Processing Speed Test — It Isn't One Number
Search for a processing speed test and you will be handed a number. The problem is that processing speed is not one ability, and a single number quietly averages away the thing you actually wanted to know.
How fast you locate a target in a cluttered field, how fast you resolve two conflicting signals, and how fast you switch between two rules are separable. People are routinely quick at one and ordinary at another. Four of the tests here measure different facets of it — playing all four tells you far more than any one of them repeated.
Four measures, four different facets
Each of these is a published paradigm, and each isolates a different part of what people lump together as speed. Play them in any order — the interesting result is the shape across all four, not any single score.
Schulte Table
Visual scanning speed with a fixed gaze. The bottleneck is how much of the grid your peripheral vision can serve up without your eyes moving.
~2 minPlay →Visual Search
Target detection among distractors. Same eyes, different bottleneck: how efficiently you reject things that are not the target.
~2 minPlay →Trail Making Test
Sequenced movement plus, in Part B, switching between two sequences. The gap between your Part A and Part B time isolates the cost of switching from raw speed.
~1 minPlay →Cancellation Task
Exhaustive scanning. Unlike the others you cannot stop at the first hit, so it also measures whether your sweep has a strategy — what you miss says more than what you find.
~2 minPlay →
What processing speed actually refers to
In cognitive psychology, processing speed is shorthand for how quickly you can carry out simple, well-defined mental operations — find, compare, decide, respond. It is deliberately about the easy operations: the tasks are trivial in difficulty so that the only thing varying is how fast you get through them.
That is also why it fractionates. A task where the bottleneck is your eyes sweeping a field measures something different from one where the bottleneck is resolving a conflict between two signals. Both are 'speed', but improving one does not automatically move the other.
Why a single score is misleading
Take two people with the same overall score. One is fast at visual scanning and average at switching; the other is the reverse. Averaged, they look identical. In practice they will find completely different tasks tiring.
The other problem with a single number is that accuracy is usually folded in silently. Going faster while getting more wrong is not an improvement, but a combined score can hide it. Every test here reports time and accuracy separately alongside the score, so you can see which one moved.
The confound nobody mentions: your input device
Every browser-based speed test is measuring your hand as well as your head. A trackpad is slower than a mouse for repeated targeting, a phone is slower still for grid tasks and faster for simple taps, and a laggy display adds a constant to every trial. None of that is your processing speed, and all of it lands in the score.
It matters less than you would think for tracking yourself and more than you would think for comparing against anyone else. Keep the device constant across sessions and your own trend stays meaningful; switch from laptop to phone mid-week and the drop you see is mostly hardware. This is also the honest reason a global leaderboard across mixed devices should be read loosely.
What actually improves it
Expect near transfer and little else. Practising a specific test makes you better at that test and its close relatives; the evidence for it generalising to unrelated everyday tasks is weak. That is not a reason to avoid these tests — it is a reason to treat them as a measurement instrument and a warm-up rather than a training programme.
The things with better evidence behind them for how quickly you process anything on a given day are unglamorous: sleep, whether you are doing one thing or three, and how long you have already been at it. A test result taken at 2am measures your sleep debt more than your processing speed.
Frequently asked questions
What is a good processing speed test score?+
There isn't a meaningful universal answer, and any site that gives you one is overstating what it knows. Scores depend on the specific paradigm, the difficulty tier, the input device, and how practised you are. The useful comparison is against your own earlier runs on the same test, which is why every score here is saved to your history if you sign in.
Why do I score well on one test and badly on another?+
Because they measure different things. Scanning a grid with a fixed gaze and alternating between two sequences share the word 'speed' and very little else. A gap between two of these tests is normal and is the most informative thing you'll get out of playing all four.
Can processing speed be improved?+
You will get better at these specific tests with practice — that part is reliable. Whether it generalises to unrelated tasks is much weaker evidence, and the honest position is that it mostly doesn't. Sleep, doing one thing at a time, and not being three hours into a session move your same-day performance far more than practice does.
Is this the same as the digit symbol substitution test?+
It measures overlapping ground, but that specific instrument is a standardised, copyrighted clinical subtest and isn't offered here. The paradigms here are ones that can be presented honestly as practice exercises rather than as clinical measures.