I built N-Back.net, a lightweight and responsive Dual N-Back task for working memory training.
I was frustrated by the clunky UI and heavy design of most existing n-back tools, so I made something clean, fast, and distraction-free. No login, no ads, no tracking – just cognitive training in your browser.
Key features:
• Dual mode: visual + auditory
• Adaptive difficulty: level up as you improve, drop down if needed
• Dark mode and keyboard-only mode
• Clearly differentiated letter sounds, carefully selected to avoid confusion
• Performance tracker (WIP: I’m working on streaks and detailed history)
• Mobile-friendly (works well on phones & tablets)
I’m planning to add:
• Custom training modes (single-modality, color, etc.)
• User-defined N range
• Optional accounts for progress saving (but always optional)
Would love your feedback on:
• UX flow and performance
• Cognitive challenge balance
• What would make you come back and use it regularly?
the buttons jumping up and down when correct or incorrect is jarring on my smaller screen, forcing a constant space there or moving the message would be a smoother UX, in my opinion.
My personal experience with n-back mainly revolves around discovering the large variation in my ability to focus. Practisibg does get me better at the n-back task (not anything else as far as I can tell) but daily form determines results to a large degree. This effect seems to transfer somewhat: on weeks I'm overall worse at focusing on n-back I appear to also produce fewer commits.
But what determines this daily form I don't know. I suspect sleep, stress, and exercise has something to do with it.
I've been using the n-back protocol for over a year and it's definitely improved my cognitive abilities. I don't know if my IQ has increased, but on days where I use this protocol(20-25 minutes) I see a marked improvement in the clarity and vibrancy of my thoughts- put another way this protocol's biggest benefit is it's effectiveness in clearing the mental fog in my head. I've tried experimenting with other protocols(solving crossword/chess puzzles, meditation,etc..)but they were not nearly as effective as n-back.
I've actually built my own variation of n-back(triple n-back), where you're tested on the colour of the stimulus in addition to the sound and position-so it's much more cognitively demanding. If you're interested, you can check it out here: https://mind-workout.pages.dev
Generalization beyond n-back tasks had not been established, at least at the last time I looked into the field, maybe 8 years ago. There was always the hope that tasks that focused on specific core cognitive processes would lead to broad cognitive improvements, but I am not too confident about it.
One problem in this field has been weak control conditions..e.g. no train conditions. I often thought that a cardio control conditions would be useful. I'd almost guarantee that for most people, 60 minutes a week jogging would lead to better cognitive improvements than 60 minutes of n-back.
Yes, though 70% is a normal cut-off, I think most versions more heavily bias the placement towards 1/2 in the past square instead of the 1/9th of real chance. Without the bias it is simpler to always guess no.
Ahh, "brain training" - games you play instead of getting better at value added tasks. This shit doesn't generalize and the opportunity cost results in making you dumber, not smarter.
> Performance on dual n-back has complicated correlations with performance on other tests of working memory or IQ, so it’s not clear what it is tapping into. (And the link between WM and performance on IQ tests has been disputed; high WM as measured by OSPAN does not correlate well with performance on hard Raven’s questions and the validity of single tests of WM training has been questioned .)
> Unfortunately, in general, IQ/g and memory don’t seem to be trainable. Many apparent effects are swamped by exercise or nutrition or by simple practice. And when practice does result in gains on tasks or expensive games, said benefits often do not transfer .
> Have I seen any benefits yet? Not really. Thus far it’s like meditation: I haven’t seen any specific improvements, but it’s been interesting just to explore concentration - I’ve learned that my ability to focus is much less than I thought it was!
It does however seem like maybe dual n-back trains concentration and focus and willpower?
> WM training helps alcoholics reduce their consumption and increases patience in recovering stimulant addicts (cocaine & methamphetamine). WM training has been shown to help children with ADHD and also preschoolers without ADHD; Lucas found behavior improvements at a summer camp.
> To those whose time is limited: you may wish to stop reading here. If you seek to improve your life, and want the greatest ‘bang for the buck’, you are well-advised to look elsewhere.
> Meditation, for example, is easier, faster, and ultra-portable. Typing training will directly improve your facility with a computer, a valuable skill for this modern world. Spaced repetition memorization techniques offer unparalleled advantages to students. [Lots more.] And all of these can start paying off immediately.
> DNB, on the other hand, requires a minimum of 15 hours before one can expect genuine somatic improvements. The task itself is unproven - the Jaeggi studies are suggestive, not definitive.
For people with brain damage it might be very very useful. At least dual-n-back.io gave me a way to train spatial and temporal memory when I had no options.
Hi HN,
I built N-Back.net, a lightweight and responsive Dual N-Back task for working memory training.
I was frustrated by the clunky UI and heavy design of most existing n-back tools, so I made something clean, fast, and distraction-free. No login, no ads, no tracking – just cognitive training in your browser.
Key features: • Dual mode: visual + auditory • Adaptive difficulty: level up as you improve, drop down if needed • Dark mode and keyboard-only mode • Clearly differentiated letter sounds, carefully selected to avoid confusion • Performance tracker (WIP: I’m working on streaks and detailed history) • Mobile-friendly (works well on phones & tablets)
I’m planning to add: • Custom training modes (single-modality, color, etc.) • User-defined N range • Optional accounts for progress saving (but always optional)
Would love your feedback on: • UX flow and performance • Cognitive challenge balance • What would make you come back and use it regularly?
Site: https://n-back.net
Thanks for checking it out!
"do nothing" is too slow, need another key to hit to skip.
I would expect a final reaction time of under 100-200ms ideally?
Yes, I would also remove the automatic progression entirely. This should be driven by user input only.
the buttons jumping up and down when correct or incorrect is jarring on my smaller screen, forcing a constant space there or moving the message would be a smoother UX, in my opinion.
Agreed, came here to say the same
This is pretty great, thank you!
I didn't know about these types of tasks (N-Back) and this was a nice introduction to it.
My personal experience with n-back mainly revolves around discovering the large variation in my ability to focus. Practisibg does get me better at the n-back task (not anything else as far as I can tell) but daily form determines results to a large degree. This effect seems to transfer somewhat: on weeks I'm overall worse at focusing on n-back I appear to also produce fewer commits.
But what determines this daily form I don't know. I suspect sleep, stress, and exercise has something to do with it.
I've been using the n-back protocol for over a year and it's definitely improved my cognitive abilities. I don't know if my IQ has increased, but on days where I use this protocol(20-25 minutes) I see a marked improvement in the clarity and vibrancy of my thoughts- put another way this protocol's biggest benefit is it's effectiveness in clearing the mental fog in my head. I've tried experimenting with other protocols(solving crossword/chess puzzles, meditation,etc..)but they were not nearly as effective as n-back.
I've actually built my own variation of n-back(triple n-back), where you're tested on the colour of the stimulus in addition to the sound and position-so it's much more cognitively demanding. If you're interested, you can check it out here: https://mind-workout.pages.dev
Generalization beyond n-back tasks had not been established, at least at the last time I looked into the field, maybe 8 years ago. There was always the hope that tasks that focused on specific core cognitive processes would lead to broad cognitive improvements, but I am not too confident about it.
One problem in this field has been weak control conditions..e.g. no train conditions. I often thought that a cardio control conditions would be useful. I'd almost guarantee that for most people, 60 minutes a week jogging would lead to better cognitive improvements than 60 minutes of n-back.
This is very nice! But I don’t quite get how 70% is considered good performance — I managed to outperform that by literally doing nothing:
Game Complete!
Accuracy: 78.6% Correct Responses: 11 / 14 Average Reaction Time: 0 ms
Yes, though 70% is a normal cut-off, I think most versions more heavily bias the placement towards 1/2 in the past square instead of the 1/9th of real chance. Without the bias it is simpler to always guess no.
I also wrote a minimal version for me, https://mostlymaths.net/nb/
Only works well on mobile portrait though, I designed it especifically for my phone. Don't use it much though.
" Works on All Devices" statement doesn't seem to be true - I cannot press "space key" on my iPad (Safari).
It’s hard to tell when a turn starts if the tile stays on the same square. Could you possibly add a quick fade animation to the tile?
"Working Memory Test" link returns 404
Sry, i am fixing it. I vibe-coded this website so there might be some bugs :<
Ahh, "brain training" - games you play instead of getting better at value added tasks. This shit doesn't generalize and the opportunity cost results in making you dumber, not smarter.
From https://gwern.net/dnb-faq:
> Performance on dual n-back has complicated correlations with performance on other tests of working memory or IQ, so it’s not clear what it is tapping into. (And the link between WM and performance on IQ tests has been disputed; high WM as measured by OSPAN does not correlate well with performance on hard Raven’s questions and the validity of single tests of WM training has been questioned .)
> Unfortunately, in general, IQ/g and memory don’t seem to be trainable. Many apparent effects are swamped by exercise or nutrition or by simple practice. And when practice does result in gains on tasks or expensive games, said benefits often do not transfer .
> Have I seen any benefits yet? Not really. Thus far it’s like meditation: I haven’t seen any specific improvements, but it’s been interesting just to explore concentration - I’ve learned that my ability to focus is much less than I thought it was!
It does however seem like maybe dual n-back trains concentration and focus and willpower?
> WM training helps alcoholics reduce their consumption and increases patience in recovering stimulant addicts (cocaine & methamphetamine). WM training has been shown to help children with ADHD and also preschoolers without ADHD; Lucas found behavior improvements at a summer camp.
But to GP's point:
> To those whose time is limited: you may wish to stop reading here. If you seek to improve your life, and want the greatest ‘bang for the buck’, you are well-advised to look elsewhere.
> Meditation, for example, is easier, faster, and ultra-portable. Typing training will directly improve your facility with a computer, a valuable skill for this modern world. Spaced repetition memorization techniques offer unparalleled advantages to students. [Lots more.] And all of these can start paying off immediately.
> DNB, on the other hand, requires a minimum of 15 hours before one can expect genuine somatic improvements. The task itself is unproven - the Jaeggi studies are suggestive, not definitive.
For people with brain damage it might be very very useful. At least dual-n-back.io gave me a way to train spatial and temporal memory when I had no options.
[dead]