SciFinance is ground-breaking software for building derivative pricing models and model calibration. Using an intuitive, very high-level programming language (VHLL) for describing financial contracts and numerical methods, SciFinance provides a friendly, versatile environment in which to make and implement modeling decisions. Specify the choices that are important, then let SciFinance handle the rest using its extensive knowledge base.
SciFinance eliminates programming by automatically translating model specifications for any financial derivative into fully documented C family source code in minutes. SciFinance generates wrapper code (in Java, Python, .xll, COM, or .NET) to automate integration without imposing proprietary data models.
With over over 15-years of input from practitioners at top tier institutions worldwide, SciFinance is a cost-effective and flexible in-house derivatives pricing modeling technology with unique features:
No limitations when defining instrument features, terms and conditions of the contract, underlying model dynamics, numerical methods, market data and its format, and model outputs.
No "black box" components and users have full control through all stages of pricing model development.
No imprecise or limited functionality; users drill down as far as they wish into modeling decisions then let SciFinance implement the rest based upon its extensive knowledge base.
SciCalibrator module of SciFinance lets users develop and integrate their own calibration routines.
Customers can create an unlimited number of pricing models with SciFinance, owning them in perpetuity. SciFinance-generated models have no run-time licenses, so they may be used by an unrestricted number of end users.
Automatically generates CUDA or OpenMP-compliant source code for any Monte Carlo-based pricing model (runs on NVIDIA GPUs). Absolutely no CUDA or parallel programming expertise is required.
SciComp Consulting, provides ready-to-use, efficient pricing models that can be precisely tailored to customer specifications. Unlike vendors who rely on pre-built libraries or toolkits, SciComp Consulting custom designs each model specification in accordance with customer requirements using state-of-the-art numerical methods. Features such as industry-standard or proprietary underlying model dynamics,financial contract provisions, and calibration routines can be combined to produce customer specified pricing models for any asset class.
Ready-n-Customizable Calibrators are a suite of robust, standalone, ready-to-use calibration functions any of which can be tailored to meet customer requirements. In addition, SciComp Consulting can quickly and economically implement custom calibrators for customers looking for either public domain or proprietary calibration routines that are tailored to their specific requirements.
Utilizes the SciComp extended classical log-normal model to incorporate volatility smiles.
Based on the work of
Ren, Madan and Qian [Risk, Sept 2007], Lipton [Risk, Feb 2002],
Jex, Henderson and Wang [J.P.Morgan, 1999], and the Bloomberg paper "Stochastic Local Volatility" by Tataru & Fisher.
A least-squares calibration of a Heston model via Levenberg-Marquardt. The Heston model assumes that the underlying asset follows a Black-Scholes process with a stochastic volatility. The Heston model may include asset jumps and be piece-wise constant.
Applicable to any portfolio of European options written on a single asset.
A least-square calibration of either a 1-factor, constant parameter Gaussian model, or a 2-factor, constant parameter Gaussian model by means of Levenberg-Marquardt.
A calibrator (in a least-squares sense) for a constant coefficient Gabillon model to a collection of at-the-money future contracts.
A calibrator (in a least-squares sense) for a constant coefficient Schwartz97 model to a collection of at-the-money future contracts.
An implied correlation for the given expected loss specified by credit tranche spreads.
Includes credit index-based structures
(e.g., DJ Itraxx, etc.)
SciComp Consulting can quickly and economically implement custom calibrators for users looking for either public domain or proprietary calibration routines that are tailored to their specific needs and requirements. Custom Calibrators are available for a broad range of pricing models including:
- Equity/FX/commodity models including many parameterized local volatility, stochastic volatility (with and without jumps), and pure jump models.
- Generic short rate models (including popular Gaussian and lognormal flavors) with interest rate and/or hazard rate calibration to volatility term structure of cap/swaptions and CDS spreads or corporate bonds.
- LIBOR market model calibration, including exact fit to caps and least square fit to swaptions. Several parameterizations of the correlation matrix are available as well as several approaches for including volatility skew.
- Exchange Options
- Credit models, including survival/hazard rates from credit spreads, base correlation and semi-analytic models with deterministic or stochastic factor loading.
Many models accommodate time dependent parameters, either exactly through numerical models of the calibration instruments, or through very fast approximate analytic techniques.
Available optimization techniques include a robust Levenberg-Marquardt algorithm and simulated annealing.