Hacker News new | past | comments | ask | show | jobs | submit login
Study suggests that the “brain training” industry may be a placebo (arstechnica.com)
225 points by Aelinsaar on June 20, 2016 | hide | past | web | favorite | 156 comments

Maybe it's just me, but I've always thought that being "smart" has everything to do with having a wealth of information and being able to draw connections between various facts and ideas.

Repeatedly solving a simple puzzle only makes you good at repeatedly solving a simple puzzle.

> I've always thought that being "smart" has everything to do with having a wealth of information and being able to draw connections between various facts and ideas.

Yes. My pet theory is that the brain is just a huge pattern matching engine. Formal education is mostly about training the pattern recognizer with new patterns.

Whenever you talk to people that are "uneducated" (or whatever word you may want to choose for that) that struggle with basic concepts, try to probe deeper. You'll find that, in many cases, they are being overwhelmed with new concepts being presented simultaneously.

People with extensive formal education will have been exposed to innumerable problems and concepts by the time they reach adulthood. There's a wealth of information the brain can use to speed up comprehension.

This is a pet theory though, being reinforced over the years by my own cognitive biases :)

I second that. And one of the main reasons I daily visit HN is exactly that. It widens my horizons, presenting me with different point of views to various problems/issues. I literally think that coming here makes me a smarter person in the long run.

I also experience that line of thinking in learning math. Mathematics seem like the perfect example of pattern matching. You progressively learn to adapt to a certain kind of thinking by solving a multitude of different exercises that broaden your conceptual understanding of a field. At least that's how it works for me. You keep exercising and at some point you experience an epiphany and then everything becomes crystal clear.

I too continue to visit HN in part because its one of the rare places that feeds my smart side and inspires/stimulates/satisfies that aspect of my existence. At the same time I've repeatedly come close to abandoning HN entirely because of another aspect interwoven seemingly equally with the first thing, which might be described as a kind of smart-idiocy, different flavors of it, whether recursive irrelevant descents into subjective trivial minutia, or points-scoring one-upmanship in the field of PC-ness, the occasional pro-VC/pro-startup propaganda, or merely the blatant uncharitable assumptions about the intelligence/knowledge of the other party, or the increasing cases of "you want to learn more about 'foo', really? Gosh golly, let me Google that term for you, here are my top 5 results, and I suggest you learn about Wikipedia, Amazon, YouTube, Khan Academy, etc.", or the thinly-disguised self-promotional posts/comments (humble-bragging, etc).

Summary: there is gold here, utter gold. There is also shit, reeking shit. Make your choice, choose your focus, then tune your dial.

You're a clever boy seeing as how you've packed all that into 33 days.

People can (and do) change accounts for many reasons. It's also quite common for people to just read articles/comments without commenting themselves, therefore no login is required.

Did these possibilities not occur to you?

I know those possibilities exist. I wrote it under the assumption that the syngrog66 has another more longstanding username. I was trying to be funny by teasing him.

this sounds reasonable

I've packed it into multiple decades. there is something else I'd like to feel allowed to say, but, barring that, hoping that @dang instantly bans you for extreme personal rudeness. If he does not, yet another notch in what I see as the dark side or decline of HN.

I don't know how long hacker news is on the go, multiple decades surprises me. However you only registered 33 days ago, that was the point I was making. If you find that to be 'extreme personal rudeness' then you must get offended a lot.

> You're a clever boy

I await your ban from @dang. otherwise I will re-conclude that HN has too high a ratio anymore of assholes-to-signal.

I'm not precious, I don't mind you calling me an asshole.

your original comment to me was rude and creepy. therefore I stand by my initial reaction. especially considering the historical ecosystem of HN

however... I've looked at your follow-up comments since, and I've spelunked through your post history. Therefore I now believe you did truly make a mistake, and that you honestly did not intend to sound the way that you did. And you gave some reasonable responses. I'm willing to give you the benefit of the doubt.

I've learned to take an increasingly hard-line against assholes in real life. But I've also learned to try even harder to show empathy and see things from other people's shoes. Therefore I apologize if you felt I over-reacted. And I look forward to seeing awesome contributions from you here in the future. ;-)

Cheers, thanks for the kind words and for being big enough to come back like this. I apologise as well, my attempts at humour fall flat a lot of the time. A little of it goes a long way according to my wife. Best wishes to you.

ha! ok cool. I hear you.

> I literally think that coming here makes me a smarter person in the long run.

Good ole' placebo effect :-)

agreed that a large part of "smart" is having a big enough pattern-matching apparatus

think there is more to it than this, especially on the highest ends

think the next biggest impact comes from having a large solid internal mental model of the world, of cause-and-effect, of action-reaction, perhaps two or more separate models, one or more models are "the way it is supposed to work" and one or more models are "the way it might work"

think the next higher level of smart/genius is the purely creative/original/innovative/brainstorming/fount-based: ideas and urges and visions and impulses that arise from somewhere that a purely logical/controlling/categorical mindset can't nail down, and yet, it... happens. and yet we feel it, we experience it, we feel it, we can imagine it. whether it is real or fantastical, possible or impossible.

I'd argue that the very smartest of us all may somehow leverage from all three of these levels

What you described is definitely how I feel when I really get into the zone of something, whether it be coding something very specific or doing overall design/architecture.

As I get into it, it suddenly snaps into place, and I have a mental picture (which is more visceral than visual) of how all the moving pieces fit (or will fit) together, and can start adding on to it. With coding in particualr, usually my brain is going faster than my ability to type.

I absolutely love that feeling.

well said. me too.

> This is a pet theory though, being reinforced over the years by my own cognitive biases :)

Imo someone "smart" is someone who is smart enough to admit what they don't know.

It's a terrible and ironic thing, that you require intelligence to analyze your own thoughts and actions, to realize that human intelligence is still all about degrees of monkeys counting on our fingers.

yep, sort of a reverse dunning-krueger effect

or you know, just the normal

Intelligence is supposed to be a measure of intellectual capability, independent of knowledge. It's difficult to define clearly, but knowledge is definitely a different quality.

Of course "smart" is even less well-defined than intelligence, so it's not a very useful word for such a discussion.

Yet, most tests of intelligence seem to be influenced by knowledge.

I.e. if you have solved a lot of "guess the next shape" or "find the odd element" you are probably going to perform better because you know how the thing work.

Likewise, literature and basic math seem to be widely influenced by education, and don't measure intelligence.

Are there tests that really measure intelligence? And if there aren't, are the qualities really separated?

> I.e. if you have solved a lot of "guess the next shape" or "find the odd element" you are probably going to perform better because you know how the thing work.

I hear this repeated a lot. I happen to have worked with someone who was into this (former something in Mensa), and as far as I understood one of the hallmarks of a good IQ test is when previous experience have small influence on the testscore.

Some people tend to say that anyone can train for this. I guess if that was true, Mensa would have a lot more members.

Once in the distant, distant past, I had a contract job. Full-time, they paid me for 40 hours a week, but usually I'd have an hour or two of down time each day. Like, 2pm rolls around, and I've actually finished everything that needed to be done for the schedule, so hey, I've got some time.

I started taking online IQ tests. I looked around for the really hard ones, and found that they fall into a couple of categories. Some focused on obscure vocabulary, featuring lists of hundreds of words that I, a native speaker with a pretty good vocabulary, had never seen. Others tested spatial awareness, logic, whatever.

What I discovered, after taking a few of these tests (I skipped the vocab ones; what's the point of learnin' fancy words no one else knows?), is that you get better at them, and fast. I only took maybe 2 or 3 a day, a couple times a week, but it didn't take more than a week to start testing a lot higher.

It got kinda boring once I started seeing that kind of improvement, so I stopped. I'm curious what it'd be like to try again, though.

  >I started taking online IQ tests.
Not at all the same as a real IQ test. A real test will be much broader and (imo) more interesting

I'm not disagreeing, but is there any reason why an online test can't test the same things that a "real" test, tests for?


Clinical IQ tests are administered by psychologists and created by publishers. Pearsons WISC-IV Basic Kit costs $1,123.40.

It's a bit like brand-name drugs and generics, except that the generics aren't allowed to copy the exact chemical structure of the brand name drug in a similar way that research substances copy controlled substances with minor chemical alterations, with unknown effects.

Now these free online IQ tests could create their own IQ test, but that requires both expertise in creating IQ tests, and a large normative sample (WISC used a normative sample of 2,200 children between 6 and 16).

Furthermore, creating more accurate IQ tests does nothing for their bottom line, or may even hurt it as some of these tests are biased to give you higher scores, so that you may share them on social media.

Whilst you could invest in creating an accurate IQ test to be cheaply taken online, you'd have to fight the perception these less accurate tests have created that online IQ tests are about as accurate as horoscopes.

As an interesting note, pearson does offer computer based IQ tests under their Q interactive brand, however, they are not cheap.

I'd bet it's at least a little the same as an offline IQ test.

>I guess if that was true, Mensa would have a lot more members.

Being in Mensa doesn't really "get" you anything, so there's little incentive to train for an IQ test in the same way there are to use the same amount of time to train for other more materially valuable things, like a license or a skill.

Even if you have a number that they would accept you for, that doesn't even mean you would want to be in a back-patting social group.

You are conflating various things.

1) it is uncontroversial that training improves performance on traditional IQ tests. It is true that people aim to minimise this when designing tests. It is also true that so far, nobody has done a great job at that. It's worth noting that Mensa presents different tests in different parts of the world - which is a pretty big hint that they think previous experience/cultural background affects how you perform on the test.

2) Anyone can improve their running performance through training. Not everyone can be a world class marathon runner. The fact that you can improve your performance on an IQ test is not disproven by the claim that not everyone is capable of reaching Mensa levels on it.

> The fact that you can improve your performance on an IQ test is not disproven

You and others here seem to think I'm out to:

* disprove that IQ scores can be trained

* say something about or recommend Mensa (not a member, don't know if I would qualify, never took the official test)

while I only wanted to offer a counterpoint to the IQ-measurement-is-more-or-less-worthless school of thought.

> Anyone can improve their running performance through training. Not everyone can be a world class marathon runner.

Same goes for cognitive skills. I think this sums it up: there are significant differences in both physical and mental capabilities.

I have no opinion on what you were "out to" do, I was responding to your comment which seemed to be expressing disbelief in the idea that IQ tests can be trained for ("I hear this repeated a lot...I guess if that was true Mensa would have a lot more members")

Thanks for a good answer. Sorry to have been unclear.

My point was that to me it seems clear that IQ tests do measure something outside of previous exposure or training.

IIRC most people will not manage to get 125 (just a number) on a well-designed test even if they are allowed to prepare with the exact same type of questions or even get trained for them.

I personally have no interest in joining Mensa, whether I'm qualified or not. I have no idea how many people generally would like to join, but my impression is that not everybody wants to.

I view intelligence as like this: in training camps for professional American football, prospective players perform physical tests: 40 yard dash, bench press reps, etc., and scouts for pro teams take note, because the results of those tests are correlated with actual professional performance. However, sometimes the guy that had no impressive stats becomes a stand out player, and sometimes the guy with amazing stats never amounts to much. Richard Feynman had an IQ of 125. JD Salinger had an IQ of slightly above average, and the guy with the highest recorded IQ worked as a bouncer for 20 years and hasn't accomplished anything intellectually ground breaking or significant in life.

Maybe cognitive tests have some correlation with what psychometricians call 'g,' general intelligence, but you're either naive, conceited, misinformed, or just plain stupid to think IQ is a meaningful number within the scope of an individual. Judging someone's intelligence is probably more complex than someone's athletic ability, and it's already not hard to see how ridiculous we're becoming when we ask: who is the best athlete: Usain Bolt or Peyton Manning? It is nothing more than a trait of the human mind that as a product of us defining important concepts such as athleticism and intelligence, we then, of course, desire to place ordinal ranking to answer the next obvious question: what is the most athletic, the most intelligence -- it is in an attempt of the human mind's desire to manipulate the thoughts and emotions of ourselves and the people around us; it is not done to understand reality.

> Richard Feynman had an IQ of 125.

Likely mismeasured. He won the Putnam mathematics competition with a stunning margin without preparation. He reportedly had the highest scores ever on the math/physics graduate admission exams at Princeton. Both of these are reliable indicators of an IQ above 145, probably by a good bit. (And this is if you ignore the dozens/hundreds of theoretical physics professors who thought of him as the fastest thinker among them, and one of the deepest.)

Well, when we get obvious deviations between IQ and g (the thing IQ is stated in the academic literature as what it's trying to measure) some will say that the IQ test was administered incorrectly. Some will say that IQ tests aren't perfect.

IQ tests are probably both not perfect and administered incorrectly (often enough).

I think it's an almost meaningless number unless we're talking about very large groups of individuals, and there's a lot of evidence and theory to support that.

There are lots of smart people with no desire to join Mensa whatsoever.

> I guess if that was true, Mensa would have a lot more members.

Mensa accepts a number of tests which are somewhat correlated with IQ but not traditional IQ tests, and for which there is significant research showing training benefits (including a number of the standardized college, graduate, and professional school entrance exams.)

Sorry for posting it as a video link, but highly relevant:


Or in text form, I think your assuming that smart people want to put in the effort required to join Mensa...when in reality it would be filled with mostly insecure people who feel they have something to prove.

Most people, geniuses or not, don't even know of/care for Mensa.

I draw much from David Krakauer and J. Doyne Farmer, both of the Santa Fe Institute, and the latter's drawing on Muth (1980s).

Krakauer: Intelligence is Search. It's the ability to find a least-energy path or solution through an arbitrary n-dimensional space.

Knowledge is having a solid backing of (valid) data and models.

Ignorance is not having that data and model store. It's a curable state, though being primed with bad data and models is very damaging to effective search.

Krakauer gave a really good interview at Nautilus a year or so back on this, it's very highly recommended.


Farmer, quoting Muth, looks at a model of engineering in which engineers generate solutions at random but are very good at identifying the better solution. I'm thinking that there are emperical reasons to suspect that this may not be far from the truth, and methods such as A/B testing seem to explicitly embody the identification element of this.

You'll find this covered in many of Farmer's recent talks (past few years), Martin School at Oxford and Nanjing University in particular, at YouTube.


<3 Krakauer.

Like everyone in his field he has probably encountered Herbert Simon before. But as far as "Intelligence is Search" goes, Herb thought that "search" is a subset of problem solving, next to "reasoning" and "constraints". In the end they could all be considered as being unified under "traversal of state-space" for a given problem. Creativity or imagination can be thought of as generating useful representations and possibilities for states within that space.

Ooh! Simon's new to me (or at least I think he is): https://en.wikipedia.org/wiki/Herbert_A._Simon

Anything of his that you find particularly singular or encapsulating these ideas?

Also, it was just Simon's centenary.

Good brief interview with Simon: https://www.youtube.com/watch?v=ErnWbP_Wztk

I too am a camp follower following the bread crumb trail.

After looking at Herbert's wiki, I feel like I've been creeping around his ideas for much of my life without realizing he was the progenitor, to the point where a few comments ago I used a word he invented. Krakauer's ideas too really resonate, these guys are really onto something.

Tentatively dipping a toe into complex adaptive systems and the Sante Fe institute also keeps coming up time and time again, they seem to have captured everybody thinking about this in depth.

Does anybody have good leads on Krakauer and Simon's ideas? e.g. books for a dive into?

Hoping for a Adaptive Complexity for Dummies!

> After looking at Herbert's wiki, I feel like I've been creeping around his ideas for much of my life without realizing he was the progenitor

I had this exact feeling as well. Herb thought at the exact same level of abstraction that I like to think in, broadly construed as "models", not too rigorous to be mathematics, but not too loose to be raw observational psychology. Just right. Marvin Minksy is precisely the same but for AI systems rather than humans.

Krakauer's website is where I would read Krakauer: http://tuvalu.santafe.edu/~krakauer/Site/Publications.html

Scott E Page and his "Model Thinking" course on Coursera is in the same vein as these two. He's a complexity theorist that deals with proper representation of social networks and distributed decision-making. His books "The Difference" and "Complex Adaptive Systems" are academic but extremely readable, and don't harken after the same Cellular Automata and Chaos examples without at least operationalizing them into some application first.


I just tracked down a whole mess of Simon's publications through Worldcat yesterday after responding on this thread. Organizations and hist book on adminstration seem like the two main ones.

I'm not sure regarding Krakauer, though I'm pretty positive he's published. SFI are _quite_ good about posting papers and such online, something I discovered after scanning in two books on a 30s/page scanner.... (I've since found faster scanners to use...) Though the published papers don't have the full formulae and typesetting of the published work.

Thanks, I'll check those out. Krakauer has a book on protocells, looks like a general overview of new ideas about self starting life plus computational ideas since Miller-Urey experiments. My trepidation is that I don't know if I'd require a PhD in biological science to understand it. I'm reading through some of his articles at the moment and they're all interesting.

> Though the published papers don't have the full formulae and typesetting of the published work.

Maybe email the author.

My POV has been that most of most of our conscious analysis boils down to utilizing a smaller set of basic cognitive tools. (Much like complex software runs on top of hardware with relatively simple interfaces.) If you can isolate and train core cognitive functions then that benefit will hopefully improve everything that utilizes it.

So for a random example: if you can go from comfortably holding 5 random objects in your working memory to 8, that's going to manifest itself in a variety of ways.

I wonder if people get different results from these puzzles based on how they approach the puzzle. Maybe they can solve the puzzle by avoiding the cognitive function the puzzle was designed to use and thus they get no/limited benefit.

As a counter argument to this, for example when I play chess sometimes I feel that the 'holding of objects in working memory,' the "playing card castle" practice of thinking ahead that I build up in my head, has some practical use in the real world, but if I'm honest about it I feel the extremity and specificity of the scenario of chess is not transferable to the real world.

I think the argument that chess has improved my underlying hardware sounds reasonable, but it's also necessary that hardware be able to adapt to different scenarios in life. If I spend too much time on chess, it feels like that would end in a place where that hardware would be specifically tailored to just playing chess.

I think that your comment undervalues the "thinking ahead" aspect, while overvaluing the "extremity and specificity" of what chess can teach.

I'm not a chess player, but I recognize that it can be a very valuable tool for understanding what I might call the "while loop" of chess

* any given position has multiple options for action

* how to anticipate what moves an opponent may make in response to each of those options

* how you will, in turn, respond to the opponents responses

IOW, it teaches tactics and strategy, both of which we engage in countless times each day, whether or not we consciously realize it.

The strategy and also the heuristics that it teaches you are valuable, I can certainly agree with that. When you start playing a lot of chess, well at least I find that the incremental improvements you can make, after a while of playing, becomes a very narrow set of problems you are solving: mainly, how much can you "hold in your head" in terms of thinking ahead. That involves how well you can visualize the board.

I just found at a point getting better meant memorizing a lot of opening game theory, as well as memorizing end game tactics. Then, I'd probably still never be world class because I don't think I'm exceptional at visualizing chess piece movements--I may be better than 99% of all people but the way the top 1% of chess players can do it is just phenomenal. I feel like it's a specific skill that some people just have and get extremely good at. You really need to spend some time playing the game and getting better and better to realize how good at visualizing the ability to think ahead the top players are. Magnus Carlsen never studied chess using a board -- it is just natural for him to think about.

Yeah, I think that what a lot of people miss in this conversation is that there really is this thing that strongly correlates better performance on IQ tests with better performance in all sorts of other otherwise-unrelated things.

And at the same time, training for better performance on IQ test questions does NOT increase performance on those otherwise-unrelated things.

So it's that thing that correlates across those unrelated things. That's what the intelligence researchers define as intelligence, even though it's popular to redefine it as something else that is entirely nurture and not nature.

There's a little bit more to it, however, since working memory is often a limitation for a wide range of tasks, and anything that could increase your short term buffer to hold a few more items concurrently could help across a wide range of tasks, and potentially yield more "fluid intelligence." There are a number of conflicting findings about whether you can actually increase this, but the general concept is not quite as trivial as overtraining on puzzles as you laid out.

There was a controversial study done in the 60s when we could still do these types of studies where they told teachers who the smart and dumb students were, the Pygmalion Effect.[0] The study suggests being "smart" has more to do with other people's perception and expectation.

[0] https://en.wikipedia.org/wiki/Pygmalion_effect#Rosenthal.E2....

The problem is, it is hard to find distinction between the word intelligence and knowledge. However, there is one metric we can use to define the difference between the two: Given two entities with equal knowledge the one who is more intelligent can perform better on an Intelligent Quotient test (aka IQ test).

In order to perform such a test, however, we will need to find two people with equal knowledge. To find two such people I propose we use a Knowledge Quotient test (aka KQ test)

The problem is, it is hard to find distinction between the word knowledge and intelligence. However, there is one metric we can use to define the difference between the two: Given two entities with equal intelligence the one who is more knowledgeable can perform better on an Knowledge Quotient test (aka KQ test).

In order to perform such a test, however, we will need to find two people with equal intelligence. To find two such people I propose we use a Intelligent Quotient test (aka IQ test).

I've spent sometime studying this. I personally believe Lumosity and the like are a waste of time (if not a downright scam) and do not make anyone smarter. Until I see scientific data that proves otherwise this will remain my belief. However that doesn't mean I don't believe in brain training

By your definition, smart is knowledge AND the ability to draw connection between facts and ideas. And although solving simple puzzles may not train you for the latter, continuously solving mathematical problems (that may otherwise not have any real world applications) will train your brain to find patterns.

If someone can make Calculus, Graph Theory and Discrete Mathematics fun enough for people to tackle daily, you'll have a real brain exercise platform.

Here [0] is an interesting article that talks about both sides of this argument. It has examples of studies that have shown real life advantages to these types of brain training. Some examples include increased scores on standardized tests in teenagers and better memory recall in adults 65+

[0]: http://www.nytimes.com/2012/04/22/magazine/can-you-make-your...

I like the analogy that the brain is a muscle and if you don't exercise it, parts of it start to atrophy.

I think these brain games actually bring a lot of value to people who have retired and are no longer required to exercise their brains in order to fulfill their basic requirements of life.

Maybe they don't make you smarter, but perhaps they help prevent you from getting stupid.

Good points, would this work with other puzzle types that do not involve math? Is there a threshold of difficulty for these puzzles to become useful?

I would agree, in broad strokes. One of my gifts has always been a near-photographic visual memory (I can remember images of certain pages of books I read decades ago), and my brain's propensity for collection and efficient indexing of random bits of trivia. Mention some historical subject, and I almost instantly pull up everything I've ever read about that subject, as well as secondary and tertiary subjects that are related...

Repeatedly solving a simple puzzle only makes you good at repeatedly solving a simple puzzle.


a.k.a programming interviews

If most people solve a problem by going from: A -> B -> C

'Smart' people will consistently go straight to: A -> C

It different than just having access to information, speed and how it's processed makes a significant difference.

> If most people solve a problem by going from: A -> B -> C

> 'Smart' people will consistently go straight to: A -> C

Cutting out intermediate steps in a chain of analysis is more typical of people with applicable, domain-specific expertise than general smarts. Being smart isn't the same as being an expert.

General smarts is reflected more in a reduction in the average number of dead-end side paths pursued before making the connection from A -> B and again before making the one from B -> C.

I was good at math growing up but not in a "#1 in the world" way. I've been to a couple of math contest national finals. So I naturally have seen many "geniuses". Some of them when they see a problem sometimes just come up with the answer immediately. These are questions that require procedural thinking so I couldn't understand how it was possible to come up with the correct answer instantly. One day I asked him. He said he doesn't know but sometimes he can just "see" the answer. I used to think that's bullshit, but after having experienced more of life I think it is not entirely impossible.

We have no idea how brain works, so we can only make hypotheses based on experience. When someone can "see" an answer without being able to explain how he came up with that answer, regardless of how useful that skill is in the real world, that is something. Maybe sometimes all his neurons connect instantly in certain circumstances. Who knows.

Another reason I think this is possible is because sometimes I come off that way (to much lesser degree) to other people. Sometimes when I come up with a solution to a problem (not at all domain specific), people wonder how I can do that. I can't explain of course. I am not trying to say I'm a genius. I'm saying things like this happen sometimes for no reason. And "highly intelligent" people tend to have more of these than "normal" people.

Anyway, these are all just theories, yours included. So I think just saying that our brains objectively work in certain ways is not a good idea, and is probably wrong.

Im not talking about being an expert, Im talking about intelligence. Intelligent people tend to 'see the forest through the trees' on a grander scale than average or unintelligent people (you can tell Im less than intelligent by my over use of bad analogies) . They don't need to go A -> B because they already know that A -> C is correct. It's a lousy example because everyone here is looking at it like its a procedural problem being solved and it's not. It's about individuals having greater cognitive ability.

I know what you are talking about intelligence, I'm disagreeing with your description and saying it doesn't capture the advantage intelligence provides, but more accurately captured what expertise provides. Now, I'm not clear whether it's just a bad articulation of what you are trying to illustrate or if your underlying concept is wrong, but your description doesn't, IMO, accurately portray the difference that intelligence creates.

Its the most generic explanation I can use in the shortest amount of text that generally gets the point across. It is a lousy example. It's obvious when people are smart/intelligent and also obvious when the are not, I believe its largely genetic and out of people's control so discussing it or creating websites to better your 'IQ' are mostly a waste of time.

I don't think it's obvious at all, and I think people often mistake a number of other things in others (philosophical agreement, expertise in a domain, certain social skills, among others) for general intelligence.

There's also considerable evidence, IIRC, that, while the upper limit a person can reach in intelligence is genetic, that environmental factors play significant roles in both how closely that limit will ever be approached and how well/long what is achieved will be maintained.

Well we'll just agree to disagree on this one :) At the end of the day though it doesn't matter if either one of us is correct, nothing really changes either way.

This is called intuition and not analysis.

I don't think that is the case. Trained people will go A -> C. Also careless people will go A -> C, because it is easier to make a mistake when skipping the partial steps.

I read a book where they talked about a Boeing (I think) pilot training exercise. They constructed checklists for new pilots of what to do in various situations, gathered from their best pilots (likely not exclusively their best, my memory is fallible). These lists worked very well - but they noticed their best pilots didn't exactly follow the list.

The book I was reading made a point that early learning on a topic is rote (You do this then this because you are supposed to), while advanced practitioners _intuit_ . "I'll check this because it feels like that kind of problem, and if that isn't it I'll hit the other more likely things in turn or question my assumptions, but I don't think those steps out explicitly, I just do them". That certainly matches my programming experience, though my anecdote does not data make.

IIRC, this book was focused on how to improve skills, and was making the point that you can't expect to jump straight to intuit without experience, you have do rote first. So in this example, "smart" people will go straight from A->C, people with only moderate experience will have to go A->B->C, people with enough experience will go from A->C, and careless people MIGHT end up A->C, but more likely never made it to B and so can't even get to C from where they are.

Thus the appeal of "brain training" - people want to up their ability to intuit with less experience, and they feel that training can do it like most skills. This study (I've not yet read article) seems to be saying "no, it doesn't work that way".

Assuming the headline and study are correct, it's shame but I don't feel people should be mocked for feeling it was worth a try. There are certainly a number of other studies that show forced mental effort helps avoid degradation at a minimum. (I recall a study about Bridge delaying Alzheimers as an example, though I can't say anything about how well that study has held up).

This is how aeroplanes fall out of the sky. Pilots think they know more than the engineers, and decide to skip the checklist, where the checklist includes a very rare edge case that the pilot has never encountered.

Reminds me of this case where airlines routinely skipped a step during maintenance because it took too much time. Skipping that step in the manufacturer's process ended up costing 273 people their lives.


Are you referring to space fortress [1]?

[1] https://www.brainturk.com/space-fortress

And this is exactly why "smart" people are counterproductive. Their "smart" solutions are not reproducible.

A smart person who is also pragmatic and takes the team into account though is superproductive. Their solutions may take a bit of learning but it pays off.

> Their solutions may take a bit of learning but it pays off.

Solution that takes a bit of learning is always worse than a solution that does not need any learning and is obvious to anyone, even to someone in my IQ range.

Not always. Otherwise we'd all be using punchcards and valves still.

I used punchcards. It definitely required more learning than a video terminal.

That's in running for the dumbest thing I've ever heard.

Try to look at a smart-ass solution and compare it to a good, solid, dumb solution. The former is A->"some-obscure-magic"->D, while the latter would be like "A->B->C->D" with each step explained in an ELI5 style. No matter what, I'd always prefer the second option.

You're not talking about smart people, you're talking about "smart" people, who are clearly not that smart if they're doing such obviously stupid things.

I'm talking about the people with high IQ. Most of them are guilty of the same "from A straight to Z" mistake, unfortunately.

Are you a python programmer? :)

My coworker is, and he bangs the same drum. I agree with him on it, mostly!

Actually, no. I believe Python is exactly the opposite thing, despite preaching some of the right ideas. Python is too low level, and therefore it invites obscure solutions that hides the real meaning behind layers upon layers of low level, irrelevant details. Exactly a tool for the "smart" people, pretty much useless for a reproducible and maintainable development.

> therefore it invites obscure solutions that hides the real meaning behind layers upon layers of low level, irrelevant details

And thats exactly why I said "mostly agree" with my coworker! This is my exact gripe. Basically its the same idea, taken to an extreme to the point where it no longer matches the original reason.

There is nothing wrong with the idea per se. It's just implemented badly in Python. Instead of allowing unlimited abstraction, which is the only way to unlock this desired simplicity on all possible levels, they're limiting the abstraction and therefore increasing the complexity needlessly.

Or smart people might decide the problem is unimportant and neglect solving it.

Even if you can't "train your brain" what you can train is your ability to focus at a task. For example, if you play chess you probably do not "get smarter." All you get is better at visualizing future board positions and maybe you learn some general strategy and opening theory. But 99% of chess is raw calculation.

Most people when they begin playing can't stand games longer than fifteen minutes or so. They have not trained their brains to focus over extended periods of times. But the more you play, the better you become at sitting still and concentrating on the position on the board. Eventually after years of practice, you'll be able to play games where both players have 120 minutes+increment and use your time without getting bored.

Replace chess with any other brain teaser, like a quiz or an iq test. If you can sit and stare at the same problem uninterrupted for hours trying to work out a solution, then you will condition your brain to be able to do just that.

I firmly believe that is useful in all parts of life. Random example, you are thinking about divorcing your wife. Now if you are able to spend several hours analyzing the possible ramifications of that decision you will be better of than if you can only think about it for twenty minutes without losing your train of thought.

So I don't think it is all placebo. Ability to concentrate and perseverance through difficulties are two very important skills to have.

It's not only chess, meditation, for example, works on the same principles.

I've read of some people who were finally set free after a long stretch in solitary confinement, and many of them attributed the preservation of their sanity to simple math or word exercises. I believe one would just count exponents of two as far as he could go and another would try remembering every word in every language they knew.

Not sure of similarities to brain training games, but I would bet there is something there.

You just gave me an interesting idea. Manually writing down my vocabulary (grouped into passive and active) for every language I know. Counting up the results and comparing would be rather interesting.

After that diffing it with a dictionary file for each language to see the gaps.

This will probably take forever, but it could be worth the time.

This study argues that people recruited to use Brain Training score higher on fluid intelligence than people recruited to participate in a study for a reward. That is, Brain Training works for people who believe it works.

So, shouldn't it be used by people who believe it works?

The self selected group may have exerted more effort in the second round of testing because of their belief they were in a study looking for certain results.

It's clear, studies should recruit participants using neutral language like the right poster of the article to remove as much selection bias as possible.

>So, shouldn't it be used by people who believe it works?

You can't ignore opportunity cost when evaluating value/ROI. It's not merely about the absolute benefit any given method/material provides, but about how that compares to whatever else might have been done instead. This stuff is very much not free (hence being a "billion dollar industry"), either in terms of money or time, so the only way it "should be used" is if there was no superior (cheaper, faster or both) way to induce the equivalent placebo effect, and no other way whatsoever to produce an actual effect beyond placebo. That seems highly doubtful, given proven benefits from other forms of actual study. With the time/money spent on this sort of pure-placebo "Brain Training" someone could have worked on learning another language for example, or invested in other valuable training.

The same argument you propose here is commonly applied to other highly profitable placebos pushed to the public, and in general the flaw in reasoning is always the same one of ignoring comparative ROI. Probably the most classic example would be homeopathy, wherein people literally just buy bottled water for hundreds to thousands of dollars per gallon. By definition there should be no negative effects, because there is nothing there but water (perhaps with a few flavors or non-bioactive aromatics), so worst case there is no harm, best case somebody experiences a bit of placebo benefit. That leads to "well what's the problem then, it can't hurt." But people don't have unlimited resources, any spending on pricey water directly subtracts from what they might have spent more productively elsewhere. Worse, choosing a non-medicine placebo may result in them failing to use actual medicine that was appropriate.

So it does matter when something like this is pure placebo. Perhaps most humanly, it matters because it's not what people were being advertised, signing up or working for. "We lied to them because it was good for them" does not cut it.

This is generally true for placebos in the psychological domain. Often times a positive outlook or expectation of improvement can achieve a somewhat better outcome. (One example among many: acupuncture in the domain of pain management.[1]) It is more a question of how much of an improvement one should expect, and whether placebo-sellers are exploiting their customers be setting false expectations.

[1] https://www.sciencebasedmedicine.org/acupuncture-doesnt-work...

I think that's a reasonable question. The answer is that there are much cheaper ways to obtain the placebo effect here than to give these companies your money.

I also guess without immediate proof that the placebo effect is likely to be temporary and represent no real, persistent change to intelligence at all.... another reason to not give them your money. There are cheaper and much more entertaining ways to obtain "no change to your intelligence".

> The answer is that there are much cheaper ways to obtain the placebo effect here than to give these companies your money.

There might be cheaper ways, but paying money has an impact too. That's why more expensive painkillers often work better than generic ones, despite having identical ingredients [1].

[1] https://www.sciencedaily.com/releases/2008/03/080304173339.h...

We don't know if it was selection bias or the placebo effect, or a combination of both.

It would be pretty simple to design another trial to test this.

> If it is selection bias then the training has no effect whatsoever.

That doesn't follow. Both factors could have an effect at the same time and both effects could, in principle, go in either direction. The only thing that we can conclude from the present study is that the joint effect of placebo and selection bias is positive.

Yes. I just meant that the entire effect could have been due to selection bias, so we don't know if there was any placebo effect at all.

What really matters is cross-over. Does the skill you are training have application to what is being simulated? It's kind of like you get what you measure, in that you get what you train. For example, through training head-to-head in math based games where speed is important, I have noticed that I am much more able to do quick calculations on checks more easily. There is a game in Elevate where you are shown avatars of people then told information about them and their names. You have to remember what they look like and basic information about them through various rounds and comparisons to the other players. I have noticed that when I play that game frequently, it helps with memorizing names of people. Dual-n-back is said to help increase working memory and I believe it because in that game you are having to constantly keep in your head the last auditory and visual records at the same time and integrate new information with them. I am annoyed by people constantly saying this type of training does not work...

At least to me, it seems like doing repetitive puzzles isn't likely to do much more than make you good at that type of puzzle, rather than building actual problem-solving skills. I used to play these games with pretty little actual impact outside of my score...

In math and logic, I've found a pretty great resource in https://brilliant.org/ with many quizzes and wikis that focus on the underlying connections and general problem-solving approaches, which has made me better at solving actual problems.

I seem to remember studies from a few years back showing Lumosity as less effective than World of Warcraft at "brain training". At least WoW can engage a few social and competitive aspects rather than just mindless puzzle solving.

CBC's Marketplace also did a show about this, pretty much crapping on the evidence for the whole thing:


"Dual N-Back" remains one of the best - it may be the only one that really works, unlike the bulk of other brain games that truly are scams.

This one has extensive research literature, and many free/ad-supported and cheap ones on android and iPhone app-stores.

For a critical discussion of the research see: https://www.quora.com/How-does-dual-n-back-actually-increase...

It seems to be one of the best if you interested in this sort of thing.

Claims include: improved working memory, improved executive functioning (great for ADHD/ADD), improved emotional control, self control and regulation, minor bump in IQ.

To anyone interested, Gwern also did some extensive research on Dual N-Back.

Gwern’s DNB FAQ[0] offers some favorable opinion on DNB (in that it’s fascinating, though much larger gains can be achieved via many other means—meditation, environmental changes, etc.—before DNB can start to give any meaningful effect). Later meta-analysis[1] links post-DNB IQ gains to methodological problems in studies, with quite small remaining positive effect that may be linked to further biases.

(Note that currently the FAQ is marked “unlikely” and the meta-analysis is marked “highly likely”.)

The results are overall not so rosy and I personally am under the impression that DNB is of little utility, if not a waste of time or highly depending on luck[2].

Some of the links from https://www.quora.com/How-does-dual-n-back-actually-increase..., however, lead to studies supporting DNB’s reputation as a great way to increase anyone’s IQ. I don’t know if newer studies avoid the biases of the ones Gwern looked at.

[0] https://www.gwern.net/DNB%20meta-analysis

[1] https://www.gwern.net/DNB%20FAQ#n-back-in-general

[2] There are many options on how to do DNB exercise, including test presentation, features, training length, and so on.

nice analysis. My experience with DNB is that it definitely slightly improved my working memory after quite a lot of practice. One side-effect I noticed was an improved ability to track people's arguments and notice inconsistencies. However, in order to get to say 4-back took a lot of practice and it only worked when I was well-rested and nourished. When I was tired I simply could not improve. And DNB is ever so boring.

Could you share how you practice DNB exercise? What software do you use, presentation options chosen (if any), anything else.

i just used an iphone app (IQBoost I think..it's been a while) & did the sound+vision version with headphones, practicing in groups of 3 with a break inbetween until & got bored

Can someone explain the downvote? Please don't downvote without explanation, i've got no idea if this persons claims are suspicious or false or etc.

edit: Also, to this thread's OP; Doesn't your link's accepted answer claim n-back doesn't work? There is critical discussion there sure, but the consensus seems to be that it doesn't work. No?

> Please don't downvote without explanation,

Downvotes exist to improve the signal to noise ratio on HN: please do, generally, downvote without explanation. If something is productive and worth discussing, that's a sign that its not worth a downvote.

Hmm, i guess i just don't understand the downvote though. Ie, clearly the person who downvoted knows something i don't. I'd like to understand that!

It seems the person downvoting would rather me not view it (as it's not "constructive"), but i can view it.. and i am, yet how am i to know why the downvote was given? Is the author lying? Is it a spam link? Is the author wrong? etcetc

Thx Fizzbatter, I indeed have no horse in this game (I'm not involved in the industry) - I've seen all of the mainstream "brain games" be thoroughly discredited.

Dual N-Back is the only one that has at least SOME indication that it may have actual benefits, and it's not owned by one of these big brain game companies - so you can at least be more impartial and less subject to hype.

You're right, I deliberately pointed everyone to a skeptical discussion of the research.

Brain games as a whole ain't great, but of a bad bunch, this is the best, both in terms of research behind it, and in terms of cost (not some monthly subscription - often a free download) and impartiality.

Yeah, the human body seems to be very good at doing the minimum required to meet the provided training stimulus. Ride a bike fast, you get better at riding a bike fast.

I started doing crossword puzzles a few months ago. Mainly I've gotten better at doing crossword puzzles. When I started, I had to think about each clue, but now I know the common words that always show up and don't even have to read the clues to get those. Meanwhile, as I never write about "MDs" or the "EPA", this has had no useful impact on my life.

Right, which is why training specificity is a thing. You can train different than you play/compete, but you have to target very carefully, especially as you progress.

The only real "brain training" is reading, learning practical and technical skills, and perhaps living a healthy lifestyle.

Other than that, forget it. Brain power and being a quick-witted genius is all about GENETICS (like most things in this life)

Couldn't disagree more. I guess intelligence might be an exception but most outcomes in life are not about genetics but are about circumstance, passion, and hard work.

That's a rather strawman response. You say you "couldn't disagree more" then state a position that the person you're replying to never stated they hold (about life outcomes).

They were talking about cognitive decline and "brain power," not life outcomes as you seem to think.

(like most things in this life) doesn't imply to you they are extending this POV beyond the discussion of brain power?

Passion. Hard work. Two personality-traits largely influenced by genetics. Try again.

I'm interested to see if you have any links to studies/research that back this up.

The reason I ask is that I personally find I am passionate/hardworking for certain things but for others I'm completely disinterested/lazy.

Likewise I have seen other people who are typically lazy and not hardworking, but get them on the one thing they are interested in (e.g. a specific sport) and they can quote match and player statistics stretching back decades. It's like they have this amazing ability but they've only applied it to a relatively useless part of their life (I say relatively useless because if they'd applied themselves with the same fervour to their work they'd have a much better quality of life).

I'm not sure how that connects that with genetic personality traits.

Here is one:


If you want more:


Scientists are still primitive in the understanding of the human brain, but as you can see, one factor in your behavior and personality (ie. motivation, persistence, ease-of-anger, compulsion, etc) is the complex relationship between neurotransmitters and their receptors (and millions of other factors of which we have no clue) These relationships are largely genetic, they are inherited to us.

I have something to add of my own: Every characteristic about us, every move or action that we make, has a relationship (or is a byproduct) of some sort of gene(s) working together. You can even call it FATE if you want.

> You can even call it FATE if you want.


I don't know if it's brain training, or just feeding an addiction, but playing music seems to make my brain work better. Maybe it functions by distracting me from other stuff that I'm over-thinking or worrying about.

I won't discount genetics.

That's basically saying it's all environment, except for the things that aren't.

I've had sharpness and focus improvements taking energy metabolism supplements. Could be that I'm deficient here or that it could also help others.

CoQ10, Nicotinamide Riboside, L-Carnitine, Magnesium Malate, B complex, N-Acetyl Cysteine

In terms of just being smarter and quicker to learn I'm not really sure that seems to be a combination of experience and natural factors. The experience part you can obviously change but the other one who knows.

I think in general if a supplement has a positive result on health/cognition, that means you were deficient. CoQ10, for example, is produced by your body. You would think then, it would produce as much as you need.

Obligatory link to gwern's experimentation and research on nootropics:


Are you actually taking Nicotinamide Riboside long term? 500$ for a yearly supply is a bit steep and that's assuming 300mg/day which is not ideal.

This research article and the arstechnica commentary are examples of exactly the overreach that they aim to criticize.

A number of large-scale multi-site randomized controlled trials have shown that specific types of brain training generalize to untrained measures of cognitive function and real-world activities. Here's two examples:

ACTIVE: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3934012/

IMPACT: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4169294/

Some brain training works, and some doesn't. Of the ones that work, some show certain types of benefits; and others show other types of benefits. Throwing the whole field out is like saying "laetrile doesn't treat cancer, so no molecules treat cancer."

disclosure: I was an investigator on IMPACT; and I work at a brain training company (Posit Science)

Well this doesn't really come as a huge surprise. these brain training apps seem to target logical repetitive tasks that would only train your brain to be better at doing just that - logical repetitive tasks. I do however think it is very possible to train the brain, it just hasn't been put in the form of an app or autonomous computer software yet.

An interesting thing about placebos is that all placebos are not equal, which to me calls into question what a placebos really is.

Unlike "brain training", good old "learning" and "education" are definitely _not_ placebos.

Learning a language might increase your overall intelligence, but mostly you "just" learn a language - like solving puzzles makes you better at solving these specific puzzles.

"brain training" industry may be a placebo but brain training is real. I have yet to understand why people think that you can have a very good brain if you don't train. In sports it's a given: if you want to run fast, you have to train, you have to eat right, you have to sleep rigth, you have to take breaks to not destroy your body.

But somehow people that work with their brain think that the brain, the organ, does not work like the rest of the body and that you are smart no matter what an so on. You are smart if you trained, if you sleep right, if you eat right and in general if you take care of yourself.

People say "it's just a puzzle". It's like a football player would say "why do I have to do squats?! I never have to do one while I play". Sure, at some point the puzzle becomes too easy. Thats why they do squats with weights.

Keeping your brain fit is important, and the study isn't arguing that. The way that these games do it is not useful.

Just liking training running will make you good at running, training those puzzles will make you good at... those puzzles. That's not what those game companies sell it as, people want better general intelligence and instead they are just getting really good at matching shapes or solving crosswords.

In your football example, a squat is full body exercise that strengthens numerous different muscle groups. These "brain training games" are not full mind exercises.

I thought it was well-proven a year or two ago that these apps/offers/classes were a scam?

It was. And at any rate claiming your app is scientifically proven or whatever was a lie from the beginning.

Chewing gum can also increase intelligence. So it's easy to see that the effect of games can be hard to isolate from other factors.


This has been both obvious and revalidated numerous times.

I'd like to know why public broadcasters, particularly PBS and NPR in the US, aren't called to task for advertising this crap heavily during their sponsoship drives.

Associating yourself with crap spreads the stains.

Isn't the point of brain training sort of to intentionally induce placebo?

No. A placebo effect is fleeting but training is supposed to have long-term effects.

If a placebo is an effect that requires no training, being able to induce it would seem superior to training.

Studies show that surgeons who play first person shooter video games are faster and better with laproscopic and robotic procedures. Does that not count as brain training?

Seriously, all training is brain training, if it isn't weightlifting. What else are you gonna train?

Exercise is great for cognitive development/maintenance and weight training in particular involves more math than an outsider may realize so even that is brain training.

Brain suggests that the "study" industry might be a placebo.

I find N Back training to be actually effective from personal experience.

The intelligence is all in my mind?


Looks like you've mastered the art of having no thoughts, and with so many words even.

The funny part is why does the placebo work? I know, there you go again....

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact