Use this method when you need to explore every possible combination of your input settings. By testing all variations of your factors together, you ensure that you do not miss how different settings influence each other. This provides a complete, clear picture of your entire experimental space in a controlled way.
Complete Experimental Coverage
A full factorial design creates a matrix that represents every possible setting combination. Because this method does not skip any combinations, it is the most thorough way to understand how your inputs affect the outcome.
By systematically varying all factors simultaneously, you capture the main effects of each input and the specific interactions between them. This is useful when you need to be certain that you have not overlooked a hidden relationship between your process settings.
Handling Multiple Levels
Lattice allows you to define different numbers of levels for each factor. Whether you are working with binary settings (high and low) or multiple variations (low, medium, and high), the tool calculates the full set of unique permutations required for your experiment.
This flexibility ensures that your experimental matrix is customized to the specific requirements of your process, allowing you to test complex setups without manually calculating the combinations yourself.
Standardization and Reproducibility
Every design generated maintains a 'standard order,' which provides a consistent baseline for your documentation. This helps in tracking your progress and ensures that your experiment can be easily audited or revisited later.
While the standard order is preserved for record-keeping, you can apply randomization to your experiment execution. This helps protect your results from time-based changes or environmental shifts, such as equipment drift or fluctuating conditions during long test sequences.
When to Choose Alternatives
While this approach is thorough, it requires more runs as you add more factors or levels. If your experiment grows to a size that becomes impractical or too time-consuming to execute, other methods can help reduce the number of runs while still providing valuable data.
For complex processes with many factors, you may consider exploring design types that focus on main effects or specific interaction patterns to maintain efficiency without sacrificing the quality of your findings.
1 · Intent → method
An LLM picks doe_generate_full_factorial from a fixed catalog.
2 · Method → numbers
Deterministic Python engine runs the math. Same input → same output.
3 · Numbers → plain language
A second LLM translates the result into your domain’s vocabulary.
Why is this considered a 'full' design?
It is called full factorial because it tests every possible combination of every level for each factor you provide, ensuring no interaction is left unexplored.
How many experiments will I need to perform?
The number of runs is determined by multiplying the number of levels for each factor together. For example, 3 factors with 2 levels each results in 8 unique runs.
Tool input schema
Schema for doe_generate_full_factorial not exported yet (run pnpm export:registry).