Brain-training apps reliably make you better at brain-training apps. Mahjong makes you better at reading a situation with incomplete information while three other people watch you decide. Neither one is a proven cure for anything, but only one of them loads memory, attention and social judgement in the same two hours.
The transfer problem, in plain language
Cognitive scientists split the question into near transfer and far transfer, the distinction summarised in Wikipedia’s article on transfer of learning. Near transfer means getting better at a task very like the one you trained on; far transfer means the improvement spreads to unrelated abilities such as reasoning, planning or everyday memory.
Brain-training products are sold on far transfer and mostly deliver near transfer. Adrian Owen and colleagues put more than eleven thousand participants recruited through the BBC through six weeks of online training and reported in Nature in 2010 that the gains showed up on the trained tasks and did not carry across to untrained ones.
In 2014 the Stanford Center on Longevity and the Max Planck Institute for Human Development issued a consensus statement, signed by dozens of researchers, warning that the evidence did not support the industry’s marketing. The US Federal Trade Commission later announced that Lumos Labs would pay US$2 million to settle deceptive advertising charges over claims that Lumosity guarded against age-related decline.
None of that makes training worthless. It makes it narrow. A broad workout needs a task that refuses to sit still.
What a single mahjong hand asks of you
You are reasoning from what you cannot see
You can see your own hand and nothing else. Everything else must be inferred from the discard pond, from what opponents have claimed, and from what they have conspicuously stopped throwing.
No puzzle app has this property, because a puzzle shows you the whole problem. Mahjong hides most of it and then changes it while you think.
The difficulty comes from other people, not a curve
An app tunes its difficulty to keep you engaged. Three opponents tune their behaviour to beat you, and they adapt faster than any algorithm you have a subscription to.
There is no undo, and the mistake is public
Discard the wrong tile into a waiting hand and you pay for it in front of everyone. That consequence is what turns idle pattern-matching into genuine decision-making under risk.
Someone is expecting you on Thursday
The most underrated cognitive feature of mahjong is the obligation attached to it. Apps get abandoned quietly; a table of four notices an empty seat, and an activity you quietly drop in week two cannot do anything for you.
Side by side
| Dimension | Brain-training app | Mahjong |
|---|---|---|
| Information available | Complete and visible | Hidden, inferred from discards |
| Source of difficulty | Algorithmic curve | Three adapting opponents |
| Consequence of error | Retry instantly | Points, and a hand you cannot take back |
| Memory load | One drill at a time | Your hand, the pond and three opponents at once |
| Social contact | None, or a leaderboard | Three people for two hours |
| Emotional demand | Minimal | Composure after a collapsed hand |
| What it costs | A subscription that renews on its own | One set to own, or a hire for the night |
| Reason to continue | A progress graph you keep alone | Three people who noticed you missed last week |
Where the apps genuinely win
An honest comparison has to concede this much. An app is available at three in the morning, requires nobody else, and never has to be rescheduled around four calendars.
It also tracks progress in a way a mahjong night never will, works for people whose hearing or mobility makes a long session at a table difficult, and asks nothing of a room. If the alternative to fifteen minutes of drills is fifteen minutes of scrolling, the drills win.
The defensible claim is narrower and stronger than the marketing on either side: for loading memory, attention and social cognition at once, four people around a table do work that no single-player exercise is built to do.
What can and cannot be claimed for mahjong
Mahjong is unusual among leisure games in having been studied rather than merely praised. Sheung-Tak Cheng and colleagues in Hong Kong have run structured mahjong sessions with nursing-home residents living with dementia. Their exploratory study of 62 residents, published in the International Journal of Geriatric Psychiatry in 2006, reported gains across every cognitive measure it took, and their cluster-randomised trial of 110 residents in the American Journal of Geriatric Psychiatry in 2014 set mahjong against tai chi and a handicrafts control group, finding that the control group deteriorated while the mahjong and tai chi groups held their ground.
Two caveats belong in the same breath. That work was done with small numbers of nursing-home residents who already had dementia, which is a long way from a weekly social game among healthy adults, and no credible study shows that mahjong prevents dementia. Our detailed look at what mahjong actually does for memory sets out which systems the game loads, and our guide to mahjong as therapy for older players covers the practical side of running sessions.
How to get the benefit: a setup that survives contact with real life
- Fix the night before you fix anything else. Same evening, same room, every week or fortnight. Adherence is the whole ball game.
- Recruit six, not four. Two reserves is the difference between a standing game and three cancelled ones.
- Teach the shape, not the scoring. Show a beginner what a finished hand looks like, then send them into a real one. Our method for teaching mahjong in a single game sets out how to do that without a scoring lecture first.
- Stay on one variant for a month. Switching rulesets early destroys the pattern recognition you are trying to build.
- Phones in a bowl. The point of the exercise is undivided attention, which is the one thing a screen reliably removes.
- Add scoring in week three. Once the hand shapes are automatic, counting becomes the next layer of the workout rather than the barrier to entry.
Practise between sessions on a phone if you like; our rundown of mahjong apps covers the good ones. Volume online, depth at the table, is a sound division of labour.
What it costs to test the idea
A subscription renews whether you open it or not. A mahjong night is a one-off: in Sydney a tile set hire starts at A$20.00 and a full table with stools and tiles at A$70.00, priced by weekday or weekend and by how long you keep it.
That is cheap enough to run the experiment before committing to owning anything. We deliver across Greater Sydney, or you can collect from Brodie Spark Drive in Wolli Creek, and every set is cleaned before it goes back out.
Pick a Thursday, find three people, and put a real table in the room for one evening. You will know by the second hand whether the app was ever the better workout.
Questions worth answering before you cancel the subscription
Do brain-training apps work at all?
They work on the exercises themselves. Practise a digit-span drill daily and your digit span improves; what the evidence does not support is that the gain spreads to unrelated thinking. The US ACTIVE trial randomised 2,832 older adults to memory, reasoning or speed-of-processing training or to a no-contact control group, and the improvements it found were largely specific to the ability that had been trained.
Has mahjong been proven to prevent dementia?
No, and nobody should claim it. Sheung-Tak Cheng’s group in Hong Kong has reported cognitive benefits from structured mahjong sessions with nursing-home residents who already had dementia, which is a long way from showing that a social game prevents it. Treat mahjong as a good use of an evening, not as insurance.
How often do you need to play for it to matter?
No study establishes a dose, so take this as practical advice rather than a finding. A standing weekly or fortnightly table is far easier to sustain than an intensive month that collapses, and the version of this you actually keep doing is the only version that can matter.
Does playing mahjong on an app count?
Partly. A digital client gives you far more hands per hour, which is excellent for pattern recognition and tile efficiency. It removes the parts that load memory hardest, because the app sorts your tiles, flags legal calls and counts the score for you.
Is mahjong too hard to pick up later in life?
The winning shape can be explained in a sentence and shown in one hand. Scoring is the hard part, and it can be postponed. A patient teacher can have four complete beginners playing real hands inside a single evening.
What if I cannot get four players together?
Three-player variants exist and play faster, and a standing fortnightly game with a reserve list is more realistic than a weekly one that keeps collapsing. Failing that, play online during the week and keep the physical table for the social benefit that screens cannot supply.
Hosting soon? Rent a complete mahjong setup delivered across Sydney, or browse our original mahjong merchandise.
More in Lifestyle & Social