Skip to content
Get in touch

Hex Life

Watch people move through a shared space. Change the conditions, follow the encounters, and compare what happens.

Instructions, details & credits
An event-space simulation

Created by Mike Besso

  • Simulations
  • At your own pace
  • Repeatable experiments

Start the model, pause to look closer, and test one change at a time.

Executive Summary

Watch a room fill with people, then look for the patterns that emerge. Hex Life is a simulation of movement and encounters on a hexagonal floor. People arrive, visit activity areas, form conversations, and leave. You can pause, inspect individuals, change conditions, and compare repeatable experiments.

This is a systems model for exploration. Its rules make behavior inspectable; they do not predict how a real venue or a particular person will behave. The colored people are model markers, not demographic groups.

Before You Play

The room starts running when you open it. Let a few people enter, then choose Pause and select a person to see their destination, needs, and recent encounters. Step advances one simulated second. Start with the default population before trying a different target or conversation chance.

For a comparison, change one condition at a time. Under Experiment record, describe what you expect, and choose Start experiment. Add what you observed and learned, then Export results to keep the configuration, seed, timeline, and measurements. Reloading starts a new session; download results you want to keep.

Controls and Accessibility

  • Use Play, Pause, Step, and the speed buttons to choose the pace. A busy device may run below the requested speed.
  • Select a person on the map to open their inspector. Use touch, a pointer, or Tab and Enter to select a person. Where people overlap, a chooser lists the group.
  • Observation mode simplifies the display. Count overlays adds cell counts. The dashboard and inspector expose written state alongside the visual map.
  • Inputs and buttons have labels and visible focus. The spatial map remains a visual task; the written diagnostics do not provide a complete screen-reader equivalent.
  • Optional Record screen uses your browser’s screen-sharing permission and saves a recording locally. You choose what to share. Experiment exports download locally; Hex Life has no account or server for experiment data.

How We Made This with Generative AI

Mike Besso directed the model and experiment workflow. Generative AI helped implement the movement engine, interface, diagnostics, and automated checks. The development process separates a frozen model from experiments so changes can be documented and compared.

The Game Room migration preserves the corrected Hexlife Baseline v1 model. The new title, learning page, and static packaging make it available as a Simulation while retaining the original experiment identifiers.

The Gory Details

The floor has 88 hexes. Traversable cells contain six standing positions, and movement reserves an exact destination. Stage and food footprints are obstacles with activity around their edges. Entrances are transition zones: admission requires an available position and an immediate inward move, and routes cannot use entrances as intermediate shortcuts.

One engine tick is one simulated second. Seeds, configuration, and command order support deterministic replay. Exported experiments retain the identifier hexlife-baseline-v1; the entrance correction is BLV1-C001. Same-color duration bias is a configurable rule to investigate, not an empirical claim about people.

The model, playback logic, baseline documents, and original tests are preserved from the source checkpoint. Automated checks exercise movement, conversations, reservations, entrances, experiments, and repeatability. Browser checks cover the Simulations listing, learning page, controls, export, and narrow-screen layout.

Credits

Creator

Created by Mike Besso.

Human direction and Generative AI

Mike Besso directed the model and experiment workflow. Generative AI helped implement the movement engine, interface, diagnostics, and automated checks. The development process separates a frozen model from experiments so changes can be documented and compared.

Read the game brief for the development account, controls, and limitations.