Methodology

Last reviewed 2026-07-26 · maintained by the Cultivar grow team · about

Short version: published horticultural ranges, narrowed by stage, cross-checked against what real logged grows actually do — and stated as ranges, because cultivars vary. Everything below is the same logic the app runs on, so a number you read here is a number your grow is graded against.

1 · Environment bands

VPD is computed from temperature and relative humidity: VPD = (1 − RH/100) × SVP, where SVP = 0.6108 × e^(17.27T/(T+237.3)) and T is in Celsius. Leaf temperature typically runs 1–3°C below air temperature under LEDs; where a leaf offset matters we say so, and the calculator lets you set it. We publish air-temperature VPD by default because that is what an inexpensive sensor measures — quoting leaf VPD from an air sensor would be false precision.

2 · Why the bands are stage-specific

A plant's ability to move water changes as it builds root mass: low VPD early, higher through veg and flower, driest in late flower to protect dense buds from botrytis. The exact table is on the VPD chart. We state ranges, never single values — cultivar, phenotype, airflow and root health all shift the optimum, and anyone publishing one exact number for all cannabis is guessing.

3 · How a grade is calculated

An environment grade is the share of a plant's readings that fell inside its stage band at the time the reading was taken — the stage is resolved from the grower's own stage dates, not from today's stage. It is not an average of averages, and it is never a comparison against other growers. A grow with fewer than three qualifying readings gets no grade rather than a misleading one.

4 · Where readings come from, and why we label them

Every environment reading carries its source: typed by hand, read over Bluetooth from a sensor, pulled from a connected device account, or fetched from local weather for outdoor grows. Automated readings are continuous; a manual one is a single moment somebody remembered to record. When we publish an aggregate we say which kind it came from and how many grows are behind it. A figure drawn from typed-once entries is a weaker claim than one drawn from sensor streams, and we will not blur the two.

5 · Genetics data

Strain records — flowering time, type, lineage, breeder — are compiled from breeder listings and public genetics databases. They are reported values, not measurements. Breeders have every incentive to publish flattering flowering times and yields; treat them as the starting hypothesis your own journal tests. Pedigree trees are built from the lineage graph, so an error in one parent propagates — every strain page carries a "Suggest an edit" button, and corrections are applied to the source data rather than patched page by page.

6 · Photographs

Strain photographs are cached from public sources and are illustrative of the variety, not of a specific phenotype or a specific grow. Where a photo comes from a grower's published journal it appears as part of that journal, with its date and context. We do not use marketing renders to represent grow results.

7 · How the assistant is grounded

The on-page grow assistant and the in-app coach are given the same reference bands published here, plus — inside the app — that grower's own logs, readings and genetics. They are instructed not to contradict these tables, not to invent conditions they cannot see, and to say plainly when an answer depends on data only the grower has. They are language models: they can still be wrong, they are a starting point rather than an authority, and they will never be the reason we tell you something we would not print on this page.

8 · Aggregate data from real grows

Where a page cites what real grows do, the figure comes from published journals in aggregate — never private accounts, and never in a form that identifies a grower. Sample size is always stated. We report medians and ranges rather than means where the distribution is skewed, and yield always is. Self-selection is real: growers who publish are more engaged than growers who do not, so read our aggregates as "what careful home growers achieve", not "what the average grow yields".

9 · What we don't do

• No yield or potency promises. Breeder-reported figures are labelled as such.
• No medical, dosing or consumption advice. Cultivation only.
• No legal advice. Growing law varies by jurisdiction and changes; check your own.
• No dressing guesses as data. Where the evidence is thin — flushing is the classic example — we say so.
• No paid placements inside guidance. Product links are affiliate links, disclosed at the link, and chosen to solve the problem on the page they sit on. Nobody pays to be recommended, and the coach's advice is never sponsored.

10 · Corrections and versioning

Guides carry a published date and an updated date, and both appear in the page's structured data. A substantive correction changes the updated date; a typo fix does not. Strain corrections go through the "Suggest an edit" button and are applied at the data layer. If you find something wrong here, telling us is the fastest way this library gets better.

Track your own grow free — the journal grades your environment against these bands →