pymetrics Games Answers: Why There Is No Answer Key

Search for the pymetrics games and you soon find "answers", "optimal strategies" and long forum threads about how many times to pump a balloon. This guide explains why the pymetrics games have no answer key, what they measure instead, why copied strategies can backfire and what honest preparation looks like. It also covers the four extra reasoning games, which are the exception.

The short answer

There are no pymetrics games answers to learn. pymetrics says its core games have no right or wrong answer, your play is compared with the employer's own high performers rather than with a key, and only your first attempt counts. What you can learn is how each game works, so the rules do not catch you out.

Why the pymetrics games have no answers

pymetrics says so

An employer-hosted copy of pymetrics' 2018 candidate FAQ says there is no right or wrong answer and that the games are not an IQ test. Each game is a short version of a psychology experiment, such as the Balloon Analogue Risk Task or a digit span test, and records how you play rather than marking you right or wrong.

You are compared with people in the job, not with a key

pymetrics' patent and a 2021 research paper co-written by pymetrics staff describe the model. For each job, an "in group" of 50 to 100 of the employer's high performers plays the games, and their results are compared with an "out group" drawn from other players. Your play is then given a fit score for that job, from 0 to 100%.

Two things follow. The "right" way to play depends on who does well in that particular job, which no outsider can see. And the same way of playing can fit one employer's model well and another's less well.

Only your first results count

The same 2018 FAQ said you could replay games, but only your first results would be used. pymetrics' terms of service, last updated in May 2019, are stricter: you may play each game only once in any 12-month period, you must not stop a game halfway to repeat it, and you must not create more than one account. Prep sites add that results are reused across employers for about 330 days; no official source confirms that figure.

The terms rule out the shortcuts

The same terms ban bots and other automated access, and any attempt to reverse engineer the site. A tool that promises to play the games for you, or to reveal how they are scored, breaks rules you accept when you create an account.

What the games do measure

Harver, which has owned pymetrics since August 2022, groups the traits into nine categories: attention, decision making, effort, emotion, fairness, focus, generosity, learning and risk tolerance. The research tasks behind the 12 core games suggest where each comes from:

• Risk and learning from losses: Balloons and Cards.
• Trust, generosity and fairness: Money Exchange 1 and 2.
• Effort: Hard or Easy.
• Speed and steadiness: Keypresses.
• Attention and impulse control: Stop and Arrows.
• Memory: Digits.
• Reward learning and fine detail: Lengths.
• Planning: Tower.
• Emotion: Faces.

This grouping is our reading of the research tasks, not a mapping pymetrics publishes. Our free pymetrics test guide describes every game, and the game guides cover Arrows, Faces, Cards and the rest.

Why "optimal strategies" can backfire

Prep sites and forums publish strategies such as an ideal number of pumps per balloon or a "safe" amount to send in Money Exchange. In our view they can do more harm than good, for three reasons.

• The target is different for every job. A strategy can only aim at one profile, but each employer's model is built from its own high performers. Cautious play might suit one role and count against you in another.
• Forced play is hard to keep consistent. If you play one game by a memorised rule and the others naturally, your profile reflects the rule rather than you, and you spend attention on the rule instead of the game.
• Some strategies rest on the wrong rules. Prep sites disagree on basic details, such as 39 or 24 balloons and 90 or 135 rounds of Arrows, and a video walkthrough reports 30 to 40 card draws where most prep sites say 80. Some describe Tower with a Tower of Hanoi rule about disc sizes, although pymetrics' patent describes a Tower of London task in which coloured discs are rearranged to copy a target. One prep site describes a completely different Keypresses game from all the others.

Even in games where one choice pays more on average, such as Balloons or Cards, pymetrics does not say that earning the most money is the goal.

What pymetrics Reddit threads can and cannot tell you

We could not access Reddit when we researched this article, so we do not quote any threads here. Forum reports can still help with practical points, such as how long the games took or which device worked well. What they cannot tell you is what a particular employer's model rewards, because the comparison group is that employer's own staff. Anyone who claims to know is guessing.

What honest preparation looks like

1. Learn the rules of every game. Knowing that Stop asks you to hold back on green, or that the colour of the arrows changes which ones you judge, is fair preparation, and it prevents mistakes you cannot undo.
2. Get used to the pace. Several games depend on quick, accurate key presses. Prep sites advise a desktop or laptop with a keyboard.
3. Play the behavioural games as yourself. Balloons, Money Exchange, Hard or Easy and Cards have no target number.
4. Concentrate where accuracy counts. Digits, Stop, Arrows, Lengths, Tower and Faces include responses that can be right or wrong, such as a correct digit list or a press on the wrong colour.
5. Set aside a quiet half hour. pymetrics says the 12 games take about 25 minutes, and you cannot pause in the middle of a game.
6. Ask about adjustments early. pymetrics' research paper mentions built-in accommodations for colour blindness and dyslexia, and the games come in more than 20 languages.

The four extra reasoning games do have right answers

Some employers add an extra suite of numerical and logical games, which pymetrics' research paper confirms. Prep sites name four of them:

• Magnitudes: pick the larger of two quantities, such as two fields of dots or two fractions (14 items, about 5 seconds each).
• Sequences: find the missing number in a series (12 items, 30 seconds each, according to one prep site).
• Shapes: find a small shape hidden in a complex figure, without rotating it (10 items, 45 seconds each).
• Letters: a 2-back memory task, where you respond when a letter matches the one two letters back (45 letters, according to one prep site).

These games have correct answers, so practising the formats helps in the usual way: you learn to estimate quickly in Magnitudes and to keep track of the last two letters in Letters. pymetrics does not publish the items, though, so a list of "answers" for these games cannot be checked either.

Practising the formats

Candidate Falcon's pymetrics practice has independent simulations of all 16 games, rebuilt from the research tasks in pymetrics' public patents and from prep-site descriptions, with original art and wording. They are not the real games, they contain no real items or answers, and the review after each game describes how you played in our own words rather than giving a pymetrics score. The games come with the Harver platform in our plans; see pricing.

FAQs

Are there answers to the pymetrics games?

No. pymetrics says its core games have no right or wrong answer, and your play is compared with the employer's own high performers. Only the four extra reasoning games have correct answers, and their items are not published.

What do people on Reddit say about pymetrics?

We could not access Reddit for this article. Forum reports can help with practical points such as timing and devices, but they cannot reveal what a particular employer's model rewards.

Can I replay a pymetrics game to get a better result?

No. pymetrics' 2018 candidate FAQ said only your first results are used, and its terms allow each game once in any 12-month period and ban stopping a game to repeat it.

Is it worth learning a strategy for the balloon game?

Learning how Balloons works is worth it; see our pymetrics balloon game example. A fixed number of pumps copied from the internet is not, because there is no target number and each employer's model is built from its own staff.

Sources

• pymetrics patent WO2015200880A1: the games, the comparison with high performers and the fit score.
• Wilson et al., "Building and Auditing Fair Algorithms", FAccT 2021 (pymetrics co-authors): the 12 games, the extra reasoning suite, languages and accommodations.
• pymetrics candidate FAQ, employer-hosted copy (2018): no right or wrong answer, not an IQ test, first results only, about 25 minutes.
• pymetrics terms of service (last updated May 2019, read October 2026): one play per game in 12 months, no restarting a game, one account, no bots or reverse engineering.
• Harver: gamified assessments and Harver acquires pymetrics (read October 2026): trait categories and ownership.
• Prep sites, for game details and the reuse window: JobTestPrep, MBB Prep, GraduatesFirst and Psychometric Success.

Candidate Falcon is independent and not affiliated with pymetrics or Harver.