Selection scorecards for pheno hunts
How you make selection processes more objective with criteria, weightings and confirmation runs.
Goal
Objective comparability
Mandatory
Confirmation run under the same conditions
Typical core criteria
10–15, weighted
Bias reduction
Blind scoring where possible
Key points
- Without a structured scorecard, selection quickly tips into gut feeling that only looks for a justification after the decision — fixed criteria before the run prevent this.
- Weighted criteria make different breeding goals (yield, aroma, stability, risk) visibly comparable, instead of unconsciously weighing them against each other.
- A candidate is only considered reliable after a confirmation run under the same conditions — single observations are too uncertain for a breeding decision.
Note
- Making a breeding decision based solely on a single, non-repeated run confuses random variance with genuine genetic quality.
Definition and classification
A selection scorecard is a standardized evaluation form that makes multiple phenotype candidates of a cross comparable according to the same, predetermined criteria.
Without a scorecard, implicit preferences (the 'prettiest' phenotype, the strongest memory) flow uncontrolled into the selection, which leads to unreliable decisions especially when there are several similarly good candidates.
Criteria axes of a robust scorecard
Agronomic axis: growth form, yield potential, stress resistance, disease susceptibility — traits that concern the grow itself.
Quality axis: aroma/terpene profile, trichome density, post-harvest quality, trimming effort — traits that concern the final product.
Consistency axis: batch-to-batch variation within the same phenotype across multiple runs — a candidate with high individual quality but large variation is riskier for production than a somewhat weaker but more consistent one.
Set weighting before the run
Weightings should be fixed BEFORE the evaluation begins, not adjusted afterward to fit the preferred candidate — otherwise the scorecard loses its actual function as a bias corrective.
Ten to 15 core criteria are usually the upper limit in practice at which consistent, fatigue-free evaluation is still possible; more criteria tend to dilute discriminatory power rather than increase it.
Checklist
- Fix the weighting of all criteria in writing before the first evaluation round
- Limit the number of criteria to 10–15 core points
- Use an identical evaluation form for all candidates
Bias reduction in evaluation
Where possible, evaluate candidates blind (coding instead of name/origin labeling) to reduce expectation effects from known line reputation.
Explicitly combine subjective sensory data (aroma, appearance) with objective lab or process data (cannabinoid/terpene values, yield weight), rather than relying on only one of the two data types.
The confirmation run
Only candidates that again perform comparably well in a second, independent run under the same conditions should be carried forward into further breeding or production.
A single outstanding run can be distorted by random effects (slightly different microclimate, position in the grow room) — the confirmation run is the actual filter, not the initial observation.
Common mistakes
Adjust weighting only after observing the favorites, rather than fixing it in advance.
Skipping a confirmation run because the initial candidate was 'clearly' convincing.
Selecting exclusively by appearance or aroma, without factoring in agronomic stability and consistency.
Frequently asked questions
Can I select without lab values?
How many criteria make sense?
Is one outstanding run enough as a basis for the decision?
Sources
→ Full registerRelated articles
Genetics & Selection
Genetics and phenotype selection
How you compare genetic lines, select stable candidates, and avoid drift across generations.
Genetics & Selection
Mother plants and clone hygiene
How to maintain healthy mother lines, avoid cross-contamination, and make clone programs reproducible.
Genetics & Selection
Assessing genetic stability across generations
How true stability of a cannabis line differs from random similarity, and what practical methods test it across multiple cycles.