How to Interpret Focus Test Results

Completion time is the elapsed interval from a defined start to a defined finish. Accuracy is the proportion of scored responses that met the rule; its denominator matters. Errors may be wrong choices during a search, while misses are required responses that never occurred and false alarms are responses made when the rule required withholding.

Span or level is the largest amount of material reached under a page's progression rule. It is not a universal memory unit. An n-level states how far back the current item must be compared; n=2 means comparison with the item two positions earlier, not that two items are remembered in every ordinary sense.

Response-time summaries should say which trials enter them. A median of correct responses is resistant to one unusually slow trial, but it still reflects display, browser, input, decision, and movement time together. A best response is especially vulnerable to anticipation and chance and should remain secondary.

A person can respond faster by accepting more mistakes, or slow down to verify each choice. That tradeoff means a shorter time is not automatically a better result. Compare time together with error rate, misses, or false alarms, and check whether the response policy appears to have changed.

The fastest isolated result is a poor anchor because it selects an extreme from a noisy series. A median across comparable trials or a small group of recent sessions is usually a more stable description. Keep the fastest value if it is motivating, but do not let it replace the complete record.

First, choose one task and one purpose. Second, keep its mode, duration, grid size, cue rate, and difficulty fixed. Third, use the same device, screen orientation, zoom, and input method. A phone tap, laptop trackpad, external mouse, and keyboard do not create interchangeable timing conditions.

Fourth, decide in advance how many sessions to run instead of repeating until a desired score appears. Fifth, keep all outcome fields, including errors and invalidity flags. Sixth, compare a recent group with an earlier matching group and describe the change in the task's own units rather than converting it into a percentage of 'focus.'

Time of day, interruptions, fatigue, discomfort, and familiarity can still differ even after these controls. A personal comparison is therefore a practical description, not a controlled experiment. Notes about changed conditions are often more informative than an extra decimal place.

Treat a session as non-comparable when the tab was hidden during a timed stream, the wrong input device was used, instructions were misunderstood, a touch was accidental, the page visibly stalled, or another person completed part of the task. Some tools pause on focus loss; others attach a validity flag. Read that page's behavior before deciding.

Do not silently delete every inconvenient result. Keeping a flagged record prevents selective memory from turning a history into only personal bests. Discard a run from trend calculations when the rule was not followed, and retain the reason when the interface supports it.

The following numbers are invented solely to demonstrate interpretation. They are not norms, expected values, or real user data. Imagine five matching sessions on the same laptop and settings:

• Session 1: median correct response 510 ms, accuracy 96%, no validity flag.
• Session 2: 495 ms, 94%, no flag; slightly faster with slightly lower accuracy.
• Session 3: 420 ms, 71%, many false alarms; speed came with a large accuracy change.
• Session 4: 505 ms, 95%, tab-hidden flag; exclude it from a clean timing comparison.
• Session 5: 488 ms, 96%, no flag; together, Sessions 1, 2, and 5 suggest a stable band, not a diagnosis or guaranteed trend.

Repeated exposure teaches the instructions, layout, response mapping, and likely stimulus patterns. That familiarity can reduce hesitation without establishing a change outside the page. Early gains are especially likely to mix learning the interface with performing the underlying rule.

Use recent groups of sessions to describe ordinary variability: broadly stable, more variable, faster at similar accuracy, or more accurate at similar timing. Avoid drawing a rising line through every result and calling it cognitive growth. The site does not supply the controls needed for that inference.

No score on focus-game.org diagnoses ADHD, a memory disorder, cognitive impairment, or another health condition. The pages do not provide an IQ, brain-age estimate, clinical cutoff, age norm, school-readiness judgment, driving clearance, or employment assessment.

If a concern affects daily life, the next step depends on that real context and may involve an appropriate health, educational, workplace, or accessibility professional. Repeating a browser task cannot identify a cause and should not delay suitable support.

Choose Reaction Time for a simple signal-to-action interval; Attention or Sustained Attention for target and withhold decisions; Stroop for competing word and color cues; Schulte or Visual Search for stationary scanning; and the memory guide when information must be retained or continuously updated.

If the real goal is to begin work, a timer or direct next action is usually more relevant than another score. Stop testing once the planned observation is complete and return to the activity that matters outside the site.