Unknowns
What the model can't explain yet.
Every miss points to something the model has wrong or doesn't know: a link nobody has measured, a number it can only guess, studies that disagree. Each unknown comes with the evidence or experiment that would settle it.
The loop on one question
A worked example: a prediction that failed against a trial, the unknowns it exposed, and the experiments that would close them. You can run the experiments yourself.
Worked example · metabolic, liver, type 2 diabetes
Explore the loop →Why does fat build up in the liver in type 2 diabetes?
2 failed predictions5 missing links5 numbers poorly pinned down3 studies that conflict
From forecasts
As sealed forecasts resolve, every confident miss will add unknowns here, grouped by disease and by what went wrong. See the trials we will forecast →