Learn the method.
Test the assumptions.
Build the capability.
A practical learning environment for investment and risk management. Eight worked models across four disciplines — yield-curve construction and forecasting, bond valuation, credit risk and interest-rate derivatives — each built so that every assumption is visible and every result can be reproduced and questioned. Intended for teams who want to understand the methods behind sound practice, not to accept a figure on trust.
Level · Slope · Curvature
Illustrative Nelson-Siegel yield curve.
Four connected models, one teaching sequence
A complete worked chain, designed to be followed step by step: decompose an observed curve into three factors, project those factors forward under competing assumptions, translate the paths into a full return distribution for any duration, then construct an optimal allocation across durations. Each stage feeds the next, so the effect of an early assumption can be traced all the way through to the final allocation — which is the point of the exercise.
Fit the yield curve
Recover the Nelson-Siegel level, slope and curvature betas from market quotes by ordinary least squares, with fit diagnostics.
Open model →Forecast the factors
Project betas forward under eight drift regimes — from random walk and mean reversion to a jointly-fitted VAR(1), a neural net and a Gaussian process — with Monte-Carlo fans and in-sample skill diagnostics.
Open model →Calculate scenario returns
Turn simulated curves into holding-period returns per bond — expected return, volatility, percentiles and Value-at-Risk — simulated jointly so bonds on different curves share a historically-correlated shock.
Open model →Optimize the portfolio
Solve the allocation across your Returns bond universe under six objectives and real position limits — minimum variance, maximum Sharpe, target return, maximum diversification and more.
Open model →Start from the bonds, not from a curve
Where the models above begin with an already-built curve of tenor and yield quotes, this pair starts one step earlier — from individual bonds with real cashflows, coupon schedules and day-count conventions — and derives the curve from them. It is the harder case to get right, and the more instructive one: it is also the situation most often faced in thin, sparsely quoted markets where a handful of bonds is all there is.
Fit the curve from a bond universe
Bootstrap an exact-fit discount curve and calibrate Nelson-Siegel or Svensson against it, with liquidity and staleness weighting, a robust loss so one bad quote can't pull the curve, and arbitrage checks on the result.
Open model →Price a portfolio off it
Value a government and corporate book against that fitted curve, applying credit spreads to corporates only — with every cash flow's spot rate, spread, discount factor and present value shown, so the price reconciles by hand.
Open model →Simulate portfolio credit risk
A separate discipline from asset allocation, and a foundation of any sound risk framework: upload positions and a correlation matrix, then simulate rating migration and default across counterparties — decomposed by issuer and by risk type, with Expected Credit Loss and CVaR at any confidence level.
SOFR futures & options Value-at-Risk
A worked example of how a derivatives book is measured: upload an options and futures book with market data, fit a factor model to the curve, simulate rates forward, and compute VaR and CVaR at the position, strategy, trader and portfolio level under base and stress scenarios.
Strengthening practice across the financial sector
The platform is intended for smaller and medium-sized asset management and investment management companies, pension funds and other financial institutions — organisations with capable people and real responsibilities, but limited access to structured training in quantitative methods. It is equally intended for onboarding new staff, retraining experienced colleagues, and giving risk, compliance and internal control functions a concrete view of how the numbers they oversee are actually produced.
Working through the models is a way to build competence in areas that underpin sound institutional practice: governance of the investment process, risk measurement and its limitations, internal controls, compliance with policy limits, and operational resilience. The methods reflect internationally recognised principles and widely accepted good practice, presented in a form that can be studied, discussed and adapted to an organisation's own circumstances.
Transparent by design, so it can be taught and reviewed
Every assumption is an explicit, editable input, and every figure can be reproduced from the inputs on screen — which is what makes the models usable as teaching material and as a basis for informed discussion. The platform is a learning and analytical aid; it is not a production system for day-to-day investment decisions.