SPE Sensitivity and Configuration
Configuration controls how SPE interprets variable states and relationship patterns. Sensitivity analysis tests whether the comparison still holds when important assumptions move.
Configuration elements
The current Albarena SPE configuration includes:
- Initial condition sets: probability vectors over NN, N, O, P, PP
- Pattern sets: matrices for Sig+, Mod+, Sli+, No, Sli-, Mod-, Sig-
- Base modifier sets: coefficients used by simulation modes that adjust base-state influence
- Model configuration: the selected combination of initial condition set, pattern set, and base modifier set
Sensitivity analysis
Sensitivity analysis asks how much outcome distributions change when a Driver or Process is forced through NN, N, O, P, and PP. This helps the team identify the variables that most influence the outcome comparison.
Use sensitivity to answer:
- Which Driver has the largest effect on the selected Outcome?
- Which Process explains most of the variation?
- Does Scenario 1 still beat Scenario 2 when a key variable is unfavorable?
- Which relationships need better evidence before a recommendation is made?
Configuration discipline
Do not change pattern sets only to obtain a desired result. Pattern and base modifier adjustments should reflect evidence, expert review, or a deliberate calibration decision.
PLACEHOLDER: calibration guidance
Future documentation should describe how Albarena teams calibrate pattern matrices and base modifiers using historical data, expert workshops, and model back-testing.