Early access — cascade metrics are real (derived from canonical token telemetry); the operator field is a curated seed. Learn more about the data

Leverage

6Production evidence

Cache reuse per fresh input. How much an operator leverages cached signal relative to new input.

Last updated: 2026-09-01Spec: SigRank Standard v1.0

Definition

Cache reuse per fresh input. How much an operator leverages cached signal relative to new input.

Inputs

cache_read, input.

Derived variables

leverage = cache_read / max(input, 1)

Claim

Leverage measures how efficiently an operator reuses cached context relative to fresh input.

Test

Verify leverage = cache_read / max(input, 1) for known inputs. Cross-check with 10xDEV = log10(leverage).

Observable

The numeric leverage value.

Falsifier

If leverage does not equal cache_read / max(input, 1), the implementation is wrong.

Evidence

Canonical test. Field data shows leverage distributions across operators.

Limitations

Does NOT establish causal leverage or business value. High leverage may indicate efficient reuse OR may indicate insufficient fresh input.

Lineage

SigRank Standard, cascade ontology (Reuse stage = cache_read / cache_create). Architecture: mos2es.com/architecture.

Cross-references