QUANTITATIVE RISK

VaR. CVaR. RWA. Monte Carlo. One platform.

The same engine that audits the contract also runs the numbers. Parametric VaR with stress overlays, Basel III SA Risk-Weighted Assets, multi-asset Monte Carlo, dynamic risk matrices and a live market-data feed โ€” all under the same audit chain as the rest of the analysis.

5 stress scenariosBasel III SAMonte Carlo IRCAlpha Vantage liveMulti-LLM consensus

VAR ยท CVAR

Parametric VaR with historical stress overlays.

Value at Risk and Conditional VaR computed on a single position or a multi-asset portfolio. Historical stress overlays apply the empirical shock of five reference dislocations so the same portfolio can be expressed at 95%/99% confidence in calm market AND under tail conditions. Output: VaR, CVaR, expected shortfall, contribution by leg.

Stress scenarios applied

Black Monday

S&P 500 -22.6 % single session

1987

Global Financial Crisis

Lehman collapse, credit-spread blow-out

2008-09

Volmageddon

Short-VIX ETPs liquidation, VIX +115 %

Feb 2018

COVID Crash

WTI negative, equity -34 % in 5 weeks

Mar 2020

Russia 2022

TTF +500 %, RUB / equity simultaneous shock

Feb 2022

RWA ยท BASEL III SA

Risk-Weighted Assets calculator (Standardised Approach).

Basel III SA risk weights applied per asset class and external rating. Used by the engine to compute the capital impact of a deal before it is signed โ€” useful for credit committees that need the RWA number alongside the legal opinion.

Asset classExternal ratingRisk weight
Sovereign exposureAAA โ†’ AA-0 %
Sovereign exposureA+ โ†’ A-20 %
Bank exposureAAA โ†’ AA-20 %
Corporate exposureAAA โ†’ AA-20 %
Corporate exposureA+ โ†’ A-50 %
Corporate exposureBBB+ โ†’ BB-100 %
Retail / mortgageregulatory35 %
Unrated exposureโ€”100 %

Risk weights from BCBS Basel III SA. The engine returns the full breakdown per leg; this table is the headline reference.

DYNAMIC RISK MATRIX

Severity ร— probability, rendered as you go.

Each finding the engine surfaces is plotted on a severity ร— probability grid. The view is reactive โ€” adding a new finding or accepting a mitigation re-renders the matrix in place. Useful for credit and risk committees that want a one-screen view of the deal posture.

MONTE CARLO

IRC scenario simulation.

Configurable variables (price corridors, FX, default probabilities) feed a Monte Carlo engine that returns the IRC distribution and stress percentiles for the contract. Same engine that powers Mode 04 in the analysis catalogue โ€” exposed standalone here for quant teams that already know what they want.

See mode 04 in /finanzas/modos โ†’

MARKET FEEDS ยท ALPHA VANTAGE

Live quotes injected into the analysis.

Live market data from Alpha Vantage โ€” equities, FX and commodities โ€” pulled at request time and injected into the relevant analysis. Bulk fetch up to 10 symbols per call. Used by the Commodities mode (mode 06) automatically when the contract references benchmark commodities (TTF, Brent, WTI, copper, natural gas, etc.).

EquitiesFXCommoditiesBulk 10 symbolsAuto-inject on commodity refs

CONSENSUS ENGINE

Three LLMs. One verdict.

Adversarial analysis with up to three LLMs running simultaneously. Outputs are compared on Jaccard similarity and JSON divergence; only what all three validate reaches the partner's screen. Designed for the few cases where the cost of being wrong is asymmetric โ€” credit decisions, sanctioned-party exposure, criminal-risk flags.

Anthropic Claude
DeepSeek V4
Google Gemini

Want to run the numbers on your own portfolio?

Bring a representative position and we will execute the VaR / RWA / Monte Carlo stack on it, with the cryptographic chain on, before any commercial conversation. Same offer as the rest of Finance: due diligence first, paperwork later.