We Are GlimmAnalytics

Providing unique and precise risk assessment and risk management solutions for financial institutions

No other risk management system vendors can offer capabilities like these.

Our system can manage large asset portfolios at frequencies from thirty minutes to two weeks. Our model is virtually universal across many asset classes. The platform is sufficiently flexible to provide the best strategy at specified risk return profiles and can be used as a stand alone product or an enhanced platform integrated with existing systems. Further, our scenario engine allows portfolio managers to trade at different frequencies and build multiple trading strategies simultaneously.

Since all our models are consistent with rational finance theory (dynamic asset pricing), GlimmAnalytics derives portfolio insurance instruments within the same financial models as used in portfolio risk estimation. We offer products of particular interest to traders in VIX futures and related products.

We go well beyond regulatory compliance requirements.

The GlimmAnalytics risk management system surpasses both current and future Basel III Accord regulatory compliance issues in two important ways: by assessing risk of more severe downturns as well as expected returns at all quantiles. These tools will help portfolio managers to manage risk and optimize gain in their portfolios.

Methodologies Have Driven Our Development

ALPHA – Profitability

Our trading strategy blends predictions across time scales ranging from two weeks to 30 minutes.  We support optimization based on expected gain vs. expected shortfall. Advanced strategies based on Bayesian methods and long range dependence are under development.

BETA – Risk

We provide risk management for a variety of time scales from two weeks and more to 30 minutes and less. The elevated risk associated with market open is identified and quantified. Risk is proven accurate by backtesting, with levels of accuracy as high as 0.1% expected shortfall tail risk.

Asset classes include equities, FX and commodities. Fixed income and options are under development. Path dependent risk measures such as maximum drawdown and margin call risk are supported. We also support measures of performance evaluation, based on predicted vs. realized portfolio returns.

Our models incorporate major identified features of market data, including heavy tails, volatility clustering, long range dependence, leverage effect and a time varying covariance matrix. Our models follow the arbitrage theory of pricing, so that the same model is used to assess risk and to optimize or mitigate it through hedging and risk budgeting.

2023 News from GlimmAnalytics

GlimmAnalytics’ CEO, James Glimm, and coauthors, Daniel Lazarev, Gui-Qiang, Hamid Said, Jarret Petrillo, and Min Lee, submitted a solution, January 2023, to the $1M Millennium Prize Mathematics, Fluid Dynamics.

This problem is one of 7 Millennium Prize problems proposed in 2000, with only one other having been solved. This problem solves for the existence of a smooth solution for the Navier-Stokes equation. The Navier-Stokes equation is the fundamental equation describing turbulent flow.

The solution, which actually has relevance to financial modeling, is that the smooth solution is identified as the average value of the turbulent flow. This solution is not turbulent, as all turbulent fluctuations are suppressed in the averages in the solution. It is on this basis that the solution acquires its smoothness.

In relation to financial models, risk models play a similar role. They suppress all volatility fluctuations and are an average or mean value for expected future financial returns. Due to the increased regularity that results from suppression of fluctuations, risk models can be developed with analytical formulas. and backtested, showing high degrees of accuracy.

Alpha models in an extreme can be purely fluctuation based, as in the Black-Scholes algorithm for option trading, which is completely independent of mean values. The real world is less extreme and in it, alpha models benefit from a coupling to beta (mean) models.