pymetrics Towers Game: Complete Practice Guide | Game Assessment Prep

What is the pymetrics Towers game?

Towers is a planning puzzle derived from the Tower of London but implemented on Hanoi-like hardware. Five distinct colored discs are distributed across three towers, each with five slots. A target pattern stays visible. Your task is to make the live arrangement match that target before the timer ends.

Only the top disc of a tower can move. You select a source tower, then a destination tower with a free slot. The discs are the same size, so there is no traditional Tower of Hanoi rule that a smaller disc must sit above a larger one. The constraint is access: a buried disc cannot move until everything above it has been relocated.

Our reconstruction labels every disc with a letter as well as a color, provides per-move Undo and full Reset controls, and uses a 120-second remaining timer. It records the pause before your first move and every successful move, undo, reset, and invalid attempt. The target is generated from the same five discs and verified by an exact solver before play begins.

What does Towers measure?

Planning efficiency is the most visible outcome. A solver can calculate the smallest possible number of legal moves between the start and target. Comparing your successful forward moves with that minimum shows how much detouring occurred. Our score expresses that comparison as an efficiency from 0 to 100, where a solved minimum-move path is 100 and an unsolved timeout is 0.

Time-to-first-move adds a second planning signal. A short pause can indicate that you inspected the target and formed a sequence before acting. Moving immediately may reflect effective rapid insight or trial-and-error; waiting too long can leave insufficient execution time. The number is descriptive, not a rule that longer is always better.

The task also involves working memory because you must preserve an intended sequence while the board changes. Undo use can reflect sensible error correction rather than failure. We show a plain-language observation from the full log without claiming to know how pymetrics combines these behaviors for any employer.

Parameters we know—and what remains uncertain

Five discs are high-confidence. A pymetrics patent explicitly describes five colored discs across three pegs, and that matches an open-source reconstruction plus preparation reports. Three equal towers with five slots and a permanently visible target are well-supported implementation details.

A two-minute limit is supported by several independent preparation sites and is the recommended default. One hobby clone used 60 seconds without a strong source, so we do not follow it. The exact production timer still needs a verified recording.

Click-to-select followed by click-destination is the best-supported real interaction. Dragging appears in a clone rather than the reported assessment. Multiple sources describe both an Undo control and a clear or restart control, but their exact behavior is not patent-confirmed. Our successful-move counter remains cumulative when you Undo or Reset so those controls cannot erase detours and inflate efficiency; production scoring behavior is not public.

How the minimum-move solver works

Every legal board is represented as three ordered stacks. From a given board, the solver creates every neighbor reachable by moving one top disc to another non-full tower. Breadth-first search explores those neighbors in layers: all one-move states, then all two-move states, and so on. The first time it reaches the target, the layer number is the exact minimum.

The five-disc state graph is small enough to solve locally and contains 2,520 connected legal arrangements. We verified that all of those states are reachable from a canonical board. Each generated start-target pair is accepted only when the solver returns a finite path, and the result is cached for the round. The minimum stays hidden during play because displaying it would be a coaching aid; it appears in the completion tile.

Six practical strategies

1. Compare towers from the bottom up

The bottom discs are hardest to change because everything above them must move first. Identify which bottom positions already match the target and avoid disturbing them unless the target requires it. Then find the deepest incorrect disc.

2. Work backward from the target

Ask what the final move must be. The target's top disc must arrive from another tower, which tells you where it needs to wait and what must be clear. Working backward one or two moves often exposes a cleaner forward plan.

3. Use one tower as temporary space

When a needed disc is buried, decide where the blocking discs will wait before moving the first one. Randomly spreading blockers can create a second obstruction. A temporary-space plan keeps the route reversible.

4. Preview a short sequence

You do not need to solve all moves mentally. Plan the first two or three, execute them, and reassess against the permanently visible target. This balances deliberation with the two-minute limit and reduces impulsive single moves.

5. Undo a branch early

If a move blocks the next required disc, use Undo immediately rather than adding several compensating moves. The action remains in the cumulative count, but one correction is still cheaper than building a long detour.

6. Reset only when the structure is lost

Reset is useful when several moves have made the route harder to reconstruct. It costs no extra time animation, but previous successful moves remain part of practice efficiency. For one localized mistake, Undo is usually better.

How to read your practice result

Moves versus minimum shows total successful forward moves and the exact BFS baseline. Time is elapsed time, while first-move think time isolates your initial planning pause. A perfect efficiency score means you solved in the minimum, not that the same path is the only optimal solution.

The session record stores the seed, initial board, target, final board, each before-and-after state, and Undo/Reset counts. That makes a round reproducible for debugging. A percentile compares planning efficiency in this simulation only after enough sessions exist; it is not a pymetrics percentile or hiring threshold.

Towers FAQ

Do disc sizes matter?

No. All five discs use the same movement rule. Any top disc can move to another tower with a free slot, regardless of color or letter.

Why did Undo not reduce Steps?

Steps records every successful forward move. Undo restores the board but does not erase the action from efficiency. This prevents unlimited corrections from looking like a minimum-move solution.

Is every random board solvable?

With these five distinct discs, three stacks, and reversible top-disc moves, the legal state graph is connected. We still verify every generated pair with BFS before showing it.

Should I wait before the first move?

Pause long enough to identify the first useful sequence, but leave time to execute and adapt. Planning latency is contextual; neither zero seconds nor a very long wait is automatically ideal.

Is there a pass mark?

No public universal pass mark or production minimum-move threshold exists. Employers use role-specific models. Practice should improve familiarity with legal moves and forward planning, not imitate a supposed preferred personality.