15 Nov

About Burns Statistics

Burns Statistics was established in 2002 as the consulting and software vehicle for Patrick Burns.

Specialties are:

  • programming in the R language
  • statistics, including quantitative finance
  • solving difficult problems via heuristic optimization (genetic algorithms and simulated annealing).



Patrick has well over a decade of experience as a statistical consultant mostly in the financial and medical industries. Assignments have included analysis of equity data with R, building GARCH modules in S, tutoring S and statistics, automated trading, risk management, and evaluating the efficacy of medical devices.

Software Development

Patrick was a lead developer of S-PLUS in its early years. His tasks included adding functionality (such as principal components and robust estimation in several settings), finding and fixing bugs, and documentation.


In addition to his consulting experience, Patrick was an employee of Schroder Salomon Smith Barney (Citigroup) in London for four years. While there, he developed statistical models for equities — both client-facing and proprietary. He also wrote in-house software for numerous applications including portfolio optimization and risk management.


In 1988 Patrick received a PhD in statistics from the University of Washington in Seattle. His thesis topic was the robust analysis of designed experiments — that is, developing techniques that perform well even when outliers exist in the data.


See also Working Papers.

Patrick Burns on Google Scholar.

A very similar version is available as Dart to the Heart

  • Burns, P. (2007). Random Portfolios for Performance Measurement. Appears in Optimisation, Econometric and Financial Analysis E. Kontoghiorghes and C. Gatu, eds. Springer.

This is based on the working paper Performance Measurement via Random Portfolios but has some additional material.

This is derived from Discussion Paper 97-30 at the University of California, San Diego Department of Economics

  • Fraley, C. and P. Burns. (1995). Large-scale estimation of variance and covariance components. SIAM Journal on Scientific and Statistical Computing. 16 , number 1.
  • Burns, P. (1991). A graphical display for choosing a transformation. Proceedings of the 23rd Symposium of the Interface: Computing Science and Statistics. p42-45.
  • Burns, P. (1990). The L1 solution set in two-way tables. Utilitas Mathematica 37 p233-250.
© Copyright - Burns Statistics