Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →yogsoth-ai/comparative-formulation
Strategy: Construct comparative research questions — systematic comparison of A vs B
Passed 19 of 22 test cases with the skill, and 13 of 22 without it.
and 11 more in the full report
8 fixed11 passed either way1 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
I'm designing a study comparing our new Transformer model against traditional XGBoost for predicting 30-day hospital readmissions. Can you help me frame structured comparative research questions? I need to know what evaluation metrics to compare, what variables to control across both models, and how to define what constitutes a meaningful performance difference.
| Checked | Without skill | With skill |
|---|---|---|
| Structure the comparative research question using the PICO framework. | Missed | Passed |
| Include a FINER criteria assessment for the comparative research question. | Missed | Passed |
The instructions your model reads when the skill loads.
Construct comparative research questions — when research requires comparing A vs B, systematically construct a fair, meaningful comparison.
Core logic: a good comparative research question requires clarifying four elements — what is compared (objects), along what dimension (metrics), under what conditions (controls), and what counts as "different" (threshold).
| Type | Example | Key considerations |
|---|---|---|
| Method comparison | Method A vs Method B | Implementation fairness, dataset selection |
| Condition comparison | With X vs Without X | Controlled variables, confounding factors |
| Group comparison | Group A vs Group B | Matching, selection bias |
| Temporal comparison | Before vs After | History effects, maturation effects |
| Tier | Comparison design | Fairness argument | Output |
|---|---|---|---|
| S | Comparison objects + clear dimensions | Basic fairness statement | ≥1 comparative RQ |
| M | + controlled variables + effect size | Fairness argument + identification of potential bias | ≥2 comparative RQs |
| L | + multi-dimensional + sensitivity | Full fairness analysis + bias mitigation strategy | ≥3 comparative RQs |
test cases passed, vs 13 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.