Local switch cost is the difference between your switch-rule and repeat-rule timing averages on a single fixed thirteen-row route; a global switch cost would compare that single-run result to a population norm the route itself does not contain. The Task Switching Game in your browser does the first job with no secret instruction. Each row shows one Amber or Violet circle or square, and the page states the current rule: Colour or Shape. Under Colour, F means Amber and J means Violet. Under Shape, F means Circle and J means Square. The rule is always visible, so you do not have to infer a mapping or remember an unseen order. The route starts with one Colour-rule starter row, then runs twelve comparison rows. Six comparison rows repeat the prior rule and six switch between Colour and Shape. After the route, the completion panel reports accuracy alongside three numbers: the average visible response time on repeat rows, the average visible response time on switch rows, and switch cost as the switch average minus the repeat average. Everything in that panel is calculated only from your thirteen visible responses on this device, so the panel describes a local switch cost, not a global benchmark. You can open the game, read the rule before the first row, and inspect the same transparent route on every attempt.
What local vs global switch costs actually mean
Working with two related ideas at once can muddy what a number represents, so it helps to keep them apart. In this context, a local switch cost is the gap between the switch-rule timing average and the repeat-rule timing average for one named run on one device, with no other comparison. It uses only the rows that one attempt produced. A global switch cost, by contrast, would compare that single-run gap against a population reference: a sample of other attempts, a published norm, or a percentile table. The Task Switching Game only publishes a local figure. The thirteen rows are product-authored and fixed; the same order, colours, shapes, response side labels, and scoring method are reused on every attempt, which makes the local summary reproducible, but reproducibility on a fixed route does not produce a population norm. There is no comparison set, no cutoff, no standard score, and no leaderboard inside the game.
Studies and article summaries that report switch cost in milliseconds typically come from a controlled task where participants, instructions, equipment, and analysis are standardised, and the result is usually compared to a control condition or to a group average. That setup is different from a single browser game on a single device. The same formula is used in both places — switch average minus repeat average — but the source data, the controls, and the limits on interpretation are not the same. Treat the browser panel as a local timing summary and treat any external study you read as a different kind of evidence rather than as a global reference for the browser game.
How the fixed route builds the two averages
The route is short and fully disclosed. It starts with one Colour-rule starter row, so that row has no prior rule to repeat or switch from. The starter row is excluded from both averages. The remaining twelve rows split into two balanced halves: six rows that continue the previous rule (repeat rows) and six rows that flip Colour to Shape or Shape to Colour (switch rows). The completion panel reports the repeat-rule average, the switch-rule average, and switch cost as the simple difference between those two visible averages.
| Row category | Rule pattern | Count | Treatment in the completion panel |
|---|---|---|---|
| Starter row | First row is always Colour | 1 | Excluded from both averages |
| Repeat row | Same rule as the prior row | 6 | Included in the repeat average |
| Switch row | Rule flips between Colour and Shape | 6 | Included in the switch average |
Because the route is short, every individual response carries visible weight in the two averages. Accurate and inaccurate responses both contribute to timing, but accuracy is shown next to time in the same panel, so a quick response that followed the wrong rule is not hidden behind a fast clock. The first starter row cannot be classified as a repeat or a switch because there is no prior rule on screen, which is why it sits outside both averages. As an illustration of the formula, if your repeat average on this run reads 1.20 seconds and your switch average reads 1.50 seconds, then switch cost equals 1.50 − 1.20, which is 0.30 seconds. A positive value means switch rows were slower than repeat rows on this run; a negative value would mean switch rows were faster, and zero is also possible. None of these directions implies a global claim — they are simply what this short fixed route produced on this one attempt.
How to run a single local route in the browser
The whole route runs in one page and does not need an account, a download, or a separate app. The numbered steps below match the visible loop on the page, from Start to the final panel.
- Open the Task Switching Game page and press Start when you are ready to begin.
- For each row, read the current rule label (Colour or Shape) and the F and J mapping shown for that rule before you choose.
- Look at the visible circle or square, decide which side of that stimulus the current rule points to, and answer with the matching button, F key, or J key.
- Continue row by row through the same fixed thirteen-row sequence, including the starter row and the twelve comparison rows.
- If the tab becomes hidden, the row pauses and Resume begins a fresh visible interval when you return, so time spent on another app is not folded into the response.
- When the route ends, read the completion panel for accuracy, the repeat-row average, the switch-row average, and switch cost as the switch average minus the repeat average — these are the local numbers for this attempt.
- Use Restart to begin the same transparent thirteen-row route again from the start, on the same device.
Each step is part of the same visible loop, and the panel is the only place a number appears. Nothing else on the page scores or records you, so the local switch cost you read at the end comes from one explicit choice per displayed row.
Why the result stays a local summary
Several properties of the game hold the panel to a local reading rather than a global one. The route is fixed, so every run shares the same fixture set, but the same fixture set across runs does not turn one attempt into a population. Different screen sizes, pointer versus touch versus keyboard input, browser timing precision, language settings, accessibility settings, interruptions during a row, ordinary daily variation, and prior familiarity with the game can each shift one attempt, which is why the same person often sees a different switch cost on different days. The hidden-tab pause is one safeguard: when the tab loses visibility, the current row pauses and Resume begins a fresh visible interval, so background time is not silently mixed into the visible clock. That makes the local timer less confusing, not more standardised.
A separate safeguard holds the score to this device. The shared game shell may retain a best local points total in browser storage, but no account is required, no online leaderboard exists, and no score is sent to Lizely. Restart begins the same transparent thirteen-row route, so the figures on a fresh attempt come from new visible answers, not from a hidden seed. Taken together, those choices mean the panel reflects one run on this device, and the local result is the only result the tool produces.
What this game does not measure
The page states plainly that the route is an entertainment exercise, not a clinical, IQ, or cognitive assessment, so the local numbers do not diagnose attention, measure intelligence, or show whether a brain has improved. Switch cost, repeat average, and switch average on this panel are not a percentile, an attention score, an executive-function diagnosis, a grade, a readiness measure, a benchmark, or a prediction about another task. There is no clinical cutoff, population norm, educational score, professional screening value, or treatment value attached to the result. Differences in screen size, input device, browser timing, interruptions, familiarity, language, or accessibility settings are reasons one attempt can look different from another, which is exactly why the same formula on the same fixed route can produce different numbers across runs. Readers who want to explore the conceptual framing behind switching research can compare this local view against the cognitive-flexibility framing in this related guide, which treats the same tool from a different angle. If you need evaluation or support for a health, learning, work, or accessibility question, an appropriate qualified service is the right place to look — not a browser game with a thirteen-row fixed route. The Task Switching Game is the right tool when you want to inspect a single transparent local route and read the four visible numbers it produces; it is not the right tool when you want a clinical interpretation, a population comparison, or a longitudinal tracking programme.
For a deeper look, see Tower of Hanoi vs Tower of London: Key Differences.