There is no single number that is your attention span, and any site offering you one is guessing. What can be measured is how well you sustain attention over time, how well you select one thing from many, and how well you stop a response you have already started — three separate abilities that fail in different ways. All three are testable here in under ten minutes, and the results are more informative than a single score would have been.
You have probably seen the claim that human attention span has fallen to eight seconds, below that of a goldfish. It is repeated constantly and there is no credible source behind it. Journalists tracing it have never found the study it is attributed to, and the goldfish comparison appears to have no basis at all.
What is true is that people switch tasks far more often in an environment full of notifications, and that each switch costs real time. That is a genuine and well-documented finding about interruption costs. It is a different claim from 'attention span has shrunk', and the difference matters because the remedies are different.
An attention problem caused by interruptions is fixed by removing interruptions, not by training attention.
Sustained attention
Staying alert when nothing is happening. Radar Scan demonstrates the vigilance decrement — the well-documented decline in detection performance that begins within the first few minutes of a monitoring task, and the reason watch-keeping roles rotate people.
Watch the pattern across a run rather than the total. A steep decline in the second half is more informative than the overall number.
Selective attention
Picking one thing out of a busy field. Visual Search and Eagle Eye measure it, and the difference between good and poor performers is search method rather than eyesight: systematic row-by-row scanning finds targets that random hunting misses.
Inhibition
Stopping a response already underway. Go/No-Go is the standard laboratory measure and the Stroop Test is the classic interference task. This is the capacity most people mean by 'focus', and it is the one most clearly degraded by tiredness.
On Go/No-Go the number that matters is your false alarm count, not your speed. A fast run with three false alarms is a worse result than a slower clean one, and the task is designed so that pure speed produces a bad score.
Run all three, on the same day. One test measures one ability, and knowing which of the three is weak is the entire point.
Compare a good day against a bad one. Inhibition in particular swings with sleep, and seeing it in a number is more persuasive than noticing it in yourself.
Ignore cross-site comparison. Presentation speed and hardware latency move these numbers enough to make it meaningless — the same caveat that applies to every cognitive benchmark.
Log them. The Score Tracker keeps a rolling average and a trend line so day-to-day noise does not mislead you.
In descending order of how well supported each one is.
Remove the interruption. A phone in another room does more for a working hour than any drill on this site. The evidence on interruption costs is far stronger than the evidence on attention training.
Sleep. Inhibition is measurably worse when tired, and it degrades before people notice any subjective tiredness.
Work in fixed blocks. Twenty-five minutes of protected time, then a real break away from a screen. A break spent scrolling does not restore attention.
Exercise. Aerobic exercise has some of the better evidence of any lifestyle intervention for executive function.
Then, if you like, practise. Attention drills reliably improve attention drills. Whether that spreads further is contested, and the honest answer is that it mostly does not.
The whole set of attention drills is on the focus and attention hub, and brain games for ADHD goes further into what the evidence supports where attention difficulties are clinical rather than situational.
No. That claim is repeated constantly and has no credible source behind it — journalists tracing it have never found the study it is attributed to, and the goldfish comparison appears baseless.
Run three tasks rather than one: Radar Scan for sustained attention, Visual Search for selective attention and Go/No-Go for inhibition. Which one you score badly on is more useful than any combined number.
The decline in detection performance over time on a monitoring task. It begins within the first few minutes, which is why roles requiring sustained watching rotate people regularly.
Removing interruptions, sleeping enough and working in fixed blocks all have better evidence than attention training does. Drills reliably improve the drills; transfer beyond them is contested.
Yes — free, instant, no account and nothing to install.
Attention is three things, not one, and the useful result is finding out which of the three is weak. Run all three, compare a rested day with a tired one, and then fix your environment before you train anything. Start on the [focus and attention hub](/focus-and-attention).