Aleksander Hansen
Well-Known Member
Just thought I'd share a few thought on the topic, and give some pointers.
A nice post on GARCH (and at the end volatility targeting) in practice: here.
Also, I would recommend not performing any optimization in Excel. It's great for a lot, but data analysis and optimization is not one of them.
Pat [who writes the blog I link to] has some interesting posts from time to time, and for those of you who are trying to implement this, it's a great resource as he was one of the early developers of the S+ language (largely superseded by R).
You can implement the models for free if you want by programming (or copying and pasting) in the R programming language (which is free and is available here: CRAN)
If you want to be a "quant" you should consider learning a language such as R or Matlab (www.mathworks.com) the latter is not free though. The two languages are almost identical and while I used to prefer Matlab, I only use R nowadays.
Reason to learn it?
Chances are that you will need to create a model that Excel can't handle (see comment above) in any quant job (actually it can be great in a number of different jobs that are not "quant like at all", that is, there are certain things you can accomplish [in R, S+ or Matlab] in a few minutes that would take you hours to do in Excel).
You can always go for a low level C-family language (be aware that C++ in particular is becoming more and more obsolete). Indeed, if you want knowledge of a low-level language, most companies today will want to see C.net/C#, Java or Objective-C instead (and many companies are paying student interns to re-write C++ code into code in those languages. That's in part why a lot of CS departments start out with C in the basic courses, and then have other optional courses where object oriented programming is used.) The main reason you want to learn a program like R, S+ or Matlab though is that a quant <> programmer, and you can complete a model in R or Matlab in a fraction of the time it takes to do it in the lower level languages. Besides, nowadays, you can actually convert your code from R or Matlab into C code automatically.
Disclaimer: Pat is an acquaintance of mine and he sells a portfolio management program. His blog posts are generally separate from that though, and I do not receive any compensation, nor do I have own any shares in any of his companies.
A nice post on GARCH (and at the end volatility targeting) in practice: here.
Also, I would recommend not performing any optimization in Excel. It's great for a lot, but data analysis and optimization is not one of them.
Pat [who writes the blog I link to] has some interesting posts from time to time, and for those of you who are trying to implement this, it's a great resource as he was one of the early developers of the S+ language (largely superseded by R).
You can implement the models for free if you want by programming (or copying and pasting) in the R programming language (which is free and is available here: CRAN)
If you want to be a "quant" you should consider learning a language such as R or Matlab (www.mathworks.com) the latter is not free though. The two languages are almost identical and while I used to prefer Matlab, I only use R nowadays.
Reason to learn it?
Chances are that you will need to create a model that Excel can't handle (see comment above) in any quant job (actually it can be great in a number of different jobs that are not "quant like at all", that is, there are certain things you can accomplish [in R, S+ or Matlab] in a few minutes that would take you hours to do in Excel).
You can always go for a low level C-family language (be aware that C++ in particular is becoming more and more obsolete). Indeed, if you want knowledge of a low-level language, most companies today will want to see C.net/C#, Java or Objective-C instead (and many companies are paying student interns to re-write C++ code into code in those languages. That's in part why a lot of CS departments start out with C in the basic courses, and then have other optional courses where object oriented programming is used.) The main reason you want to learn a program like R, S+ or Matlab though is that a quant <> programmer, and you can complete a model in R or Matlab in a fraction of the time it takes to do it in the lower level languages. Besides, nowadays, you can actually convert your code from R or Matlab into C code automatically.
Disclaimer: Pat is an acquaintance of mine and he sells a portfolio management program. His blog posts are generally separate from that though, and I do not receive any compensation, nor do I have own any shares in any of his companies.
