What the score is, and is not
The planning index is a conservative decision-support heuristic. It is not a measurement of ammonia production and it is not a promise that a particular number of fish will be safe. Real aquariums vary with feeding, metabolism, filter media, maintenance, oxygen, plant growth and individual animals.
1. Adult biological pressure
Each species contributes load units based on adult body length raised to a super-linear exponent, then adjusted by a relative waste factor and activity. This deliberately avoids the old “one inch per gallon” rule: body mass and waste do not scale linearly with length.
2. Usable water, footprint and surface area
The model estimates usable water below gross glass volume, then separately considers footprint and water-air surface area. That means two aquariums with the same water volume can receive different spatial and gas-exchange signals.
3. Filtration and maturity
Rated filter flow is converted to turnover against estimated usable water, with a modest filter-type adjustment. A newly cycled system receives less modeled resilience than a mature tank. This is intentionally conservative because a filter's advertised flow-rate says nothing about media condition or actual biological capacity.
4. Compatibility is multi-axis
Pairwise checks cover temperature and pH overlap, size/predation, fin-nipping, shared territorial zones and selected species-specific rules. “Good” means no major conflict was found by the model; it never means guaranteed compatibility.
5. Social groups and swim zones
Schooling and social fish are checked against a typical minimum group starting point. Occupancy is also accumulated by top, mid and bottom zones so spatial crowding can be flagged independently of biological pressure.
Data quality
StockingLab uses two visible data tiers. The curated core contains the original manually assembled StockingLab profiles. Extended baseline profiles use conservative detailed planning defaults to broaden coverage, but are explicitly marked for species-specific verification and reduce the plan confidence score. Baseline records are not allowed to drive automatic “what next?” recommendations. This avoids pretending database breadth and data certainty are the same thing.