One digit on screen, and two rules that keep swapping: odd or even? and bigger or smaller than 5? The rule is always written above the digit — and we measure what each change costs you.
It measures mental flexibility: how cheaply you can drop one rule and load another.
The same digit can be even under one rule and bigger than 5 under the other. So when the rule changes you cannot answer from what you were just doing — you have to unload the old rule and install the new one. That operation takes time, and research calls the time it takes the switch cost.
It is the headline number in your result:
Switch cost = your time on trials where the rule changed − your time on trials where it stayed
The subtraction does the same job here as in our Stroop test: it separates flexibility from raw speed. A slow hand is slow on both kinds of trial, so it cancels out, and what remains is what the change of rule cost you. Lower is better.
Two of the six axes on this site already depend on colour. If this one did too, three different scores would partly measure the same thing — colour perception. Digits are known to everyone who reads, and judging them is immediate with nothing to learn.
The rule is also written in words every single trial, with colour only as an extra cue. So a person with colour vision deficiency loses nothing, and everyone else can see the switch at a glance without reading each time.
Thirty-two trials, with the switch rate fixed at 50 percent rather than left to chance. Pure randomness can produce a long run of same-rule trials, and a long run lets you relax and inflates the contrast; a run of switches does the opposite by making the change predictable.
The same digit never appears twice in a row either, because repeating the stimulus speeds you up for reasons that have nothing to do with rules. The first trial is excluded, since nothing came before it to switch from, and incorrect trials are excluded from the timing and counted in accuracy only.
If too few trials of either kind survive, we show no number at all instead of showing a confident one. An unreliable measurement is worse than a missing one.
Switch cost is one of the most state-sensitive measures on this site. Tiredness, interruptions, and how settled your attention is on the day all show up in it. Read the trend across weeks, not the gap between two attempts.
Your switch cost on this task, mostly by learning the two rules so well that loading them becomes cheap. Whether that makes you better at juggling real tasks is a much bigger claim, and the evidence for that kind of transfer is weak — so we do not make it.
It is free and runs in your browser with nothing to install and no sign-up. Signing in is optional and only saves your results across devices.