Interactive first
Every applet exposes a small set of meaningful inputs, updates its visual output immediately and explains the change in plain language.
Transparent examples for difficult ideas.
Quant$ense separates the rule, the calculation and the teaching assumption so learners can see what is prescribed and what is illustrative.
Traceable calculations
FRTB lessons show intermediate rows for weights, correlations, tails, liquidity scaling, desk tests and capital assembly.
Source-aware
FRTB references link to PRA2026/1. Stochastic lessons point learners to Shreve’s books and identify the scope of each example.
Risk sensitivities
The risk path compares simplified full repricing with local derivatives. Bonds use annual coupons and flat continuous discounting; swaps use a single curve; CDS uses a flat hazard and continuous premiums; vanilla options use Black–Scholes or Garman–Kohlhagen. Units and sign conventions are shown beside each experiment.
FRTB experiments
The optimisation sandbox isolates a named SA delta component, rather than total bank capital. Data tests distinguish prescribed cadence and permission requirements from adjustable teaching tolerances; a statistical diagnostic does not certify regulatory compliance.
Educational boundaries
Examples use synthetic portfolios, simplified distributions and selected prescribed parameters. They do not replace the source documents, professional judgement, model validation, legal advice or regulatory approval.
Explore the FRTB calculations →
XVA model conventions
The XVA learning path uses deterministic teaching models to isolate the economics of each adjustment. Constant-exposure CVA and DVA integrate independent exponential default timing analytically. Funding, margin and capital examples discount assumed cash-cost streams. The wrong-way risk lesson uses two joint states with fixed average default probability. Portfolio and clearing examples are explicitly limited cost proxies.
The overview is an additive economic bridge, not a universal fair-value or bilateral replication formula. Its funding spread excludes own-default compensation; initial margin funding is kept in MVA, and ColVA is zero at matching benchmark and remuneration rates. KVA and tax examples must not be combined when the same tax gross-up is already included. Prudent valuation and model reserves are explained separately. Research and regulatory reference links accompany each lesson.
IRRBB Lab conventions
The IRRBB Lab uses PRA2025/20 currency shocks effective 1 July 2026, PRA SS31/15 Chapter 2 and Basel SRP31/SRP98. It values dated runoff cash-flow proxies with continuous discounting. Floating principal is represented at first repricing; mortgage payments amortise monthly with scenario-dependent prepayment; non-maturity deposits use a core-balance proxy.
NII is a deterministic monthly earnings projection with runoff or constant-balance replacement, deposit beta and repricing delay. Parallel scenarios are distinguished from additional curve sensitivities. This is not the full Basel standardised 19-bucket calculation: automatic-option valuation, full multi-curve dynamics and institution-specific behavioural validation are not implemented. The 15% EVE-to-Tier-1 screen is not an automatic capital charge. Each applet supplies its sources, assumptions and calculation trace.
Yield curves and liquidity
The curve lab bootstraps five annual par instruments with unit accruals and continuous zero-rate displays. Calibration instruments are repriced as a consistency check. Interpolation, overnight compounding and multi-curve examples use explicit simplified conventions, rather than reproducing the Bank of England fitted curves.
The liquidity lab implements selected Basel LCR and NSFR categories. Eligibility, haircuts, composition caps and inflow limits are identified in each example. It excludes specialised treatments and is not a complete PRA reporting calculation. Daily cash survival, concentration and collateral scenarios use separate user-defined stress assumptions.
Credit risk, modelling and IFRS 9
The Credit Risk & Loss Lab uses synthetic portfolio profiles and annual survival-weighted loss projections. Twelve-month ECL limits the default-event window, not the timing of resulting shortfalls. SICR thresholds are editable teaching policy; qualitative evidence and rebuttable delinquency presumptions are distinguished. Simplified receivables, POCI, modifications and credit-impaired assets have separate lessons.
The variable workbench fits penalised logistic models on 480 invented earlier-cohort records and evaluates them on 240 later records, with train-only imputation and scaling. IV, PSI, redundancy, calibration and leakage are diagnostics, not approval cutoffs. Portfolio tails use an exact two-regime binomial mixture with fixed marginal PD; this is not a regulatory capital model. IFRS 9, Basel ECL guidance and current PRA SS1/23 provide reference context. All presets, macro coefficients and decision thresholds are explicitly illustrative.
SIMM lab methodology
The SIMM path uses original educational explanations and reduced calculation examples referencing ISDA methodology 2.8+2512 (effective 11 July 2026). Adjustable weights, correlations and thresholds are illustrative unless a prescribed value is explicitly identified. The applets do not implement the complete calibration, all mapping exceptions or regulatory eligibility. Calculation traces show the assumptions and can be exported.
Explore the SIMM learning path →
Treasury Risk Lab methodology
Reduced bank-treasury models use synthetic balances and scenario assumptions with visible cash-flow, earnings and value calculations. General Basel LCR and NSFR rules are distinguished from selected category examples and internal behavioural parameters. References include PRA SS24/15 (current January 2026 publication), Basel liquidity principles, IRRBB and FX settlement guidance. The ALCO capstone keeps the initial balance sheet balanced and reports earnings, stressed cash headroom and a linear economic-value approximation separately.