Our community has grown to 12,900+ registered players since 2024. Every round on Cleverini leaves a few anonymous numbers behind: how long it took, which answer was picked, which level it was on. We study those patterns and publish what we find here. No personal data, no medical claims, just the shape of play since February 2024.
12,900+registered players
7.1Krounds played
92Kanswers recorded
33countries
Live country ranking
Which country’s players score highest in Cleverini’s logic and math games?
Every answer counts for the country it was played from. Play to move yours up.
Updated November 15, 2026 · 516 players scored · 38 countries seen
1🇫🇷France2 players · 7,026 answers+5.2pts–No change since yesterday
2🇬🇪Georgia1 player · 40 answers+4.4pts–No change since yesterday
2🇳🇱The Netherlands1 player · 125 answers+4.4pts–No change since yesterday
4🇸🇦Saudi Arabia1 player · 46 answers+4.3pts–No change since yesterday
5🇳🇬Nigeria2 players · 24 answers+3.6ptsNew
6🇪🇬Egypt1 player · 35 answers+3.3pts▼1Down 1 since yesterday
7🇧🇬Bulgaria1 player · 41 answers+3.2pts▼1Down 1 since yesterday
7🇮🇹Italy1 player · 103 answers+3.2pts▼1Down 1 since yesterday
9🇨🇦Canada1 player · 204 answers+3.0pts▼1Down 1 since yesterday
10🇮🇱Israel1 player · 45 answers+2.8pts▼1Down 1 since yesterday
Pick any logic or math game and your answers count for the country you play from.
Country comes from the network location the answers were sent from. Only the country is kept; addresses are never stored or shown.
Games: Mathematics, Sudoku, Logic Challenge, Brainy Digits, Math Matrix and Pyramid.
Every answer is compared with the average answer at the same game and level, so hard games and high levels do not count against anyone.
Each player counts once: their answers are averaged first, so one very busy player cannot carry a country. Players with only a few answers count as closer to average.
A country’s score is the average of its players, pulled toward the overall average when it has few players. It is shown in percentage points: +3 means its players answer right 3 points more often than an average player at the same games and levels.
A country needs at least 1 player to be ranked. Most active counts the rounds of these games played in the last 30 days.
Ranks are compared with the ranking about a day earlier. Equal scores share a rank.
Countries
Where is each country strong, and where can it grow?
Country profiles for math, logic, geography, science, art, words, memory and attention: speed, accuracy and timeouts compared with the average player, updated every hour.
Only 53% of answers in Twin cards are right, against 90% in Colors
Share of correct answers per game across 92,416 answers. Memory games (Twin cards, Cubes) and number puzzles (Math Matrix) sit at the bottom; quick perception games (Colors, Figures) and Word list sit at the top. In Sudoku each move counts as one answer, in Wordy each guess, in Twin cards each pair of cards turned.
Twin cards52.6%
Math Matrix53%
Wordy54.9%
Art59.3%
Flags62.6%
Capitals63.8%
Cubes65.6%
Word order67.3%
Chemistry67.7%
Logic Challenge72.2%
Sudoku73.9%
Pyramid76.8%
Brainy Digits83.6%
Mathematics85.1%
Figures88.5%
Word list88.8%
Colors89.6%
Accuracy depends on how many options a game offers (3 to 8) and on its time limit, so it is a measure of how hard the game feels, not of the players.
Based on 92,416 answersaccuracygames
How long one answer takes
A Math Matrix answer takes 22 seconds on median, a Twin cards answer 3
Median seconds between two consecutive answers of the same player in the same round. Memory games include the time the items are shown before the question, so part of their time is memorising, not deciding. Gaps longer than 120 seconds are treated as breaks and left out.
Math Matrix21.7 s
Logic Challenge13.6 s
Pyramid11.1 s
Wordy9.8 s
Cubes8.9 s
Sudoku8.9 s
Brainy Digits8.5 s
Word list8.4 s
Word order8.3 s
Art7.1 s
Capitals5.8 s
Chemistry5.2 s
Flags4.9 s
Mathematics4 s
Figures3.2 s
Colors3 s
Twin cards2.9 s
Measured from server timestamps, so network delay is included; the first answer of each round has no gap and is not counted.
Based on 85,987 answersspeedgames
Flags players struggle with
The flag of Gabon was named correctly in only 33% of 15 tries
Every flag asked at least 15 times, ranked by the share of correct answers. The easiest in this set is Sweden at 93%. Each question offers three country names, so random guessing would score about 33%.
Flag
Accuracy %
Answers
Timed out
Gabon
33.3%
15
0
Yemen
35.3%
17
0
Togo
35.3%
17
1
Austria
40%
15
0
Rwanda
40%
15
0
Luxembourg
43.8%
16
0
Bangladesh
50%
18
2
Sierra Leone
50%
16
3
Benin
53.3%
15
0
North Korea
53.3%
15
0
Portugal
56.3%
16
0
Belarus
58.8%
17
0
Showing 12 of 25 rows.
Small samples: each flag has 15 to 20 answers, so a few lucky or unlucky guesses move a flag several places. Wrong options are random, so there is no stable 'most confused with' pair.
Based on 405 answersquizflagsgeography
Famous paintings: the picture or the painter?
Players name the painting's title 70% of the time but its painter only 48%
The Art quiz asks two kinds of questions about the same pictures: what the painting is called and who painted it. Picking the title from three options is far easier than picking the artist. Based on 1,493 answers from 62 players.
Both question types offer three options, so chance is about 33%.
Based on 1,493 answersquizart
The most recognisable painters
Shown one of their paintings, players name Vincent van Gogh 80% of the time but Edouard Manet only 33%
Share of correct answers to 'who painted this?' when the right answer is each painter, for painters whose works were asked about at least 15 times. Three painters are offered each time, so 33% is chance level.
Vincent van Gogh79.7%
Pierre-Auguste Renoir73.7%
Rembrandt70.6%
Claude Monet62.5%
Pablo Picasso58.2%
Rene Magritte40%
Edouard Manet33.3%
Small samples per painter; treat the order as indicative. Wrong options are random, so no stable confusion pairs exist.
Based on 201 answersquizart
Chemistry: symbols are easy, properties are not
74% of players match a symbol like Fe to its element, but only 19% can match a physical property
The Chemistry quiz mixes question types. Matching a chemical symbol to its name is the common one and the easiest; questions that give the group, the atomic mass, the melting point, the density or the electron configuration are rare and much harder. Each question offers four options, so chance is 25%.
Property questions have few answers; they are grouped together so that the bucket is based on enough players.
Based on 989 answersquizchemistry
One more operation, more mistakes
Accuracy drops from 86% on a single addition or subtraction to 80% when a second step is added
Mathematics shows an equation such as 7+5 or 9-6+3 with ten seconds to pick the result. Single-step sums are answered right most of the time; chaining two operations costs several points of accuracy and more timeouts. Rounds with a second step appear at higher levels, so these players are also the more practised ones.
Multiplication and division appear only at higher levels and have far fewer answers than addition and subtraction.
Based on 8,723 answersaccuracymath
Wordy: how many guesses it takes
72% of Wordy rounds end with the word found, after 6 guesses on median
Wordy gives ten tries to find a five-letter word. Of 456 rounds by 168 players, 72% were solved; solved rounds took 6.4 guesses on average and 24% of them were cracked in three guesses or fewer. Unsolved rounds mostly end with the player leaving early rather than using all ten tries.
72.4%Rounds solved
6guessesMedian guesses when solved
23.9%Solved in 3 guesses or fewer
A guess is counted each time the server records an attempt; abandoned rounds count as unsolved.
Based on 456 answerswords
Sudoku, move by move
About 1 in 4 Sudoku moves is wrong, and the longest round ran to 1,180 moves
Sudoku records every placed digit. Across 32,791 moves in 1,117 rounds, 26.1% were wrong; a typical round takes 31 moves and 19% of rounds are played without a single wrong digit.
26.1%Wrong moves
31movesMedian moves per round
18.7%Rounds with no wrong move
1,180movesLongest round
Short rounds where the player left after a few moves are included in the median.
Based on 32,791 answerspuzzles
The first question is the one people miss
15% of first questions run out of time, against 4% from question 5 onwards
Accuracy by position of the question inside a round, for the games that ask 15 questions per round. Question 1 is answered right 67% of the time and question 15 84%. The gap is not fatigue but a slow start: the first question times out far more often, while the share of genuinely wrong picks stays roughly level through the round. Part of the rise is also that players who struggle leave early, so later positions are answered by those still in the game.
Sudoku, Wordy and Twin cards are excluded because their rounds have no fixed number of questions; the timer for the first question may also absorb loading time.
Based on 52,257 answersaccuracyfatiguepressure
The sweet spot is a few seconds
Answers given in 2-4 s are right 94% of the time; take 60-120 s and it drops to 60%
Accuracy by how long the player took, measured as the gap since their previous answer in the same round, counting only answers where the player actually picked an option. Taps within 0-2 s are mostly guesses (33% right); after that, the longer the pause, the lower the accuracy: long gaps are hesitation on a question the player cannot solve. Games with different time limits are mixed here, so part of the effect is which games allow long thinking.
Sudoku, Wordy and Twin cards excluded; timeouts, skipped questions and the first answer of each round (no gap) are left out; gaps over 120 s are treated as breaks.
Based on 43,417 answersspeedaccuracy
Higher level, steadier hands
Rounds at level 1 are answered right 73% of the time; rounds at level 3+ 92%
Accuracy of answers by the level of the round being played. A game's level only rises after wins, so the players at higher levels are the ones who already answer well, and the questions there are no easier. Only a small group has climbed that far.
Sudoku excluded (moves, not questions). This is selection, not training: the level is earned by answering well, so it does not show that playing more makes anyone better.
Based on 59,625 answersaccuracylevels
Beaten by the clock
9.0% of all answers run out of time; in Sudoku it is 15%
Share of answers where the timer ran out before the player chose anything. Sudoku gives 30 seconds per move and Chemistry 10 seconds per question, and both see the most timeouts; quick perception games almost never do.
Sudoku15.1%
Math Matrix14.3%
Chemistry14%
Logic Challenge8.8%
Wordy8.7%
Mathematics7.7%
Capitals7.5%
Word order7%
Cubes5.4%
Colors4.6%
Pyramid3.5%
Figures3.2%
Flags3.1%
Word list2.9%
Brainy Digits2.2%
Art2%
Twin cards0.9%
A timed-out answer also counts as a wrong answer in the accuracy cards.
Based on 92,416 answerspressuregames
Head to head
Only 2.6% of rounds are played against another person, and 24% of those end in a draw
Most rounds are solo. In the 104 rounds between two real players, the player who started the challenge won 49% of the decided games and the invited player 51%: starting the duel is no advantage.
2.6%Rounds played against someone
24%Draws in two-player rounds
49.4%Challenger wins (decided rounds)
Two-player rounds where the opponent was a test or bot account are counted as multiplayer but excluded from the draw and win figures.
Based on 7,094 answersmultiplayer
Who comes back
22% of players return to play a week or more after their first round
Of the players who played at least one round, 37% played on two or more different days, 22% were still playing seven or more days after they started and 13% a month or more later. The median player has played 5 rounds; the top tenth 22 or more, and the most active player 351.
37%Played on 2+ days
21.7%Back after 7+ days
13.1%Back after 30+ days
27.5%Played 10+ rounds
5roundsMedian rounds per player
Players who joined recently have had less time to come back, so the return shares are lower bounds.
Based on 7,094 roundsretention
Climbing the levels
21% of players have reached level 3 or higher, 8% level 5 or higher
Each player has a single-player level that rises with wins. Most players are still at the first steps; the top bucket, level 10 and above, holds 9 players.
Levels 5 to 9 and 10 and above are grouped so that every bucket has at least five players.
Based on 7,094 roundslevels
When people play
Play peaks at 18:00 UTC with 6.0% of all rounds, and is quietest at 04:00
Rounds started per hour of the day, in UTC. Most addresses in the server logs are European, so an evening peak at home lands in the late afternoon UTC. Accuracy per hour is shown alongside and barely moves through the day.
Times are server time (UTC), not the player's local time; accuracy excludes Sudoku moves.
Based on 7,094 roundstime
Which day people play
Sunday is the busiest day with 16% of rounds, Friday the quietest with 12%
Rounds by day of the week (UTC). Weekdays average 999 rounds per day of the week versus 1050 on weekend days. Accuracy per weekday is shown alongside.
Server time (UTC); a late evening round in the Americas lands on the next UTC day.
Based on 7,094 roundstime
Rounds and players per month
7,094 rounds since 2024-02, with the busiest month (2024-04) at 568 rounds
Rounds started and distinct players per month since the first recorded round. The current month is partial. Months with fewer than five players are left out.
Only rounds by real player accounts; test and QA accounts are excluded.
Based on 7,094 roundsactivity
Where players connect from
United States accounts for 13% of the addresses that played in the last two weeks, from 33 countries in all
Unique network addresses that called the game API in the nginx logs kept on the server (about two weeks), located to a country with the free DB-IP database. 94 addresses from 33 countries; countries with fewer than five addresses are grouped as Other.
🇺🇸 United States13%
🇪🇸 Spain9%
🇹🇭 Thailand8%
🇵🇱 Poland8%
🇧🇾 Belarus7%
🇷🇺 Russia7%
🇷🇴 Romania6%
🇦🇹 Austria6%
🌐 Other38%
Addresses are not linked to accounts or games, so this says where players connect from, not how each country plays; one person can appear from several addresses and VPNs shift countries.
Based on 94 addressesgeography
The new instant games, first days
Schulte Table is the most opened of the new browser games, with 73 sessions out of 738 so far
Anonymous usage tracking for the browser-only games started on 2026-11-08. In that time 738 sessions were opened, adding up to 47.3 hours of active play; 86% of started runs were played to the end and 9% were abandoned. Games with fewer than five sessions are not listed.
Schulte Table73 sessions
Chess vs Computer61 sessions
Honeycomb57 sessions
Arrow Escape39 sessions
Chess Puzzles31 sessions
Robot Route29 sessions
Echo Match29 sessions
Potion Sort26 sessions
Shell Game26 sessions
Shoal25 sessions
Rail Dispatch24 sessions
Pipe Rush22 sessions
Shop Keeper21 sessions
Station Rush20 sessions
Goalkeeper20 sessions
Shape Spin18 sessions
Lighthouse17 sessions
Switcheroo17 sessions
Mahjong Pairs16 sessions
Knit Unwind16 sessions
Bolt Out16 sessions
Laser Mirrors15 sessions
Sum Blocks14 sessions
Town Folk14 sessions
Market Rush12 sessions
Fresh Pick11 sessions
Gate Slide10 sessions
Word Hunt10 sessions
Letter Wheel10 sessions
Treasure Tower9 sessions
Pixel Blaster7 sessions
Bubble Pop6 sessions
Grid Clear5 sessions
Hive Stack5 sessions
Very early data from a few days of tracking; sessions are not tied to accounts, so one player can open several sessions.
Based on 738 sessionsbrowser-gamesnew
Answers that look alike
Not enough data yet: this card fills in as the knowledge games are played
This card will show which pairs of options in the knowledge games players mix up most often. A pair is listed once the two were offered together at least 20 times and one was picked for the other at least 5 times by at least 5 different players.
Questions in the new knowledge games are in English only for now (beta). Pairs of two countries are not listed. Only real player accounts; no names or accounts are shown.
quizknowledgebeta
Updated November 11, 2026. Findings describe patterns in game data from people who chose to play; they are not a clinical study and do not measure or claim any medical effect. Questions about the methods: contact us.
Add your own rounds to the dataset
Play any game and you are part of the next update. Everything stays anonymous.
