Every puzzle app implies the same promise: play this and your brain gets better. The research is less flattering and more interesting than either the promise or the backlash. Here is what has actually been found, with the studies attached.
The claim, and why it is hard to test
"Brain training works" can mean two very different things. Narrow transfer means you get better at the trained task — undeniable, and not very interesting. Far transfer means the gain shows up somewhere you never trained: your job, your memory for names, your ability to follow an argument. That is the claim worth testing, and it is the one that keeps failing.
Testing it properly needs an active control group. If your control does nothing while the training group plays a game, any improvement might be enthusiasm, attention or simply expecting to improve.
The big negative results
The largest early test was run with the BBC: Owen and colleagues, published in Nature in 2010, trained 11,430 people online for six weeks. Participants got better at the trained tasks. On untrained tests of reasoning and memory, the improvement was no greater than the control group's.
In 2014 a large group of scientists published a consensus statement objecting to the industry's marketing. A rival group answered with their own letter. Both are worth reading precisely because they disagree while looking at the same literature.
The most careful review of the whole field, Simons et al. (2016) in Psychological Science in the Public Interest, concluded that the evidence for improvements in everyday cognition is weak, while the evidence for getting better at the game itself is strong.
What the regulator said
This is not only an academic dispute. In 2016 the US Federal Trade Commission fined Lumosity's maker over its advertising; the FTC's own announcement says the company "preyed on consumers' fears about age-related cognitive decline" without the science to back the claims.
A useful filter: if a puzzle product talks about your brain age, dementia risk or IQ, it is selling something the evidence does not support.
What does hold up
Some findings survive scrutiny, and they are narrower than the advertising:
- Practice makes you good at the practised thing. Deduction is a skill: reading a clue as a filter, checking a name once, not re-scanning. That improves with reps, reliably and quickly.
- Speed-of-processing training has the strongest long-run data. The ACTIVE trial followed 2,832 older adults for ten years and found the trained groups reported less difficulty with daily tasks — the closest thing to far transfer anyone has produced.
- Physical exercise beats puzzles for general cognition. Reviews of aerobic exercise in older adults consistently outperform brain-training apps on the same outcomes.
- Engagement matters more than the specific game. Learning a demanding new skill — an instrument, a language, photography — looks better than repeating a trained task.
The honest case for a daily puzzle
So why play one at all? Because "it will not raise your IQ" is not the same as "it is worthless". A daily deduction puzzle is ten minutes of deliberate, unhurried attention in a day that offers very little of it. It has an end. You either found the answer or you did not, and no algorithm decides how you feel about that.
There is also a real, small skill being built: the habit of not guessing. Working from evidence to a conclusion, noticing when you have crossed something out on a hunch, checking before committing — that is the transferable part, and it does not need a neuroscience claim to be worth ten minutes.
The strong claim is that puzzles fix your brain. The defensible claim is that they are a good way to spend ten minutes and they teach you to stop guessing.
How to read a brain-training claim in ten seconds
- Look for an active control group. No control, no claim.
- Ask what was measured. "Improved on our own test" is narrow transfer wearing a lab coat.
- Check the sample size and who paid for the study.
- Be suspicious of any number attached to your brain: brain age, cognitive years, percentile IQ.
Applied here: this game claims nothing about your brain. It claims that every clue is true, that exactly one name survives, and that you never have to guess. Those are testable, and you can check them yourself in why crossing out works.
Related reading
- Why one puzzle a day beats a marathon — Spacing, streaks and the shared-puzzle effect: why ten minutes a day builds more skill and more enjoyment than a two-hour session once a fortnight.
- How to solve a deduction puzzle without guessing — A five-step method for elimination puzzles: read the clue that cuts deepest first, work name by name, slow down for counting, and use the guarantee that exactly one answer survives.
- The families of logic puzzle, and what each one trains — Zebra puzzles, logic grids, nonograms, Sudoku, Minesweeper and elimination puzzles share one skeleton. What separates them, which skill each one builds, and where to start.