Guppy Vitality Lab

Method

What the pond is really doing

Guppy Vitality Lab dramatizes a few intuitions from complex-systems science with cartoon fish. The simulation is deliberately simple; the numbers are playful proxies, not measurements.

⚡ The one real measurement: Live CE

It asks one honest question: does the group of fish predict its own next move better than the individual fish do? When the answer is yes, the pond as a whole carries information its parts don't — that surplus is what scientists call causal emergence, measured in bits. The meter reads it live from the last few seconds of pond life.

It runs small here — about 0.1–0.2 bits. That's real and roughly 10× what random noise would give, but modest: the group view is only ~15% more predictive than the fish-by-fish view. Read it as 'above chance,' not 'strongly emergent.'

You can watch it as the Live CE meter on the pond, or see the full computation — transition table, bits, and an honest above-chance check — on Measured Mode. Everything else on this page is playful proxy, not measurement.

More parts is not better organization

A pond crammed with one giant school is not 'more organized' — it is overgrown. Healthy systems balance how much structure they grow against how much they can support.

Healthy self-organization needs pruning and resilience

Living systems prune stale structure and keep reserves to bounce back from shocks. Turn pruning all the way down and watch a single school crowd out the pond; turn danger up without healers and watch resilience fall.

Scale matters, but it is a lens — not a score

The same pond looks different as individuals, schools, neighborhoods, or one whole pond. Switching the scale lens changes what you notice, not what is 'true'. Real multiscale causal-emergence work asks which scale best predicts the system; here it is just a way of looking.

These 'vitality' metrics are playful proxies

Vitality, Causal Spark, Scale Spread, Flow Cleanliness, Resilience, and Overgrowth Risk are cute numbers computed from on-screen behavior. They borrow names from systems science (effective information, causal emergence, dynamical reversibility) but compute none of that real math. The one exception is the Live CE meter on the pond — and Measured Mode — which do compute the real causal-emergence math on the toy's own data, and report it honestly (it runs small: ~0.1–0.2 bits).

The toy metrics

Each is a bounded 0–100 number computed from what the fish are doing on screen. The names wink at systems science; the math is not real.

  • VitalityOverall pond liveliness: balanced energy, structure, and few pathologies.
  • Causal SparkPlayful nod to 'the whole coordinating the parts' — how coherently schools move together. Not real effective information.
  • Scale SpreadHow many organization scales coexist (lone fish, small schools, big schools). A toy multiscale spread, not a real measure.
  • Flow CleanlinessHow orderly and low-noise the motion looks. A playful gesture at 'clean dynamics', not real reversibility math.
  • ResilienceCapacity to bounce back after shocks: healers present, energy in reserve.
  • Overgrowth RiskHow close the pond is to one school swallowing everything. Higher is worse.

The scale lens

Switching the lens changes what you notice, not what is true. Real multiscale causal-emergence research asks which scale best predicts a system; here the lens is purely a way of looking.

  • IndividualSingle guppies: mood and role per fish.
  • SchoolSchools as glowing groups with size labels.
  • NeighborhoodNearby schools grouped into a coarse neighborhood.
  • PondThe whole pond as one system.
  • PathologyHighlight only the troubled, pathological regions.

Pathologies you can grow

These are recognizable failure shapes of self-organization, dramatized with cartoon fish — not diagnoses of anything or anyone.

  • OvergrowthOne school grows too large and dark, crowding everything out.
  • WitheringGuppies drift apart and fade as structure dies back.
  • InfightingDefender guppies turn on the helpers they should protect.
  • FrenzyToo many flashing, transient schools form and dissolve at once.
  • StagnationOld grey schools persist long past their usefulness.

Plain-language glossary

The few terms worth knowing. None of them require math — they're ways of asking “does the whole add something the parts don't?”

  • Effective information (EI)How much watching one moment tells you about the next — a system's predictability, in bits. High EI = the present strongly pins down the future.
  • Causal emergence (CE)When zooming out to groups predicts the future better than tracking every individual. If the coarse 'group' view beats the fine 'fish-by-fish' view, the whole has emerged above its parts.
  • Coarse-grainingChoosing how to group the parts — e.g. lumping similar fish moods into a few 'group moods.' A good grouping reveals structure; a bad one just blurs it.
  • BitsThe unit of information. One bit is one yes/no question's worth. Causal emergence here is fractions of a bit — small but real.
  • Live CE meterThe glowing meter on the pond. It runs the real causal-emergence math on a rolling window of the last few seconds, so you can watch the number move as you tune the sliders.

Honest about the inspiration

Inspired in spirit by research on emergence and causality in complex systems — including work on effective information and causal emergence, and newer takes framing causal emergence through dynamical reversibility. Guppy Vitality Lab implements none of that mathematics; it is an art toy that gestures at the ideas.

What this is not

  • Not a measurement of any real system — the causal-emergence math here runs only on the synthetic pond, and the on-screen 'vitality' scores are playful proxies, not even that.
  • Not a clinical, diagnostic, treatment, safety, or emergency tool.
  • Not a model of any real person, patient, operator, or live AI system.
  • Not connected to any other project, agent, runtime, or dataset.

🐟 Back to the pond