Executive Function
Pond Stones Game
Also called: trail making game, number and shape switching game, connect the numbers game
Measure how many stones adults connect in the first minute of Part B, where numbered stones are touched in order and the kind of stone alternates between round and square.
6-9 minTimed: desktop or laptop with a mouse or trackpad
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 3,944
- Evidence grade
- A
- Registered clinical studies
- 2,293
Figures for the Trail Making Test B, the published test this game is compared with.
How these are countedWhat it measures
A timed game for research on cognitive flexibility in adults, older adults first. Stones with numbers lie in a pond. In Part A the player touches the stones from 1 to 25 in order. In Part B each number is on a round and a square stone, and the player alternates the two kinds. The score counts the right connections in the first minute of Part B. It is a raw research measure with no percentile.
Constructs
- Cognitive flexibility
- Switching between two kinds of stone
- Visual search and sequencing
Research uses
- Research on processing speed and cognitive flexibility in adults, older adults first
How the paradigm works
The Pond Stones Game is a timed game for research on cognitive flexibility in adults, older adults first. Stones with numbers lie in a pond, drawn in code, and no stone carries a letter. In Part A the player touches the stones in number order, 1 to 25. In Part B each number from 1 to 25 is on a round stone and on a square stone, and the player alternates the two kinds: 1 on a round stone, 2 on a square stone, 3 on a round one, and so on. The other kind's stone with the same number is not the next stone. The main score is the number of right connections in the first minute of Part B, with Part A as its control. A count of this kind from the first minute correlated with the time of the whole part in a published study of older adults (Zhao and colleagues 2013). Each part has a practice of 5 stones first, with a ring on the right stone after a wrong touch, and the practice runs once more after more than 2 wrong touches. A touch counts only if it starts on a stone, so a palm or a miss is counted and never blocks the next touch, and a touch within 250 ms of the last one is ignored and counted. The stones, the layouts and the words here are our own.
More about the paradigm
The paper counts and grades on this page are those of the Trail Making Test B, the published test the game is compared with, not the game's own. The validation plan also compares the game with the Trail Making Test A, against which Part A's time is checked. The game has six alternate forms whose path length, crowding and crossings are equal by construction, and they are not pooled until a check on first sessions shows them equally difficult. Try it now shows a seventh form that no study uses.
The game is built for repeated visits. Add a run-in session before the baseline, give each person the six forms in their own order, keep one device class and one input at every visit, and run a retest as a second visit of the same study 7 to 28 days after the baseline. It takes a mouse, a finger or a pen and has no keyboard route. It runs on computers and touch-screen computers now, and on tablets once the timing check passes. A mouse, a finger and a pen do not move alike, so counts and times are compared within an input type, and the first desktop data are the first mouse data. Its reliability and norms come from use.
Scores
- Right Connections in the First Minute of Part Bcount
- The stones connected correctly in Part B in the 60 seconds after its first frame, by each touch's own time stamp. A connection stamped below 60,000.0 ms counts, and one at or after it does not. Empty when Part B has no scored row.
- Higher means more stones connected in the minute; compared within a study and within an input type
- Right Connections in the First Minute of Part Acount
- The same count for Part A.
- Control: the speed of touching stones in order, with no switch between kinds
- Part A Timeseconds
- Seconds from Part A's first frame to its 25th connection. Empty when Part A was not finished within 300 seconds.
- Secondary: lower is faster
- Part B Timeseconds
- Seconds from Part B's first frame to its 25th connection. Empty when Part B was not finished within 300 seconds.
- Secondary: lower is faster
- Part A Wrong Touchescount
- Touches on a stone that was not the next one in Part A.
- Quality check
- Part B Wrong Touchescount
- Touches on a stone that was not the next one in Part B, of either kind of mistake.
- Quality check
- Part B Wrong-Kind Touchescount
- Touches in Part B on the right number on the other kind of stone.
- Secondary: the switch between round and square shows here
- Part B Wrong-Number Touchescount
- Touches in Part B on any other stone than the next one and its copy on the other kind.
- Secondary
- Switch Ratioscore
- Part B's time over Part A's time, when both parts were finished.
- Informational: a larger ratio means Part B took longer compared with Part A
- Part B Finished0/1
- 1 when Part B reached its 25th connection within 300 seconds, 0 when it did not.
- Quality check
- Repeat Touchescount
- Touches within 250 ms of the last accepted touch, which are ignored.
- Quality check: a double touch lands here
- Extra Contactscount
- Second contacts on a stone while one was held, and mouse buttons other than the main one, which are ignored.
- Quality check
- Touches off the Stonescount
- Contacts that started on no stone, such as a palm or a miss, and touches within 100 ms of a part's first frame, which are ignored.
- Quality check
- Share of Touch Parts%
- Share of the scored parts whose first accepted touch was a finger, not a mouse or a pen.
- Informational: costs are compared within an input type
Use and evidence
3,944 papers from 2021 to 2025 name this task in their title or abstract, up 22% compared with 2016 to 2020. That ranks it 5 of 711 test families in our census. These counts belong to the Trail Making Test B, the published test this game is compared with. This game is a new measure, so its own reliability and norms come from use. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 944 |
| 2024 | 803 |
| 2023 | 841 |
| 2022 | 648 |
| 2021 | 708 |
| 2020 | 720 |
| 2019 | 667 |
| 2018 | 646 |
| 2017 | 627 |
| 2016 | 576 |
| 2015 | 595 |
| 2014 | 496 |
| 2013 | 449 |
| 2012 | 421 |
| 2011 | 389 |
| 2010 | 330 |
| 2009 | 273 |
| 2008 | 274 |
| 2007 | 199 |
| 2006 | 208 |
| 2005 | 138 |
| 2004 | 98 |
| 2003 | 102 |
| 2002 | 90 |
| 2001 | 85 |
| 2000 | 92 |
Use grade A
1,000 or more papers from 2021 to 2025.
Evidence grade A
At least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome.
- Papers on test-retest reliability
- 161
- Papers on reliability or validity
- 1,323
- Registered clinical studies using it as an outcome
- 2,293
Measured on ConductCognition
Pooled across the studies that feed our platform norms, by device group. A figure needs 50 or more people.
No pooled figure yet.
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on the Trail Making Test B, the published test this game is compared with, not the game's own.
Source: ConductCognition test census 0.58.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Stones
- Round and square stones drawn in code, each with a number from 1 to 25 and no letter. A round stone is a circle with a smooth face, and a square stone is a rounded square with two ridges. Both have one sand fill, one dark edge and one dark number, so colour tells nothing and no colour vision is needed. The game look adds a pond behind the stones, the same for every part; the plain look has a flat ground. Both looks draw every stone the same way.
- Part A and Part B
- Part A has 25 round stones, touched in number order. Part B has 50 stones: each number from 1 to 25 is on a round stone and on a square stone, and the person alternates the kinds, 1 on a round stone, 2 on a square one, 3 on a round one, and so on. The other kind's stone with the same number is not the next stone. Stone 1 carries a small flag, and no other stone is marked.
- Layouts and forms
- Six alternate forms and a seventh for Try it now, each with its own layouts for both parts. The layouts are equal in path length, crowding and crossings by construction, no two forms are turns of one another, and no path sweeps across the pond. Each participant gets the forms in an order drawn once for them, and each session takes the next form after the sessions they began before it. The forms are not pooled until a check on first sessions shows them equally difficult.
- Touches
- A touch counts only if it starts on a stone, so a palm, a knuckle or a miss is counted and never blocks the next touch. A second finger on a stone while one is held is counted and ignored, and so is a touch within 250 ms of the last one. A touch on a stone already connected is ignored and not counted. A wrong stone is counted as an error, by kind or by number. A part with a hidden tab or a resized window is left out.
- Marks
- A connected stone fills in, and a line joins it to the one before. A wrong stone is outlined for 300 ms. Nothing else moves in a timed part, no line runs ahead, and no count, time or score shows. A part ends at its last stone or 300 seconds after it opens, and the first minute is counted from each touch's own time stamp.
- Practice
- Each part has a practice of 5 stones first, with a ring on the right stone after a wrong touch. The practice runs once more when it was not finished or after more than 2 wrong touches.
- Scores
- The stones connected in the first minute of Part B (the main score), the same count for Part A (its control), the time of each part, the wrong touches (Part B's by kind), the switch ratio, whether Part B was finished and the counts of repeat, extra and stray touches. The split-half is the six 10-second segments of that minute, odd against even.
- Window and devices
- The pond scales from 1.5 down to 0.85 and turns a quarter on an upright tablet, so every form keeps its layout on every device. A stone is 56 px at scale 1 and 47.6 px at the smallest, and older-adult mode changes the words and the buttons, never the stones. A stage (the window less the progress bar) under 714 by 510 px (510 by 714 upright) is held with a line until it fits. The game runs on desktops, laptops and touch-screen computers at release, with the input recorded. Tablets open once two filmed tablets pass the device timing check, and phones do not run the game. It takes a mouse, a finger or a pen and has no keyboard route.
- Age and use
- For adults, older adults first (ages 50 and over). Counts and times are compared within an input type, since a mouse, a finger and a pen do not move alike, and the first desktop data are the first mouse data. Keep each participant on one device class and one input at every visit.
- Schedule
- Built for repeated visits: a run-in session before the baseline, one session at each visit, and for a retest a second visit of the same study 7 to 28 days after the baseline. The six forms come in each person's own order. The evidence plan reports the split-half and the retest from the stones connected in a minute, the practice curve over the first sessions, and whether each form is within 0.3 standard deviations of the others.
The game takes the paradigm of connecting numbered targets in order and, in Part B, alternating two kinds, as the trail making task does and as a published variant does that shows each number in a circle and in a square (Zhao and colleagues 2013). The stones, the layouts, the touch rules, the words and the looks are our own, and no letter is used. This is a new measure for research with adults, older adults first, built for repeated visits: its reliability and norms come from use, the figures published in those studies are not its own, and it computes no percentiles. Use a mouse, a finger or a pen for the whole study, and compare counts and times within one input type.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: pond_stones_game test parts A and B, each once, from its first row (a part with a hidden tab, a resized window or a row of the wrong shape is left out, and so is a run holding two forms). A part with at least 12 touches of which under a quarter were right invalidates. Fewer than 3 connections in Part B, a part not finished within 300 seconds, and extra contacts and stray taps on over 10% of the touches or over 20 repeat touches are flagged. No response time is scored, so the fast-response rule is not used.
Why these rules
A touch on a random stone is the next stone about 1 time in 25 in Part A and 1 time in 50 in Part B, so a person who was connecting the stones is right on far more than a quarter of the touches, and a run that is not is touching without following the numbers; the line is ours. A connection's time is a stamp from the part's first frame, so no response time is scored and a response under 100 ms cannot enter a score. A part during which the tab was hidden or the window changed size was not played as the others were, so it is left out. Counts of stones in one minute need Part B to have started: very few connections mean the count rests on little. A part not finished at the 300 second cap is the plain test's discontinue case, so it is flagged and not invalid. A palm or a second finger on the screen, or a double touch, can change which stones were touched, so a high share of them is flagged for a person to read beside the count.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Look | Game | Game, Plain |
Every plan can change the look, which never changes how a result is scored.
Need a setting this test does not have?
Devices and languages
- Timed test: participants need a desktop or laptop with a mouse or trackpad. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Pond Stones Game in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Reitan, 1958), with the standard settings (look: Game). The main score was right connections in the first minute of part B (the stones connected correctly in Part B in the 60 seconds after its first frame, by each touch's own time stamp).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Zhao, Q., Guo, Q., Li, F., Zhou, Y., Wang, B., & Hong, Z. (2013). The Shape Trail Test: Application of a New Variant of the Trail Making Test. PLoS ONE, 8(2), e57333. https://doi.org/10.1371/journal.pone.0057333The paradigm: Part B shows each number from 1 to 25 twice, once in a circle and once in a square, and the person alternates the two shapes; the number connected in the first minute of Part B correlated well with the time of the whole part (full text read). The game takes the paradigm and the one-minute count, not the figures or the shapes.
- Reitan, R. M. (1958). Validity of the Trail Making Test as an Indicator of Organic Brain Damage. Perceptual and Motor Skills, 8(3), 271-276. https://doi.org/10.2466/pms.1958.8.3.271The origin of the trail making task, cited for the paradigm only: the paper itself was not read, and Crossref and OpenAlex give no abstract.
Source papers
Questions
How long does the Pond Stones Game take?
About 6 to 9 minutes, as listed in the test library.
What device does a participant need?
Timed test: participants need a desktop or laptop with a mouse or trackpad. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Right Connections in the First Minute of Part B, Right Connections in the First Minute of Part A, Part A Time, Part B Time, Part A Wrong Touches, Part B Wrong Touches, Part B Wrong-Kind Touches, Part B Wrong-Number Touches, Switch Ratio, Part B Finished, Repeat Touches, Extra Contacts, Touches off the Stones and Share of Touch Parts. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Trail Making Test B used in research?
These counts belong to the Trail Making Test B, the published test this game is compared with. This game is a new measure, so its own reliability and norms come from use. 3,944 papers from 2021 to 2025 name the task in their title or abstract, up 22% compared with 2016 to 2020. That ranks it 5 of 711 test families in the ConductCognition test census (use grade A). Source: ConductCognition test census 0.58.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Trail Making Test B have?
These counts belong to the Trail Making Test B, the published test this game is compared with. This game is a new measure, so its own reliability and norms come from use. In the published research, 161 papers report the task's test-retest reliability and 2,293 registered clinical studies use it as an outcome (evidence grade A in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Look (Game, Plain). Every plan can change the look, which never changes how a result is scored.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.
Related tests