Candidate ranking identifies interesting worlds. Expected information gain asks a different question: which supported measurement would reduce a declared uncertainty the most?
25 candidates × 6 actions · 83 supported calculations
25
candidate worlds
leading ranked sample
6
measurement actions
explicit synthetic likelihoods
83
supported actions
67 remain undetermined
bits
information unit
expected KL divergence
What the analysis found
A precise parameter offers little information if it is already known well. A weaker parameter can dominate the decision if a plausible measurement sharply contracts its posterior. Unsupported actions remain withheld because information gain requires both a calibrated prior and a defensible observation model.
Decision laboratory
The calculator compares a declared current uncertainty with a synthetic future precision. It reports expected posterior shrinkage and expected KL information in bits.
Expected update
SIMULATEDplanet mass or minimum mass
Prior σ
0.055 Earth masses
Synthetic σ
0.106 Earth masses
Expected posterior σ
0.0488 Earth masses
0.1734 bits
Expected KL information · target-local action rank 2
Observing time and cost: not modelled. This is a precision requirement, not an observing proposal or instrument guarantee.
