Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →yogsoth-ai/discriminating-prediction-design
SOP: design key predictions and observation plans that can distinguish competing hypotheses
Passed 8 of 23 test cases with the skill, and 15 of 23 without it.
and 2 more in the full report
3 fixed10 passed either way5 still failing5 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 3 test cases the skill fixed. Same model, same prompt, with and without the skill.
We are evaluating a single hypothesis regarding dark energy (H1: cosmological constant from vacuum energy, mechanism: zero-point quantum fluctuations induce negative pressure, unique prediction: equation of state parameter w remains exactly -1 over cosmic time). Output a JSON array with comparison, divergence_point, h1_prediction, ch1_prediction, discriminating_observation, method, feasibility, and feasibility_notes.
| Checked | Without skill | With skill |
|---|---|---|
| The output reports an error indicating at least 2 competing hypotheses are required. | Missed | Passed |
| The output halts processing without generating fake hypothesis comparisons. | Missed | Passed |
The instructions your model reads when the skill loads.
Design the key predictions that can discriminate between competing hypotheses, providing direction for subsequent experiment design.
Not satisfied → stop and return an error: at least 2 competing hypotheses are required to design discriminating predictions.
[
{
"comparison": "H1 vs CH1",
"divergence_point": "The condition or measurement where predictions differ",
"h1_prediction": "What H1 predicts in this condition",
"ch1_prediction": "What CH1 predicts in this condition",
"discriminating_observation": "The observation or experiment that would distinguish them",
"method": "Suggested research method (experiment/survey/natural experiment/meta-analysis/etc.)",
"feasibility": "high | medium | low",
"feasibility_notes": "Why feasible or what barriers exist"
}
]
Covers all major hypothesis pairs (primary vs. each competing).
test cases passed, vs 15 / 23 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.