Pymetrics-Style Games

How it is scored

Each game pays points of its own: Towers by how close to the fewest moves you got, Cards by where the balance finished, Keypresses by presses, Easy or Hard by what you earned, Lengths, Magnitudes and Stop by right answers (Stop takes off a point for every press on green), and Money Exchange a fixed amount for finishing, because trust has no right answer. Around 189 is a strong run.

The points are ours. The real battery never shows a score: it turns the same games into a trait profile, matched against people already doing the job, and there is no pass mark. Read the line under each game at the end rather than the total.

Where we had to choose

What is published: twelve games in twenty to twenty five minutes, played in a browser; the result is a trait profile rather than a mark; it is kept for twelve months and shared with every other employer on the platform, so you sit it once a year, not once a firm. The game list (Balloons, Keypresses, Digits, Arrows, Lengths, Cards, Towers, Money Exchange 1 and 2, Easy or Hard, Stop, Faces, and in some batteries Magnitudes, Sequences, Shapes and Letters) comes from candidate reports and prep sites, and is consistent between them.

What is not published: trial counts, clocks, deck schedules, reward amounts, or how any trait is derived. Every number here is ours: Towers uses the classic peg sizes, Cards uses the amounts from the original Iowa gambling task, Easy or Hard the three probabilities from the original effort task. Balloons, Digits and Arrows are already on the site (linked below) so they are not repeated, and Faces needs photographs of real expressions, so it is not built.

What each game is believed to measure

• Towers. Planning: whether you work the route out before the first move or discover it by moving. Plan before the first move. The clock is generous and the moves are not.
• Cards. Learning from outcomes: whether the running balance overrides the size of the wins. Switch away from a deck after a big loss, and trust the balance over the win size.
• Keypresses. Processing speed and effort: raw motor pace, and whether it holds up in the second round. Just go. There is nothing to work out.
• Easy or Hard. Effort against reward: whether the chance of payout changes how hard you are willing to work. Pick hard when the chance is high and the reward is big, easy when the chance is 12%.
• Lengths. Learning from feedback, and whether a reward changes your attention. Read the feedback. The first few are guesses and the rest are not.
• Magnitudes. Number sense: judging quantity at a glance, and comparing fractions without a pen. On dots judge the share, not the count. On fractions cross multiply if in doubt.
• Stop. Impulse control: withholding a response once responding has become the habit. Accuracy over speed. A press on green costs a hit.
• Money Exchange. Trust and fairness, and how you send after a partner keeps most of it. There is no right answer. Send what you would send.
• Balloons. Risk against expected value. See Battery: the expected value peaks about halfway up the range.

None of this is a way to game a trait. A profile is matched to a role, and the trait a bank wants in an analyst is not the one it wants in a trader. The honest gain from practice is smaller and real: no game is new to you when its clock starts.

How to work through one

Read each brief properly, because the real one does not repeat it and the game does not pause. Sit it once at a normal pace rather than rushing: the battery is not timed as a whole, only the games inside it are. Use a keyboard where you can (space to respond, arrows for left and right, number keys for pegs and decks), and on a phone use the big panel. When it finishes, read the readouts and run it again only if a game's rules still surprised you.