How to Build a Repeatable Skills Baseline

A baseline is a starting point, not a ranking

A first score is useful because it gives future practice somewhere concrete to return to. It is not a measure of intelligence, employability, or a fixed limit. Browser games measure a narrow task under the conditions of one session: a particular device, passage, input method, and level of alertness.

Make the task repeatable before you make it ambitious. Use the same device when possible, a similar time of day, and the same mode. Then record several attempts rather than treating the best or worst run as your ability.

Use a small sample

Take three calm attempts, with a short break between them. For example, type three one-minute passages, note net WPM and accuracy, then use the middle result as a provisional baseline. The middle is less distorted by an unusually lucky run, distraction, or a slow warm-up.

For reading, keep speed and recall together. A 260 WPM run with one correct answer is not equivalent to a slower run with four correct answers. This site's reading efficiency combines the two for practice, but it is still a task score, not a population comparison.

Compare like with like

Scores differ when the task changes. A familiar typing passage, a dense instruction manual, and a numeric keypad drill use different skills. A score can also shift with keyboard layout, screen size, mouse settings, or interruptions. Label each session with the mode and any unusual condition.

Do not convert a game score into a percentile unless the exact game has a transparent, relevant reference sample. A personal trend is usually more reliable: compare your own median across sessions under similar conditions.

Turn the baseline into practice

Choose one observable constraint. If typing accuracy falls when you speed up, make accuracy the constraint. If reading answers miss names and causes, pause after each paragraph and name its point. If pattern matching errors cluster on one-character changes, compare position by position instead of scanning for a general resemblance.

One repeatable session can include a slow accuracy pass, a normal-pace pass, and one measured attempt. Review what changed before deciding whether to repeat the method or test a different one.

• Track median and accuracy, not only a personal best.
• Change one variable at a time.
• Stop when discomfort or repeated guessing changes the task.

Worked example

Suppose Maya completes three reading runs at 215, 240, and 225 WPM, with one, two, and four answers correct. The 240 WPM run is not automatically the strongest because the question result is incomplete. She can calculate eWPM for each run and use the middle combined result as a baseline.

In a later session, Maya can compare medians from the same passage length and question format. A repeatable improvement with the same or more correct answers is more informative than one unusually fast result.

Frequently asked questions