Cognilo
How puzzle scores and adaptive difficulty work
Cognilo shows numbers, so it owes you an explanation of what they mean and what they don’t. This page covers how difficulty is chosen, why scores move between sessions, and why practice and benchmark results are kept apart.
How do adaptive puzzle games select difficulty?
Each game has levels. Cognilo keeps a rating for each skill and picks the level that sits at your current rating, so a typical round is challenging but winnable. Clear a round cleanly and the next one steps up; slip, and it eases off.
One lucky round is not enough to jump a level: a rating only moves a limited amount per round, and a shifted range is only reported when several sessions agree. If you say a round felt too easy or too hard, the next round of that game moves one level, and that nudge never touches your rating.
Why do game scores change between sessions?
Sleep, time of day, interruptions, a new device and simple chance all move a single session. That is why Cognilo shows your recent range rather than treating any one number as the truth, and why a lower session is described as “below your recent range” rather than a failure.
What is the difference between practice and a benchmark?
Practice rounds are what you play every day, so you become familiar with them and your Training Rating reflects both skill and familiarity. The Cognilo Benchmark uses standardised forms built from held-out seeds that never appear in everyday practice, at consistent conditions, so the comparison over time is fairer.
Repeating the Benchmark still builds familiarity with the puzzle types, so it is a fairer comparison, not a perfect one.
Are timed and untimed scores compared?
No. Precision (untimed) is the default, and the Training Rating is built on it. Sprint (timed) scores are shown separately and never count toward the Training Rating, so playing faster never outranks playing carefully. Zen and Review replays are unrated.
What the ratings do not mean
Ratings describe performance on Cognilo’s own puzzles. They are not IQ, not a clinical measure and not a verdict on how well you think in everyday life. Practice reliably improves the practised task; evidence for broad carry-over to daily life is limited, so Cognilo does not claim it.
Frequently asked questions
Why did my score go down after a good week?
Single sessions vary for ordinary reasons. Cognilo limits how far one session can move a rating and only reports a shifted range when several sessions agree, so one lower day is usually just one lower day.
Can I play puzzles without affecting my rating?
Yes. Zen mode and Review replays are unrated, and Sprint scores are kept out of the Training Rating.
Does Cognilo compare me with other people?
No. Cognilo does not publish percentiles or population comparisons. Your numbers are compared only with your own earlier rounds. Optional Game Center leaderboards exist for anyone who switches Competition on.