What you get, and what you do not
At the end of a full run you have six numbers: a reaction time in milliseconds, an aim speed in milliseconds per target, a digit span, a sequence level, a visual memory level and a colour discrimination level. Two of them carry a percentile. Reaction time is placed on the distribution from Woods and colleagues, and sequence memory on forward Corsi span from Facchin and colleagues. The other four show your score alone, because no published norm matches the way they are played here, and we would rather show you nothing than a number we invented. Every source is listed on the methodology page.
What you will not get is a single combined figure. We could easily average the six into one "brain score" and it would be more shareable than six separate results, but it would also be meaningless. The tests measure abilities that are only loosely related to each other, and averaging a millisecond, a digit count and a level number requires inventing weights that no research supports. A composite would look authoritative precisely where it is least justified.
Why these six, and why this order
The order is not arbitrary. Reaction time comes first because it is the most sensitive to fatigue and the least dependent on strategy — you want it while you are fresh, before six minutes of concentration have worn on you. Aim follows because it is the other speed measure, and grouping the two speed tests together keeps the mental gear-change to one point in the run rather than four.
Then the memory block: numbers, sequence, visual. These three are placed in the middle because they benefit from being warmed up but suffer from being exhausted, and the middle of a six-minute run is the widest part of that window. They are ordered from most verbal to most spatial, which spreads the load rather than asking the same system three times in a row.
Colour discrimination goes last for a specific reason: it is the only test where staring is actively harmful. Your retina adapts to a field of similar colour within seconds, so the test rewards fresh eyes on each level rather than accumulated concentration. Putting it at the end also means that if the run has tired you out, the test least dependent on sustained attention is the one that absorbs it.
What we measured while building this
Some of what follows is not in any paper. It is what we found while making the tests work in a browser, and it changes how much your numbers mean.
A browser adds roughly 80 milliseconds you did not spend. Between a display refreshing, the operating system noticing a click and JavaScript receiving the event, a measurable delay accumulates that has nothing to do with you. Anwyl-Irvine and colleagues quantified this, and we subtract a documented 80 ms offset before comparing your reaction time with laboratory norms. This is also why a browser reaction test and a laboratory one are not the same measurement, and why sites that report raw browser milliseconds against laboratory averages make everyone look slower than they are.
A background tab destroys timing. Browsers deliberately throttle timers in tabs you are not looking at, to save battery. We ran the full six-test suite through an automated harness in a background tab and rounds that normally finish in two seconds exceeded a twenty-five second budget. The effect is large enough that a measurement taken while you were reading something else in another tab would be worthless. That is why every timed test here refuses to start when the tab is hidden and shows you a notice instead, rather than quietly recording a number.
Your screen is part of two of these tests. Colour discrimination depends on panel quality, brightness and ambient light more than it depends on the player. And on a 320-pixel-wide phone, a six-by-six grid physically cannot give every tile a 44-pixel touch target — the arithmetic simply does not fit — so tiles shrink on small screens. We measured this and sized the grids from the actual stage width so that tiles never overlap; an earlier version overlapped neighbouring tiles by three pixels, which meant a click near a tile edge could register on the wrong square. If you play the grid tests on a phone and a desktop, you are playing two slightly different games.
Making your run comparable to your next one
The comparison that carries the most information is you against yourself, and that only works if you hold the conditions still. Use the same device. Use the same pointing method — a mouse, a trackpad and a touchscreen produce genuinely different aim and reaction numbers, and the difference between them is larger than most day-to-day variation in you. Keep the tab in the foreground for the whole run, which the tests now enforce. Try to run at a similar time of day, since alertness follows a daily rhythm that moves reaction time by more than a good night's practice will.
Then run it more than once. Single results are noisy across all six tests, and the shape of five runs tells you something a single run cannot. Your scores are kept in this browser only, so the history builds up on the device you actually use.
Frequently asked questions
Can I stop partway through?
Yes. The run records each test as you finish it, so you can stop after any test and the scores you already have are kept. Coming back to the run page offers to resume from the first test you have not done.
Do I have to do the full run to see percentiles?
No. Each test works on its own and shows the same result panel, curve and percentile whether you reached it through the run or directly. The run exists to collect the six into one summary, not to unlock anything.
Where do my scores go?
Into this browser's local storage, and nowhere else. There is no account, and no score is sent to a server. Clearing site data clears them, and a different browser or device starts from scratch.
Why does my second run differ so much from my first?
Partly practice and partly noise. The first run of any test includes learning what the test wants, which is why first attempts are usually the worst ones you will record. Beyond that, span and reaction measures are genuinely variable within a person from trial to trial. Treat your typical level across several runs as your result, not your best or your first.