The pymetrics test is a set of 12 behavioral games, now on Harver's platform, that measure traits like risk tolerance and attention. See how to prepare.
Dr. Russell T. WarneChief Scientist
Share
The pymetrics test is a game-based behavioral assessment: a series of 12 short online games that record how you behave (how long you wait, how much risk you take, how you divide effort) rather than whether you pick right answers. Candidates typically complete the games in about 25 minutes on the web or a phone. The assessment was built by the company pymetrics, which was acquired by the hiring platform Harver in 2022, and employers still commonly call the exercise "pymetrics." This article explains where the games came from, what they measure, how employers use the results, and how to approach them.
Who owns the pymetrics test
Harver announced its completed acquisition of pymetrics on August 11, 2022, calling it "the leader in unbiased soft skills assessments" and noting that millions of job candidates around the world had gone through the pymetrics platform, with technology "compliant in 100+ countries, 30 languages and across web, Android, and iOS apps." After the acquisition, Harver reported over 100 million candidates processed and more than 1,300 customers across its combined platform. The games now run as Harver's gamified behavioral assessments, but the pymetrics name has stuck with candidates and recruiters alike.
Format: what the games are like
The Oxford University Careers Service, a useful non-vendor source, describes the Pymetrics Games as "12 online games used as an assessment tool for major employers worldwide" that together "measure over 90 cognitive, social, and behavioural traits." Harver's own product page advertises "12+ interactive, gamified experiences" and says candidates complete them "in approximately 25 minutes," receiving immediate results and feedback at the end.
The traits fall into nine categories, per both Oxford and Harver: attention, decision making, effort, emotion, fairness, focus, generosity, learning, and risk tolerance. The Oxford guide walks through each of the 12 games; a commonly described example is a balloon-pumping task where each pump earns money but risks popping the balloon, so when you choose to cash out says something about your risk tolerance. Most of the games work this way: the behavior itself is the data.
That design differs from a personality questionnaire like the Hogan assessment, which asks you to describe yourself, and from a timed cognitive screener like the Predictive Index cognitive assessment, which scores right answers per minute. The games watch what you actually do.
How employers use the results
Harver says each assessment "captures thousands of data points" that feed "profiles of what makes a person and job unique." The result is a trait profile matched against a job-specific model, not a single pass-fail score. Test-taker guides report that candidate profiles are compared against benchmarks built from strong performers already in the role, though that detail comes from prep communities rather than the vendor. Either way, the same profile can fit one role well and another poorly, so "failing" pymetrics is the wrong mental model.
According to Oxford's careers service, the games are used by "major employers in the world, including JP Morgan, BCG, Morgan Stanley, AstraZeneca, and more," typically early in high-volume graduate and professional hiring funnels.
Fairness claims and the published audit
Algorithmic hiring tools attract fair scrutiny, and pymetrics is unusual in having a published audit. A 2021 paper at the ACM Conference on Fairness, Accountability, and Transparency presented "a framework for algorithmic auditing by way of a case-study of pymetrics" and reported the results of an audit of its candidate screening tool, including how the software implements adverse impact testing. One honest caveat: the author list combines outside auditors from Northeastern University with pymetrics employees, so it is best described as a collaborative audit rather than a fully arm's-length one. Harver's marketing adds that the assessments are "validated for fairness across gender, ethnicity, and socioeconomic status," a vendor claim that should be read as such.
Is the pymetrics test an IQ test?
No. The games measure behavioral traits, and a trait profile is a different kind of measurement from a normed IQ score. Several games do tap cognitive functions such as attention, memory, and reaction speed, which is why the comparison comes up, but sampling those behaviors in a game is not the same as measuring the underlying abilities with normed subtests. In the Cattell-Horn-Carroll framework that modern intelligence testing is built on, a full battery like the RIOT IQ test (the Reasoning and Intelligence Online Test) measures six indices directly: Verbal Reasoning, Fluid Reasoning, Spatial Ability, Working Memory, Processing Speed, and Reaction Time. You can try that approach with the free RIOT IQ test.
Why do employers screen on cognition at all? Schmidt and Hunter's review of 85 years of selection research found general cognitive ability among the strongest predictors of job and training performance, and a 2022 reanalysis by Sackett and colleagues, while revising validity estimates downward, still ranked it among the more useful tools. Behavioral games and ability tests answer different questions about the same candidate, which is why some employers use both.
Can you prepare for the pymetrics test?
Only up to a point, and the honest advice is mostly about conditions rather than answers. The classic practice-effect evidence (a meta-analysis finding roughly a quarter of a standard deviation gain from retesting) covers cognitive ability tests, and no equivalent meta-analysis exists for behavioral games, so treat prep claims modestly.
• Play rested and focused: The games measure attention and effort in real time, so fatigue shows up in the data in a way it would not on an untimed questionnaire.
• Use one quiet sitting and a reliable device: The assessment runs on web, Android, and iOS; pick whichever is most stable and finish in one session.
• Do not try to game it: Oxford's guidance notes the games "monitor your behaviour" throughout, and traits are matched to a job model rather than scored as good or bad, so playing a character mostly adds noise.
• Learn the common paradigms: Knowing that a balloon game rewards calibrated risk or that a keypress game measures effort keeps surprise from distorting your first attempts.
And if you want a normed measure of your cognitive abilities to sit alongside a behavioral profile, the RIOT IQ test at riotiq.com is normed on adults 18 and older and publishes a margin of error for every version.
Frequently asked questions
Can you fail the pymetrics test?
Not in the usual sense. The games produce a trait profile that is matched against a job model, and the same profile can fit one role and not another. There is no universal passing score.
How long does the pymetrics test take?
Harver says candidates complete the gamified assessments in approximately 25 minutes, with immediate results and feedback at the end.
Does pymetrics still exist?
The company was acquired by Harver in August 2022, and the games continue as Harver's gamified behavioral assessments. Candidates and employers still widely use the pymetrics name.
Are there right answers in the pymetrics games?
Mostly no. The games record behavior, such as how much risk you take or how you allocate effort, and traits are interpreted relative to a job model rather than as correct or incorrect choices.
3. Hausknecht, J. P., Halpert, J. A., Di Paolo, N. T., & Moriarty Gerrard, M. O. (2007). Retesting in selection: A meta-analysis of coaching and practice effects for tests of cognitive ability. Journal of Applied Psychology, 92(2), 373-385. doi.org/10.1037/0021-9010.92.2.373
4. Oxford University Careers Service. (n.d.). The Pymetrics Games: Overview and practice guidelines. careers.ox.ac.uk
5. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040-2068. doi.org/10.1037/apl0000994
6. Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124(2), 262-274. doi.org/10.1037/0033-2909.124.2.262
7. Wilson, C., Ghosh, A., Jiang, S., Mislove, A., Baker, L., Szary, J., Trindel, K., & Polli, F. (2021). Building and auditing fair algorithms: A case study in candidate screening. Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency. doi.org/10.1145/3442188.3445928
