed what I already think: speed reading works!!

I have learned to read both slower and faster. The trick is that you can use both in the same session, and for many materials, this is by far the best approach.

My comprehension went up massively when I started reading out loud in the evenings, despite how childish it feels sometimes. I noticed I got a lot more out of audio than text (I tend to read too fast and not let things sink in), so I started recording my own audio for on the go.

I realized while doing this that reading out loud works great for me because it forces each word to sink in, rather than rushing ahead as was my habit.

Another thing I learned from listening to audio is, I come back to it many times, so I end up "reading" the same books sometimes dozens of times. And this is precisely what is at the core of many speed reading systems: multiple exposure.

Scott dismisses "read as fast as you flip the pages" courses. It's important to clarify this. This is actually his first tip later in the article, "Skim before you read!" The difference is, beyond just a skim and a read phase, there is an extra "photo-read" phase added on the front.

Have you ever opened a page, and a word or passage jumps out at you, visually? The word "sex" for example, has just such an effect. The idea here is you must read with purpose, with intention, with questions, and if you do, then while you are scanning the book with your eyes, you are building a map of the material, so when you come back to it later, you will remember where to look to find what you need.

In many cases, a book is large and we are only interested in a small fraction of the material. This is the case for most textbooks and their exams, or for a personal project only some aspects will be necessary to extract. The multi-pass approach to reading allows you to locate everything you need, learn where it is, ask questions and keep coming back to it as needed.

Reading slower to build comprehension when you need it can integrate perfectly well with this nonlinear approach: you find what you need faster, and then you give it all your concentration, making sure you understand each part before continuing.

(Of course, this will give rise to yet further questions you can explore the rest of the text with! :)

A thread from 2016: https://news.ycombinator.com/item?id=12023791

The practice of speed reading is not the same as the ability to read quickly. I don't speed read, I read very quickly and often at the rate of page at a glance (I often tell people not to leave confidential papers exposed so I won't accidentally read something I shouldn't be privy to). When people ask me how they can learn to read more quickly, I ask them if they would consider reading more.

Speed reading works for low-information content. For instance, try reading that article aloud to your kids. Nobody will ask you to slow down.

My personal experience with speed reading is that the limiting bottleneck factor for me is the speed at which i can internalize complex grammatical structures.

I also sub-vocalize some words but not others. In fact, the claim that one must sub-vocalize every word is contradicted by my experience reading Chinese characters, sometimes I can understand characters without knowing how to say them.

This article kinda confirms what I've felt when I read about speed reading: if you're reading anything other than fiction/romance, comprehension is the bottleneck.

So reading half a page of a math book at the time would have been completely useless (I was in university at the time).

In a similar vein: the millennial fascination with speed listening: consuming podcasts and audiobooks at 1.5 to 3x speed. The point about comprehension is key: if you complete Altlas Shugged in 8 hours but don’t understand the motivation of the characters, what was the point?

About three years ago a buddy of mine suggested that listening to course lectures or general podcasts/audiobooks/Youtube tutorials saved him a massive amount of time counted in hours by simply putting the playback speed of everything he listened to as high as he could muster. At least 1.5x for Youtube videos, often time he listens to podcasts and books in 2x. I argued that this is tantamount to speed reading and that he’s losing half the comprehension anyway, may as well slow it. I’ve tried it and believe it’s about as successful as speed reading.

My speed reading method is just to skip every other paragraph. It's really surprising how effective this is for a lot of texts, and you can adjust how much you skip as you go, dependent on the text.

Not familiar with the author, but good on him for admitting he was wrong. I find that most people are not able to admit their faults (without endless hedging), and those who can deserve some respect.

Mind map the book starting with the conclusion then toc.

If a book takes more than 4 hours to mind map it's either a good book or poorly organised.

The reading speed isn't the constraint. It's placing the knowledge in your knowledge tree without tiring yourself out. Only read the key parts of the knowledge tree slowly in depth.

Use A3 paper and a click pencil. Speeds the reading right up. When the books mapped, explain the mind map to someone.

speed reading?

for me, reading a book is a treat. Something I rearly manage to make time for, and when I do, I'd rather enjoy every paragraph.

You don't chew a great food as fast as you do with simple one, right?

The full article claims that it's impossible to reduce saccades by reading multiple lines simultaneously.

I know that claim is false because I have experienced reading an entire book by perceiving two to four lines simultaneously.

I've only had that happen once, on incredibly easy double-spaced material, but humans are capable of it.

Note that I am a natural speed reader and aphantasiac - those may be contributing factors (I've noticed the most talented visualizers I know are all slow readers, and discussion leads me too believe it's partly because they can't help stopping to visualize, while I cannot stop to visualize).

> This was masked because the books I was reading had enough redundancy to make following along possible with impaired comprehension.

Clever trick speed reading book writers!

Usually, in reaching "genius levels" of knowledge and understanding, textual input is not the bottleneck.

So, don't bother with Speed Reading.

When I don’t subvocalize my brain basically turns off and I stop understanding what I’m reading.

Hyperlexia is real and quite nice. Sadly nothing you can do to develop it.

I could do 20,000 wpm after reading something the first time in 200wpm.

I tried the Spritz link up to 700 wpm, after closing the page my eyes felt weird for a while. This may be due to the black background and the fact that I was focusing a lot on the text, but in any case I would not recommend it...

I've been reading (and writing) considerable amounts, pretty much every day all my life since I was 12-ish (and ~15), 37 as I speak. I'm cognitively no genius by any metric but certainly 'well capable' relatively to most people (strong points being speed of thought, clarity/precision to a fault, powerful synthesis). I however have a bad factual memory in general, I need to 'do' or 'understand' things to really memorize (e.g. I suck at poems, lyrics, can't even remember mine, but I love systems, physics, logic).

Without explanations or hazardous 'hypotheses' (hum) as to why, here are my observations and personal 'tricks'. Spoiler: no magic, just directed effort, solving for X.

- There is an inverse relationship between speed and depth of information input. The faster you take it in, even if you feel like you understand all of it (and may, factually), the shallower your actual intake will be in the long run. Generally. Emotions act as a catalyzer for memory, but cloud thinking temporarily.

- The more massive a piece of content (book, article..), the slower I want to go if I aim at a constant result (comprehension, food for thought, memory, etc). I may speed up when cruising if the content lends itself to that (repetitive, very structured, etc). Essentially, whenever recent memory won't cut it to think and remember, because there's too much content, or it's too complex, I need my long-term memory sparked into activation.

- I read fast in general, but it has nothing to do with 'techniques' related to the act of reading itself: it's entirely a mental 'mode' or 'state', that I only ever get otherwise in sports, this moment when you just trigger the beast and commit entirely to victory. When you just won't fail, when it's 'all-in'. That level of focus, of concentration, is best exercised with training, of any kind. So indeed you need to think, not just read, and actually want to even write and use the substance if you are to maximize its results on you.

- It's very important for me to begin slow, whatever the average pace later on. I need to get acquainted with the author, their go-to vocabulary and grammar subsets, the point they're making. Sometimes it never clicks and the whole piece is tedious for me. Sometimes it's so clear I feel like I could have written it myself, at least the structure of it. It's a thing, hard to explain. I guess why all we have our own favorites and dreaded authors.

I like people who speak relatively fast because I don't get easily distracted or bored when they talk to me, so speeding up audiobooks is a personal treat; but there's a limit if I want to retain anything more than fuzzy snapshots of partial points. I may understand 100% on the spot (I actually do nothing else but listen and focus), but it's evading memory, evaporating in the long run. Ultimately it means that book won't change me beyond a few impressions, it just won't make its mark, it never had a chance to imbue my consciousness — only my superficial cognition. However sharp, it can't substitute for the real thing.

An exercise I do if I really want to use or remember (be 'changed') by some piece of content — article, book, chapter, quote, podcast, piece of code... — is to do the "good student duty" of writing a short essay: summarize what you read, raise a few interesting points about it, elaborate a bit beyond (make that new material become food for your thought, use it like it's a new word you just learned, or a new math technique).

It only takes ~30 minutes or so, easily 10 minutes if the material is easy, an hour if it blew your mind, I don't know... well worth it, because it has a lasting power for years. But do it well, because that exercise you 'submit' to no one but yourself will become part of the memory of that content, you will see your own words sometimes before/above the actual author's.

Hopefully this helps someone trying it my way. The hard way. The no-genius no-magic no-trick way. The way that lets one play in the present with stuff learned 10, 20 years before.

TL;DR: the faster you read, the shallower it gets beyond some sweet spot, probably biological but trainable; speed is good when it forces you to focus, but not beyond that — it must not become a distraction from actual understanding. If that is your goal. Using at least in thought what you read is my preferred way to really internalize something, make it mine, thus remember it for long enough usually. At least I probably won't be able to forget that it exists, and then I can search it for finer points.

I have some problems with this article because of personal experience which differs. They talk about the constraints of 'working memory' which makes it unlikely, or even impossible to scan more than a few sentences at once and comprehend them. It just so happens that i sometimes scan a whole page at once, just by looking at it, and have understood it. But almost only on paper, and without formulas or diagrams in it. But not every time, or 'at will'. It just happens.

Not in the sense that i memorized it, and could repeat it word for word, but the gist of it, the point, message, whatever.

It even varies, because sometimes, when i need to recheck some datum mentioned like a point in time, or an amount of something, or a cross reference, i recall exactly where on the page that was, and where that page was. This happens less on screens of any sort, but more so with real paper. Makes no difference if fiction, or non-fiction.

I guess it happens from anywhere from 0,3 to 0,5% of everything i read, it would be up to 5% if i'd only read on real paper. So much for the 'haptics' of 'cyber'...

Anyways, i could read before Kindergarden, because my older Sister tought me, just for fun. I remember the first word i could read was the neon signage of some supermarket called 'Hit', because i asked what it meant. That's how it began.

At the same time this introduced me to the concepts of different languages, different spellings for the same words, and different words for the same things in different languages. Because i'm german, and hit is english.

Early school was often boring torture for me because of that. I could read books and newspapers, and my classmates stuttered along excrutiatingly slow. This lasted maybe to the 6th grade, then they finally caught up(mostly).

Also around that time i finally realized that i had this ability, and it was something special. Interestingly again not in my native language, but in english.

Teacher said go to page xyz and read this, say ready when you are done. I did, and maybe 2 secands later said: 'Ready!' (Was something about the double decker buses in London)

I never read that before (at least not conscious) because everything in that book was boring to me, and of course got reprimanded, which i considered as unfair.

It's like having a photocopy before the 'inner eye' which slowly fades, but not exactly like a photocopy, because then i could read from it and repeat it letter for letter, and word for word, which i can't.

So i'm wondering what do they mean when they talk about 'speed reading', does this include the necessity of memorizing everything, being able to do it anytime, or just the 'lossy' compression of meaning like i (sometimes) do?

In case of the latter, then from my point of view that is like blind people (from birth on) talking about color.

some edits: spelling, structure

I read extremely slowly compared to most, but seem to come up with more insights as well as debunkings.

A relative of mine is a natural born speed reader. She can read several 500+ page dense works a day. This is suggested to be impossible. I've personally tested her on her recall after having her quickly digest a work she's never seen before. She has perfect recall and can tell me what page a given passage appeared on, and recall from memory what comes next.

It's a savant ability. Probably less than 1 in a million, or perhaps billion.

She can't do anything with the ability. It's useless to her practically. She works minimum wage jobs, is married to an unemployed guy, and is always poor. She doesn't have any marketable talents. She continues to read huge amounts of material when she can find it. She long ago read all of the books in her small town library's fiction section. Discussing a topic she can cite references to it as if she is a search engine, but like the search engine, she doesn't have any insights into what any of it means.

I've suggested to her she try to market the ability, do a tour and show off. But she hates attention and craves nothing more than a quite reclusive life free from other people. Being publicly acknowledged as a freak would be the worst possible outcome for her, as far as she is concerned.

I've never understood speed reading. I've never believed in the claims because there's just no way someone is comprehending by doubling their speed.

"I Was wrong About Snake Oil"

Wrong guy is wrong.

It is possible to speed-read TFA with no loss of comprehension, because it is information-light. Most things available to read are, and when you are obliged to look at them, often a glance at each paragraph tells you all you need to know.

Things that are really worth reading are information-dense. Then, the time to think about what is presented dominates, and the time to take in words is negligible. Sub-vocalizing can aid comprehension by slowing you down, and by providing additional information via the rhythm of the presentation.

The last book I read by Robert Sapolsky took me 30-60 minutes for each of the last 30 pages, because they were tying together all the material in the rest of the book, and really gave full value for those minutes.

So, speed read what you must read, but choose material that delivers for careful study, and read it as slowly as it deserves.