Mandates
Calculations

Versioned, explainable, replayable risk numbers.

We design calculation and reporting systems where logic is versioned, inputs are traceable, outputs are replayable, and model definitions are explicit. When a number changes, it is clear whether the inputs changed, the logic changed, or both.

Signals
Model Instability
Standards
Logic Releases Blocks

Scope

We cover the quantitative and analytical infrastructure that produces the institution's numbers: pricing models, risk measures including VaR, Greeks, and stress scenarios, P&L attribution and analysis, and regulatory and compliance calculations. Anywhere the institution relies on a computed result to make a decision or meet a reporting obligation is in scope.

  • Mark-to-market, VaR, sensitivities, and scenario analysis across asset classes.
  • P&L attribution and analysis.
  • Capital adequacy, large exposure, and transaction reporting computations.

Approach

We treat calculation logic with the same rigor as production code. Every model has a formal definition with explicit input and output schemas. Version identifiers are tied to every result set so that any figure can be traced to the exact logic and inputs that produced it. Replay capability allows any historical date to be recalculated using the logic version that was in effect at the time, or compared against the current version to isolate the impact of logic changes.

  • Each model registered with its input schema, output schema, and version identifier before it can run in production.
  • Results tagged with the exact model version so that any historical figure can be attributed to a specific implementation.
  • Side-by-side comparison between model versions to isolate the P&L impact of logic changes from market moves.

Outcomes

The institution gets numbers that hold up under scrutiny. Results are consistent across teams, reports, and time periods. Every reported figure has a clear audit trail from the final number back to the underlying logic and source inputs. Decision-makers and regulators see the same numbers because they are produced by the same infrastructure.

  • A single set of numbers that trading, risk, finance, and compliance all reference.
  • Regulators and auditors can trace any submitted figure to the model, inputs, and approval that produced it.
  • Model changes go through the same review and promotion process as code changes.

Where we've applied this

We applied this mandate at BatteryOS for storage optimization and dispatch calculations; in our institutional energy trading engagement for mark-to-market, P&L attribution, and risk analytics; and at Greenflash for energy portfolio valuation. The signal that drives this mandate is Model Instability.