quant skills ( VAR, greek analysis?)
- Thread starter congoboy
- Start date
alex7,
truly multidimensional reflecting...
Only an ignorant wreckless individual, totally unaware of the subject at hand, would expose his knowledge starvation and show off his profonity skills in the lack of proper vocabulary and claim an advance knowledge of programming is not necessary in C++ or the likes to become a proficient "Kuant(bet this is how he spells it)" when his total ignorance does not permit him to realize that any quant analysis is 100% Matrix based and the one most important aspect of Matrices are, the index and the pointer!
The biggest sigh... I wish the admin would control such foul language thread trashing artists and confine them to the Hooligan threads only.
"Quants often come from physics, engineering or mathematics backgrounds rather than finance related fields, and quants are a major source of employment for people with physics, mathematics, and engineering Ph.D's. Typically a quant will also need extensive skills in computer programming."
The biggest sigh... I wish the admin would control such foul language thread trashing artists and confine them to the Hooligan threads only.
"Quants often come from physics, engineering or mathematics backgrounds rather than finance related fields, and quants are a major source of employment for people with physics, mathematics, and engineering Ph.D's. Typically a quant will also need extensive skills in computer programming."
??? This has to be one of the most misinformed and strangest posts I have read here.
I don't think you have any idea of what quantitative finance requires, and you are completely wrong on almost all counts.
Firstly, it used to be the case that Quants came from Maths, Physics (particularly astro and particle physics) and Engineering (Control and DSP), but this is no longer the case. These days there exist excellent programs in Mathematical Finance and Financial Engineering, as well as that many modern finance courses have a substantial mathematical content available to students. The Chicago course with Steve Shreve comes to mind.
Secondly, and although linear algebra is a significantly important aspect of quantitative analysis, it is not 100% of the mathematical skillset needed by quants, as your post appears to imply. Some might say that modern mathematical probability is more important, others might say partial differential equations are. I have worked with quants whose PhD's were in Group or Number Theory, not really applied stuff, so a varied lot. These days we are seeing more come from specialized graduate finance courses with good math stat and coding skills.
Thirdly, and I believe I know the posts you are alluding to, I don't see where any one said that programming is not necessary to do the work of a quant. Most quants are quite proficient in C++, Matlab, Mathematica,VBA, Excel, Gauss - depends on where and what you work on. What a quant does not need, which I believe was the idea conveyed in the post, is extensive knowledge of the dustier corners of C++. There's no need for that. Having encyclopedic knowledge of really obscure C++ does not mean you can code. I have seen many examples of this, and this is normally a terrible interviewing tactic. Any good quant who can code should have a good knowledge of sound software engineering principles and apply them to write fast, efficient and robust code. This is tough to do if you are a FO quant and you have a trader yelling at you to get the job done like yesterday! Some interviewers might ask you to code a couple of classes and price something - fairly basic stuff. These are the good ones. The bad ones will ask you very specific dusty corner C++ questions - this is utter rubbish. If you get that in an interview, then get up and run. its probably a second rate place and not good to work in. This has been my experience. When I was interviewing potential candidates a few years back, I changed the C++ questions. No longer would we ask rubbish like bitfield structures in C++ (which is straight C anyway) and trick questions like what is a virtual constructor (there is none, this is the stupid name sometimes given to the factory method pattern), etc, etc. Actually asking deceptively simple questions allowed me to gain significant insight whether the person had textbook knowledge, some experience, or significant experience and sound software engineering skills.
HTH.
I don't think you have any idea of what quantitative finance requires, and you are completely wrong on almost all counts.
Firstly, it used to be the case that Quants came from Maths, Physics (particularly astro and particle physics) and Engineering (Control and DSP), but this is no longer the case. These days there exist excellent programs in Mathematical Finance and Financial Engineering, as well as that many modern finance courses have a substantial mathematical content available to students. The Chicago course with Steve Shreve comes to mind.
Secondly, and although linear algebra is a significantly important aspect of quantitative analysis, it is not 100% of the mathematical skillset needed by quants, as your post appears to imply. Some might say that modern mathematical probability is more important, others might say partial differential equations are. I have worked with quants whose PhD's were in Group or Number Theory, not really applied stuff, so a varied lot. These days we are seeing more come from specialized graduate finance courses with good math stat and coding skills.
Thirdly, and I believe I know the posts you are alluding to, I don't see where any one said that programming is not necessary to do the work of a quant. Most quants are quite proficient in C++, Matlab, Mathematica,VBA, Excel, Gauss - depends on where and what you work on. What a quant does not need, which I believe was the idea conveyed in the post, is extensive knowledge of the dustier corners of C++. There's no need for that. Having encyclopedic knowledge of really obscure C++ does not mean you can code. I have seen many examples of this, and this is normally a terrible interviewing tactic. Any good quant who can code should have a good knowledge of sound software engineering principles and apply them to write fast, efficient and robust code. This is tough to do if you are a FO quant and you have a trader yelling at you to get the job done like yesterday! Some interviewers might ask you to code a couple of classes and price something - fairly basic stuff. These are the good ones. The bad ones will ask you very specific dusty corner C++ questions - this is utter rubbish. If you get that in an interview, then get up and run. its probably a second rate place and not good to work in. This has been my experience. When I was interviewing potential candidates a few years back, I changed the C++ questions. No longer would we ask rubbish like bitfield structures in C++ (which is straight C anyway) and trick questions like what is a virtual constructor (there is none, this is the stupid name sometimes given to the factory method pattern), etc, etc. Actually asking deceptively simple questions allowed me to gain significant insight whether the person had textbook knowledge, some experience, or significant experience and sound software engineering skills.
HTH.
The above quote was an extract from the below just referring to the programing part being as important as all other aspects and requirements for becoming a quant.
On the issue of 100% of the task at hand being Matrix based, I was not referring to the part that purely requires mathematical operations and meant the more complex analysis which are all data mining related, similar to sophisticated "pattern recognition" analysis and, then all data mining ops are 100% Matrix based.
"Education
Quants often come from physics, engineering or mathematics backgrounds rather than finance related fields, and quants are a major source of employment for people with physics, mathematics, and engineering Ph.D's. Typically a quant will also need extensive skills in computer programming.
This demand for quants has led to the creation of specialized Masters and PhD courses in mathematical finance, computational finance, and/or financial reinsurance. In particular, Masters degrees in financial engineering and financial analysis are becoming more popular with students and with employers. Carnegie Mellon's Tepper School of Business, which created the Masters degree in financial engineering, reported a 21% increase in applicants to their MS in Computational Finance program, which is on top of a 48% increase in the year before[4]. The University of California Berkeley's program in Financial Engineering through their Haas School of Business admits 60 students each year. Other well known programs are provided by the University of Chicago, Cornell University, Indiana University, Columbia University, Purdue University, and New York University. This surge in popularity has led other schools (the University of California at Los Angeles, Rutgers University, Polytechnic University, the University of Michigan, the University of Minnesota) and the Nanyang Technological University (Singapore) to add Masters level degrees. These Masters level programs are generally one year in length and more focused than the broader MBA degree."
On the issue of 100% of the task at hand being Matrix based, I was not referring to the part that purely requires mathematical operations and meant the more complex analysis which are all data mining related, similar to sophisticated "pattern recognition" analysis and, then all data mining ops are 100% Matrix based.
"Education
Quants often come from physics, engineering or mathematics backgrounds rather than finance related fields, and quants are a major source of employment for people with physics, mathematics, and engineering Ph.D's. Typically a quant will also need extensive skills in computer programming.
This demand for quants has led to the creation of specialized Masters and PhD courses in mathematical finance, computational finance, and/or financial reinsurance. In particular, Masters degrees in financial engineering and financial analysis are becoming more popular with students and with employers. Carnegie Mellon's Tepper School of Business, which created the Masters degree in financial engineering, reported a 21% increase in applicants to their MS in Computational Finance program, which is on top of a 48% increase in the year before[4]. The University of California Berkeley's program in Financial Engineering through their Haas School of Business admits 60 students each year. Other well known programs are provided by the University of Chicago, Cornell University, Indiana University, Columbia University, Purdue University, and New York University. This surge in popularity has led other schools (the University of California at Los Angeles, Rutgers University, Polytechnic University, the University of Michigan, the University of Minnesota) and the Nanyang Technological University (Singapore) to add Masters level degrees. These Masters level programs are generally one year in length and more focused than the broader MBA degree."
I believe you are quoting a Derman article if memory serves, perhaps you should give credit/reference where it is due.
OK, two things.
Firstly, pattern data mining is only a subset of the work a quant might be engaged in, and I would not say it is the most complex work for a quant either. Typically, the higher ranking quants work on FO mathematical models.
Secondly, no-one has argued that software construction is not an important aspect of a quant's work so I don't know where you are inferring this from, although there are pure quant developer roles today where the emphasis is on the coding not the modeling as much. What has been stated is that interviewing a quant and asking questions from the dusty corners of the C++ language standard (as has been the a case many times) is a stupid interview process as it is virtually uncorrelated to a quant's programming abilities, and that requires writing efficient, fast and robust code. Some places use this technique because that is what was done to them, and the stupidity is propagated further. This is all nonsense and is one of the reasons why most quant code is utter crap to debug and re-engineer, having a trader yelling at you doesn't help this either. I prefer to leave someone in a room with a laptop and a C++ compiler and ask them to write a simple monte carlo pricer, implement Black Scholes, or something like that, for a simple option (asset or nothing, european call, etc). The good ones are finished well within an hour (yes, quant interviews take many hours).
That gives me a better idea of what the person can do in a real situation, rather than some idiotic C++ puzzle which is nothing more than pure macho BS.
Just my 2c.
OK, two things.
Firstly, pattern data mining is only a subset of the work a quant might be engaged in, and I would not say it is the most complex work for a quant either. Typically, the higher ranking quants work on FO mathematical models.
Secondly, no-one has argued that software construction is not an important aspect of a quant's work so I don't know where you are inferring this from, although there are pure quant developer roles today where the emphasis is on the coding not the modeling as much. What has been stated is that interviewing a quant and asking questions from the dusty corners of the C++ language standard (as has been the a case many times) is a stupid interview process as it is virtually uncorrelated to a quant's programming abilities, and that requires writing efficient, fast and robust code. Some places use this technique because that is what was done to them, and the stupidity is propagated further. This is all nonsense and is one of the reasons why most quant code is utter crap to debug and re-engineer, having a trader yelling at you doesn't help this either. I prefer to leave someone in a room with a laptop and a C++ compiler and ask them to write a simple monte carlo pricer, implement Black Scholes, or something like that, for a simple option (asset or nothing, european call, etc). The good ones are finished well within an hour (yes, quant interviews take many hours).
That gives me a better idea of what the person can do in a real situation, rather than some idiotic C++ puzzle which is nothing more than pure macho BS.
Just my 2c.
Chicago Mellon University?
LOL, too true. Yeah, for some reason I keep thinking Shreve left CMU and went to Chicago.Quote from caementarius:
Chicago Mellon University?
Apparently Neil Chriss is at Chicago. I envy any current PhD students who have him as a supervisor. Top notch researcher and practitioner (GS, SAC).