The lenses
Missing Force lens
The Missing Force lens identifies the highest-impact variables a model has not accounted for. It returns a ranked list with marginal contribution estimates, scoped to the buyer's substrate and verdict.
What it answers
The lens answers a single question, namely which variables outside the model are most likely material to the outcome. The answer is a ranked list. The top entry is the variable whose inclusion would have moved the structural reading the most. The remaining entries trail behind by marginal contribution.
The lens does not name an external dataset, and does not recommend a specific feature to acquire. It names the structural shape of the missing force; the buyer's team translates that shape into a feature.
When to run it
Run the Missing Force lens after any non-additive verdict where the buyer wants to widen the model. A multiplicative verdict points at a domain where the variables on file compose multiplicatively and the variables off file may compose differently. A mixed verdict points at a domain with two structural shapes coexisting, and the lens helps identify which is missing coverage.
The lens does not run on an additive verdict, where the variables on file already compose independently and the structural reading has no shape to extend. The lens does run on model-inadequate and data-insufficient verdicts, where the result page also renders the Improvement block; the two are complementary and do not duplicate.
Public-shaped output
The result page renders a Missing Force panel below the Model Comparison panel. The panel header reads Missing Force lens and carries the buyer's tier badge. Each row in the panel renders the variable name, the structural shape of the missing force, and the marginal contribution estimate.
Tier gate
Growth shows the top three entries. Business and Enterprise show the full list. Free and Pro do not unlock the lens. The cap is set by the engine flag missing_force_lens_full on the tier gate, and the page copy follows the engine.
| Tier | Unlocked | List depth |
|---|---|---|
| Free | ||
| Growth | ✓ | Top three |
| Pro | ||
| Business | ✓ | Full list |
| Enterprise | ✓ | Full list |
What buyers do with it
Buyers feed the top suggestions into the next modelling iteration. The lens does not name an external dataset; it names the structural shape of the missing force. The buyer's team translates the structural shape into a candidate feature, acquires the feature, and runs a fresh diagnosis.
Reviewers reading the report under SR 11-7 and PRA SS1/23 Principle 3.4 treat the Missing Force panel as a documented widening request, not a binding modelling instruction. The panel evidences the gap between the on-file variable set and the structural reading.