In a flanker-style browser game, accuracy and reaction time are reported side by side because they describe two different things about the same twelve responses. Accuracy counts how many centre-arrow answers match the direction of the outlined middle arrow; reaction time measures how long each visible row stays on screen before a single response is accepted. The Flanker Test keeps both metrics on the same fixed route so you can see whether a fast session was also a clean one, or whether extra speed came at the cost of wrong answers. Each row has five arrows but only the outlined centre arrow counts, so the surrounding arrows can point the same way or the opposite way without changing the rule. The fixed twelve-row route is split into six matching rows and six conflicting rows, which is what makes a within-route comparison possible at all. That pairing is the core of any flanker-style accuracy vs reaction time reading: the numbers come from one private browser round, not from a leaderboard, a population table, or a diagnostic cutoff.
Why Accuracy and Reaction Time Are Reported Together
In any flanker-style challenge, accuracy and reaction time answer different questions, and answering only one of them hides part of the picture. A perfect accuracy score tells you every centre arrow was identified correctly, but says nothing about whether the responses came in slowly because of careful checking or quickly by accident. A fast average reaction time tells you the rows moved quickly, but says nothing about whether speed traded against correctness. Putting the two numbers on the same panel makes the trade-off visible inside the route, rather than burying it behind a single combined score.
This is the practical reason the Flanker Test reports both. The game accepts exactly one response for each visible row and immediately advances, so the time recorded for a row ends at the moment of that single accepted response. There is no second chance, no correction, and no extra click that can add another response to the same displayed row. That single-response rule is what keeps the time figures honest: each reaction time corresponds to one and only one centre-arrow decision, and that decision is either right or wrong exactly once.
What the Fixed Twelve-Row Route Reveals
The route in the Flanker Test is fixed at twelve product-authored five-arrow rows. Six of those rows are matching rows, where the surrounding arrows point in the same direction as the outlined centre arrow. Six are conflicting rows, where the surrounding arrows point in the opposite direction. The route never changes between sessions, which is what allows the result panel to compare matching and conflicting numbers drawn from the same fixture every run.
This matters for any accuracy vs reaction time reading because the only thing that changes between the two halves of the route is the relationship between the surrounding arrows and the centre arrow. The centre arrow still answers the question by itself; the surrounding arrows only provide context that either helps or hinders quick identification. A within-route comparison is therefore not hidden by row-by-row variation in difficulty, since every row uses the same five-arrow layout and the same one-response rule.
| Route feature | What it controls | Why it matters for an accuracy vs reaction time reading |
|---|---|---|
| Twelve fixed rows | Same fixture every run | Lets matching and conflicting averages be compared on equal footing |
| Six matching rows | Surrounding arrows align with centre | Provides the baseline reaction time |
| Six conflicting rows | Surrounding arrows oppose centre | Provides the slowed reaction time |
| One response per visible row | Single accepted answer | Each reaction time pairs with exactly one accuracy decision |
| Hidden-tab pause | Background time excluded | Keeps reaction time figures tied to a visible interval |
How to Play the Flanker Test
- Start the fixed twelve-row route while the tab is visible. Make sure the page is the active tab before you begin; switching away mid-row pauses the current row rather than recording a hidden-tab response.
- For each five-arrow row, choose only the direction of the outlined centre arrow. Use the large labelled Left and Right buttons, or the Left and Right Arrow keys as an optional shortcut. The surrounding arrows do not change the rule and should not be treated as a second target.
- Read the matching, conflicting, and interference figures as results from this private browser game only. After the twelfth response, the result panel reports accuracy alongside the matching-row average reaction time, the conflicting-row average reaction time, and the interference index. Treat the figures as a record of this session on this device.
Reading the Matching and Conflicting Averages
Once the route finishes, the result panel splits reaction time into two within-route averages: the average response time on matching rows and the average response time on conflicting rows. Accuracy is reported separately, as the share of centre-arrow answers that were correct across the full twelve rows. Each figure has a single, narrow meaning that does not extend beyond the route that produced it:
- Matching-row average — the mean reaction time, in browser milliseconds, across the six rows where the surrounding arrows pointed the same way as the centre arrow.
- Conflicting-row average — the mean reaction time across the six rows where the surrounding arrows pointed the opposite way.
- Accuracy — the count of correct centre-arrow answers out of twelve, shown alongside the reaction time figures rather than blended into them.
None of these three numbers is compared with a population table, a published reference, an age range, or a clinical sample. The averages describe what happened in this one round on this one device, including any ordinary session-to-session variation caused by screen size, input method, browser scheduling, or familiarity with the route.
Working Out the Interference Index
The interference index is reported as a single number on the same panel as the accuracy and the two averages. It is defined plainly as the conflicting-row average reaction time minus the matching-row average reaction time, calculated only from the twelve rows in this run.
For example, suppose a session ends with the following within-route figures: matching-row average = 600 ms, conflicting-row average = 720 ms. Substituting into the formula gives:
Interference index = conflicting average − matching average = 720 ms − 600 ms = 120 ms
A positive interference index means the conflicting rows took longer, on average, than the matching rows in that session. A negative interference index means the conflicting rows were, on average, faster than the matching rows, which is unusual but possible. A number close to zero means the two halves of the route took about the same time. The figure is not ranked against another player, age range, classroom, workplace, clinical sample, or published reference; it is purely the within-route difference from this device.
Why a Faster Route Is Not Automatically Better
The Flanker Test keeps accuracy visible on the result panel for a reason. A session that finishes the twelve rows quickly but with several wrong centre-arrow answers is not presented as better than a slower, cleaner session. The result panel does not collapse the two metrics into one combined score, nor does it reward raw speed on its own. Accuracy and reaction time remain separate columns of the same summary, so the trade-off between them stays legible at a glance.
This is also why the route accepts only one response per visible row. A repeated key event or an extra click cannot add another response to the same displayed row, so the recorded reaction time always corresponds to the single decision that counted. Without that one-response rule, a faster reaction time figure could be the result of accidental double-taps rather than a deliberate centre-arrow answer, and the trade-off reading would be less honest.
What These Figures Do Not Measure
The accuracy and reaction time figures on the result panel are local browser-game numbers. They are not a clinical or IQ assessment, they do not diagnose attention, they do not measure intelligence, and they do not show whether a brain has improved over time. There is no health threshold, no norm group, no educational score, and no professional use attached to the numbers. They are also not a substitute for a qualified assessment if support is needed for health, learning, work, or accessibility.
The game runs entirely in the browser. It does not ask for an account, profile, name, camera, microphone, or upload, and the figures stay on this device. The shared game shell may keep a local best score in browser storage, but it is not an online leaderboard and is not sent to Lizely. For readers who want a closer look at how matching and conflicting rows are built in this same browser game, the congruent vs incongruent flanker trials guide walks through the trial types in more detail.
If you're weighing options, Go No Go vs Continuous Performance Test: Browser Routes covers this in detail.