Above the 75th percentile
N-back is not a recall task but an updating one: you hold the last two letters and shift the window on every new one, dropping the oldest. Twenty-three and up out of twenty-five means the window did not fall apart across all twenty-five steps in a row. Both halves count: pressing on a match, and holding your hand still when there is none. A false press costs exactly what a missed match costs, which is why the top of this scale belongs to the steady rather than the fast. The caveat that belongs specifically here: training n-back reliably improves n-back, and that is the one thing firmly established about it.
Middle: 25–75
Twenty to twenty-two out of twenty-five is the ordinary result, and it almost always breaks in one place: where two identical letters sit close together but not one apart. That pair pulls your hand toward the button even though the rule says it is not a match. The other common break comes after an error — one wrong press knocks the window count out, and the next two or three steps run blind until you rebuild what came before. Which is why what helps here is not speed but the habit of not chasing a miss: accept the lost trial and carry on from the next letter.
Below the 25th percentile
Seventeen to nineteen out of twenty-five is the bottom of the scale, and the first thing to weigh is that the n-back rule confuses almost everyone on first contact. The comparison worth making is not against what you should have scored but against whether you had the task straight before the stream began — it cannot be re-explained mid-run, the letters arrive every two and a half seconds and do not wait. One session, one device, one minute of attention; short sleep and a conversation in the background move this score more than any other in the battery, because the task rests on holding rather than on knowing. The number has no clinical meaning.