How we rate cannabis strains

What we rate

Cannabis flower in German pharmacies is usually described by two numbers: cannabinoid content (THC, CBD) and botanical class (indica, sativa, hybrid). THC/CBD content and the indica/sativa label only explain the expected profile so far. The chemical makeup of cannabinoids, terpenes, and their ratio says more. Dose, route of administration, tolerance, and personal context still matter too.

A 25% THC sativa and a 25% THC indica can have different effect profiles, because strains with similar THC content show very different chemical profiles. Terpene profiles are a key axis of difference, but not the only predictor of effect.

We rate strains by their terpene profile against six effects: relaxation, sleep, mood, focus, pain, and inflammation. Each strain gets a per-effect value on a 1-to-10 scale, calibrated per effect. The effect axes describe how close a profile sits to pharmacological dimensions of action, not indications or treatment recommendations.

The rating runs in three steps: an editorially maintained effect inventory per terpene (what individual terpenes do), soft assignment plus antagonist damping per strain (how the profile spreads out), and a display scale calibrated per effect (what the number looks like in the listing).

Source hierarchy: three tiers

What makes it into our inventory and what does not is decided by a three-step source rule. Tier A is mandatory; Tier B and C fill gaps, always in a clear order and without overriding Tier A.

Tier A: pharmacological foundation

Peer-reviewed pharmacology. Conditions: original work or systematic review, mechanistically sound (receptor binding, in-vivo effect), and ideally independently replicated. Databases: PubMed, PubMed Central, Europe PMC. Substance identities from PubChem and ChEBI. This is the foundation of the effect inventory.

1. Russo EB. Taming THC: potential cannabis synergy and phytocannabinoid-terpenoid entourage effects. British Journal of Pharmacology 2011; 163(7): 1344–1364. PMID 21749363 · DOI: 10.1111/j.1476-5381.2011.01238.x
2. Gertsch J, Leonti M, Raduner S, et al. Beta-caryophyllene is a dietary cannabinoid. Proceedings of the National Academy of Sciences USA 2008; 105(26): 9099–9104. PMID 18574142 · DOI: 10.1073/pnas.0803601105
3. Santiago M, Sachdev S, Arnold JC, et al. Absence of Entourage: Terpenoids Commonly Found in Cannabis sativa Do Not Modulate the Functional Activity of Δ9-THC at Human CB1 and CB2 Receptors. Cannabis and Cannabinoid Research 2019; 4(3): 165–176. PMID 31559333 · DOI: 10.1089/can.2019.0016
4. LaVigne JE, Hecksel R, Keresztes A, Streicher JM. Cannabis sativa terpenes are cannabimimetic and selectively enhance cannabinoid activity. Scientific Reports 2021; 11(1): 8232. PMID 33859287 · DOI: 10.1038/s41598-021-87740-8
5. Bidwell LC, YorkWilliams SL, Mueller RL, Bryan AD, Hutchison KE. Association of Naturalistic Administration of Cannabis Flower and Concentrates With Intoxication and Impairment. JAMA Psychiatry 2020; 77(8): 787–796. PMID 32520316 · DOI: 10.1001/jamapsychiatry.2020.0927

Tier B: cannabis context and market data

Tier B has two roles: pharmacological context evidence (cannabis-specific reviews, mechanistic findings below clinical RCT level) and market/product data (real strains, batches, terpene profiles from pharmacy stock). Pharmacological plausibility and product data are judged separately and only brought together in the score. Tier B does not override stronger Tier A evidence.

1. Smith CJ, Vergara D, Keegan B, et al. The phytochemical diversity of commercial Cannabis in the United States. PLoS ONE 2022; 17(5): e0267498. PMID 35588111 · DOI: 10.1371/journal.pone.0267498
2. Watts S, McElroy M, Migicovsky Z, et al. Cannabis labelling is associated with genetic variation in terpene synthase genes. Nature Plants 2021; 7(10): 1330–1334. PMID 34650265 · DOI: 10.1038/s41477-021-01003-y
3. Hazekamp A, Fischedick JT. Cannabis: from cultivar to chemovar. Drug Testing and Analysis 2012; 4(7–8): 660–667. PMID 22362625 · DOI: 10.1002/dta.407
4. Herwig N et al. Classification of Cannabis Strains Based on their Chemical Fingerprint: A Broad Analysis of Chemovars in the German Market. Cannabis and Cannabinoid Research 2025; 10(3): 409–419. PMID 39137353 · DOI: 10.1089/can.2024.0127

Tier C: cross-context evidence

Aromatherapy and essential-oil studies where the terpene is the main component of a blend. Used only as a directional signal, with a clear main-component label and a transfer discount (different matrix, different dose, different route, different population). Tier C never raises a score category on its own; it only supports a direction that is already plausible. Lavender studies (linalool as the main component) serve as a cross-reference for the anxiolytic linalool rating; chamomile studies (bisabolol as the main component) for bisabolol against relaxation/sleep.

1. Moss M, Oliver L. Plasma 1,8-cineole correlates with cognitive performance following exposure to rosemary essential oil aroma. Therapeutic Advances in Psychopharmacology 2012; 2(3): 103–113. PMID 23983963 · DOI: 10.1177/2045125312436573

Score logic across the tiers

Tier A comes first. Tier B and Tier C only count where Tier A has no value, and both carry reduced weight, with Tier B above Tier C. Tier A is weighted by how close the model is to humans: human studies above animal models, animal models above pure in-vitro data. Where direct human or mechanistic evidence contradicts a direction of effect, we do not derive that direction from indirect evidence.

How we rate: two axes

For each terpene and effect we keep two separate values. This is new in v2.1 and replaces the single-axis rating of the previous version, which mixed effect strength and evidence into one number.

• Effect strength (0 to 3): how pronounced the effect is when it shows up. 3 is a canonical main effect (linalool and anxiolysis, myrcene and sedation), 2 a clear secondary effect with a mechanism, 1 a weak side signal, 0 no solid finding.
• Evidence (0 to 3): how well the effect is documented. 3 is a human study or repeatedly replicated animal models, 2 a solid single model with a clear mechanism, 1 a single plausible finding, 0 no finding.

The cell score is effect strength × evidence factor. The evidence factor steps down gently: 1.0 at evidence 3, 0.9 at 2, 0.8 at 1, 0.7 at 0. That keeps the scale between 0 and 3.0.

The point of splitting them: a well-documented strong effect ranks high, a weak but well-documented signal ranks lower. Effect beats paperwork. A strong effect with just one good animal study (3 times 0.9 equals 2.7) beats a broadly documented mini-effect (1 times 1.0 equals 1.0).

We do not award effect strength from indirect evidence. Sedating terpenes get no focus points just because a study calls them "anxiolytic." The anxiolysis of a sedating terpene goes into relaxation, not mood or focus. Wakeful terpenes get no sleep points for stress-reduction findings. That keeps the inventory clean.

From terpene to strain score

A strain has a terpene profile made of several terpenes. We combine the cell scores into one value per effect in two steps:

Soft assignment. Each terpene spreads its score budget across the six effects. A specialist (nerolidol concentrates on relaxation) keeps its strength in its main effect. A generalist (caryophyllene works across several axes) gets diluted across them. This stops a chemically broad strain from sitting at the top of almost every filter.

Antagonist damping. Certain effect pairs are pharmacologically incompatible. When a strain sedates heavily, we cap its mood score at forty percent of its unconstrained value. So sedating strains do not rank high in the mood filter, even if they nominally have high mood values from individual terpenes. The antagonist pairs run between focus and sleep, and between mood and sleep. Pain and inflammation have no antagonists, because physical effects are not antagonistic to mental ones.

Display scale 1 to 10, calibrated per effect

The internal scores range differently from effect to effect. Focus sits lower in absolute terms than inflammation, because fewer terpenes have strong focus profiles and generalist terpenes get further diluted by soft assignment. If we showed the raw internal scores, the top focus strain would look weaker than the top inflammation strain, even though it leads its own axis.

So we calibrate the display per effect against the real distribution in the database: the twentieth percentile becomes a 2, the median a 5, the top percent a 9. The top strain per effect then lands consistently between 9.3 and 10.0, and the median per effect around 5. In the listing, the focus leader sits level with the inflammation leader, even though the absolute values differ.

What we deliberately do not use as a source

Pharmacy and manufacturer data are used only for product identity, availability, price, and reported lab values (e.g. terpene profiles from a CoA). Effect claims from shop copy, marketing tags, or strain descriptions do not feed the effect score.

• Pharmacy and manufacturer marketing as proof of effect.
• Reddit and forum reports (useful for hypotheses, not for claims).
• Strain-recommendation platforms whose effect labels come from marketing tags rather than a source stack.
• Aromatherapy studies without a clear main component (e.g. blends with no isolated active ingredient).

Six effects

From the inventory we derive, for each of the 21 terpenes, which direction it pushes which effect axis and how solid that statement is. That builds a multi-axis effect profile per strain. The six effects shown publicly are relaxation, sleep, mood, focus, pain, and inflammation.

The focus axis is deliberately broad: mental clarity, alertness, attention, memory, learning and work performance, non-sedating activation. So terpenes with a cognitively activating effect count toward the focus score, not just one single marker. Sedating profiles get pulled down in the focus filter through antagonist damping.

The effects energizing and appetite from the first version are currently off. In a cannabis context, energy is more about daytime suitability than a standalone direction of effect, and it sits under the focus score. Appetite is mainly a CB1 effect of the cannabinoids and only secondarily a terpene effect.

How THC enters the score

THC does not enter the score as a positive effect factor. Higher THC content is not a reliable predictor of a sleep, focus, mood, or pain profile. The pop-science assumption that "more THC = sleepier" or "more THC = more focused" does not hold up in the Tier A studies: Bidwell & Hutchison 2020 showed that concentrates with a much higher THC blood level produced subjective effects similar to flower. THC does stay a central dose, intensity, and risk factor (intoxication, anxiety, sedation, cognitive impairment), so we report it separately as a filter. More detail under Does THC content tell you everything about the effect?

Data sources for availability and price

We aggregate eight German pharmacy platforms plus their connected pharmacies at once. Most comparison sites show only their own pharmacy network. We give the lowest price actually offered per strain, with shipping and prescription fee as visible tags.

Outdated offers are hidden. You can see how current the data is per strain right on the product. Prices and availability can change at short notice.

Why we update daily, not hourly

Other comparison sites advertise hourly or real-time updates. We deliberately do not. Eight platforms, each with its own data structure, naming conventions, and encoding quirks, with up to a thousand pharmacies behind them, cannot be merged reliably every hour. The result would be fast and wrong: strains listed twice, pharmacies mismatched, prices from stale caches, profiles from last year.

Instead, one full pass runs each day with cleanup: cross-source matching, strain normalization, pharmacy-identity consolidation, shipping and prescription-fee reconciliation, and batch dedupe. The result is reliable to the day rather than shaky by the hour. If you see a price from six hours ago, you can trust which way it points.

Batch variability

Terpene profiles are rated as close to the batch as possible, where CoA or pharmacy data exist. If only generic manufacturer or platform profiles are available, the score is based on a strain profile, not a batch profile. Terpenes are volatile: drying, storage, irradiation, and age can change the profile. For the exact values of a specific batch, the pharmacy label or the certificate of analysis (CoA) always applies.

Ratings: where the stars come from

On every strain page we show an aggregate of ratings from several pharmacy platforms. The value is the weighted average across all sources that carry the strain, weighted by the number of ratings on each platform.

The current rating sources are medcanonestop, flowzz, canngo, and Privatrezept. All four show a star rating and a rating count on their product page, anonymized without real names. We take only the aggregate numbers, no review text, and name the source on each strain page. A strain appears with an aggregate only when at least five ratings exist.

On a strain hub (the genetics page) the aggregate is averaged across all strains of that genetics. There we show the number of aggregated strains rather than the sum of individual ratings, because that sum is methodologically shaky (double-counting across platforms, the same person rating several strains).

Once we have our own rating system with registered users, we will add our own ratings to the platform aggregates and phase out the external sources when our sample grows larger than the external one.