Interview: Patrick Burns Quantitative Finance in R

[This article was first published on Quantitative Finance Collector, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Patrick BurnsDr. Patrick Burns is the founder of Burns Statistics, providing consulting and bespoke software specializing in quantitative finance, programming in the S language, and optimization via genetic algorithms and simulated annealing. Patrick has written many papers on quantitative finance and statistics, he is also the author of the book The R Inferno and the R package BurStFin.  

Tell us a little background info about yourself. Where are you from? What’s your education background?


I was born on the edge of a wheat field in the Empty Quarter.  I made my way to Seattle for university where I received a PhD in statistics (with an emphasis on computing and a smattering of economics).  Much later I moved to London.

In graduate school one of my office mates was Robert Gentleman, who would a few years later be half of the team that originated R.

How long have you been using the R language and to what extent? What are the main reasons you choose to run analysis in R rather than other languages?


I first touched R in the early 90’s when Robert came around with it on his laptop.  However I didn’t seriously make the switch from S-PLUS to R until I started Burns Statistics in 2002.

A big reason I use R is because I used to be a developer of S-PLUS (R’s sibling) and so I’m naturally fluent in R.

But there are what I think are good reasons why others should also use R.  The key thing is that the S language – and so R – was specifically designed for data analysis.  A lot of experience, thought and arguments have gone into creating what R is now.  Data analysis is really what quantitative finance is about, hence R was designed for quant finance.

Another reason is that it is looking like R will soon be the dominant data analysis platform.  I continue to be astounded at the growth of R, especially in finance.  It’s not good to use an inferior tool just because everyone else uses it (read Excel).  But if you have a selection of reasonable tools, then there are advantages to using the most popular of the good choices.
One of the advantages that R has over Matlab is that it is free.  You can think about installing R on all the machines on a trading floor so that everyone on the floor can use some functionality (probably without knowing that it depends on R).  Such a scheme would be very expensive with a commercial language.

An advantage that R has over other free languages is its wealth of functionality.  There are over 3000 packages in the main R repository. The development of packages specific to finance is quite active.  R most definitely has the attention of the statistics community, so new statistical techniques are most likely to appear first in R. Also R tends to play nice with others – you can often easily mix R and tasks done in other languages.

In short sentences, what are the major differences between Portfolio Probe developed by Burns Statistics and other free or paid packages / software?


The most important thing in Portfolio Probe is the generation of random portfolios.  That is, sample from the population of portfolios that obey some set of constraints (for example: sector constraints, weight constraints, number of names).  Generating random portfolios is a general quant tool  that has many uses, most of them yet to be discovered.  I think their main impact will be in performance measurement. Currently they are mainly used to test risk models.

The other thing that Portfolio Probe does is portfolio optimization.  The optimization uses a form of genetic algorithm.  This has the advantage of flexibility over optimizers that use mathematical algorithms.  The optimizer can adapt to the problems people actually have rather than trying to make the problem adapt to the optimizer.  The Portfolio Probe optimizer is better than other heuristic optimizers because a lot of work has gone into making it work well.  Some of that work was successful.

What’s the biggest change you feel the credit crisis has brought to the development of quantitative finance?

  
One change is that it has pushed quants away from the illusion that their models are true.  That’s a good thing, but unfortunately probably temporary.  People will be lulled into complacency once their models have worked well for a while.
The other major change I see is that it has prompted more thought on hard but important problems.  I’m thinking of things like understanding herding risk, and the real dynamics of markets.

Correlation between assets has become higher since the breakout of financial crisis, which makes portfolio management more difficult (for example, less diversification), what’s your advice for an investor to better manage his portfolio? What’s the impact on portfolio optimization?

  
I don’t have advice for investors – at least none that they should take.  However, this is an example where random portfolios can be informative.  The actual opportunity that a fund manager had over historical periods can be seen by generating a set of random portfolios that satisfy the fund’s constraints and then getting the distribution of returns for the portfolios over time.  The opportunity given a certain set of constraints may not follow average correlation very well.

As for its effect on portfolio optimization, that’s a good question to which I don’t know the answer.  Perhaps you’ll see a blog post on that some day.

What accomplishments so far are you the most proud of?


My greatest technical achievement is probably the speed of optimization and random portfolio generation in Portfolio Probe.  I of course want it to be still better, but I think it’s pretty good now.

One of my best conceptual achievements has been realizing the immense potential of random portfolios.  The achievement that I’m still missing is convincing the rest of the world of that.

What is the single toughest challenge you’ve had to face in your past projects, and how did you get through it?


I’m not convinced I’m “through it” but I’ve made progress.
Consulting has a technical component and a social component.  I’m okay with technical, but human society is a foreign language to me.  Some people take that to mean that I’m anti-social.  That’s not it – I like people, I just don’t understand the rules of interaction very well.  Things that seem to be innate to elementary school children are mysterious to me.  However, I’m massively better than I used to be.  The less observant might even confuse me for normal now.

What have you been up to recently? What projects are you working on?


I’m in the final phases of putting out the official release of the new version of Portfolio Probe.

Can you describe a typical work day? How do you like to spend your free time?


My typical day begins by waking up to the battle with Alfred – the cat who owns the house where I live – over strategic territory on the bed.  Once I admit defeat, I brace myself for the arduous commute of going one flight down the stairs.
I spend the morning playing with the buttons on my computers.  I have lunch with my wife (and of course Alfred).  Then I go for a walk in the park, and say hello to my dog friends.  This is often the most productive part of the day.  Once back, I have another session of playing with computer buttons.

How can people contact you for business? Do you have a website or Twitter account or Facebook “Like” page?


There is a website for Burns Statistics: http://www.burns-stat.com
Among other things, this has information on using R including an introductory set of pages and The R Inferno.
There is also the Portfolio Probe website: http://www.portfolioprobe.com that includes the blog.

On Twitter I’m @portfolioprobe  – this is mainly announcements of blog posts, but sometimes has something about R, and occasionally a joke that hardly anyone gets.
Tags – r , quant , portfolio , random
Unclear about this post? Read the full post at Interview: Patrick Burns Quantitative Finance in R or Asking questions and receiving answers.
—supported by Best selling investing books

To leave a comment for the author, please follow the link and comment on their blog: Quantitative Finance Collector.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)