Verbal Memory Statistics
Each attempt is recorded as the number of words judged correctly before three mistakes. This page is the distribution of those counts, with the breakpoints written out.
No curve yet — the sample stands at 0 of the 200 attempts needed before a distribution is worth drawing. The scale below is the range results land in.
The breakpoint table appears once there are enough attempts to compute one. There are 0 so far.
A count, not a level, so the axis is longer
The other memory tests on this site end on a level. This one accumulates a point per correct judgement and can run into the dozens or the low hundreds, which is why the chart is a wide discrete axis rather than a short staircase.
That width is also why the percentile moves more smoothly here than on the grid tests. Clearing one more word does not jump you past a whole band of people the way clearing one more level does.
Very low scores are a mix of genuine early failures and abandoned runs that spent the three lives on the first words while working out the buttons. They are real attempts and they are counted; they are not a picture of anybody’s vocabulary.
Language is part of the sample, and it is not separated
A run in Japanese and a run in English are not the same task, and they currently share one distribution. Splitting by language would leave every curve starving for a long time, and a starving curve is worse than a mixed one that says it is mixed.
The practical effect is that the middle of the chart is a blend. A reader comparing themselves to it is comparing themselves to everyone who finished a run, in every language the site serves.
If a language split ever has enough attempts to stand on its own, the chart will say so. Until then the honest caption is the attempt count and nothing dressed as a per-language norm.
What the sample is made of
Finished runs only. Closing the tab before the lives run out produces no score, so the distribution is not padded with half-attempts.
Repeat visits count separately, as everywhere on this site. People improve over the first few tries as they stop confusing similar words, which shifts the sample upward relative to a population of first attempts.
Input modes are pooled. The score is a count, so a tap versus a click cannot move it, and there is nothing gained by splitting.
This is a measurement toy, not a medical or psychological test. No score here says anything about your health, attention or cognition, and it is not a screening tool of any kind.
Questions
- What is the average verbal memory score?
- The chart and the breakpoints on this page are the current figures from the attempts recorded here. They will move as more arrive, and they mix every language the site serves.
- Why are very low scores on the chart at all?
- Because a run that spends three lives in the first handful of words is still a finished attempt. Leaving them out would pretend everyone understood the buttons on the first try.
- Are the languages split?
- Not yet. There are not enough attempts in any one language to draw a curve from, and a thin curve is worse than a mixed sample that admits it is mixed.