Which of these courses is most valuable?

[Flex]

Hi,

could you rank the following courses, regarding how valuable they are for becoming an excellent financial mathematician (and for getting into an internship next year ;-) )

1. Algorithmic mathematics (based on Cormen et al. - "introduction to algorithms" and Korte/Vygen - "Combinatorial Optimization")

2. Advanced course in calculus (featuring Lebesgue-integration over n-dimensional Lebesgue measure, Fubini Theorem, Stoke's Theorem,...)

3. A course on partial differential equations and functional calculus (Lax-Milgram Theorem, Theory of L2-regularity, Hanarck's Inequality...)

4. Introduction to stochastic calculus (martingales, ruinproblem, regularity for time-continous martingales, ito-representation of brownian motion, ito integration, feynman-kac-formel, girsanovtransformation, solving the heat equation

5. An introductory course on coding (using either MatLab, C or C++ , depends on the lecturer)

I bet that most of you will favor the stochastic calculus course which I will definitely take. However, I would like to know how many and which other courses I should take.

Thanks in advance!

 

[bigbadwolf]

The pseudo-code of the (current) 3rd edition of Cormen, et al is C-like. If you're going to implement the algorithms you should ideally know enough about C (or equivalently C++) to code them -- basics like functions and arrays, and later on things like linked lists.

The introductory course on C/C++ coding -- how far does it go? On what text is it based?

 

[DominiConnor]

The most "valuable" course is a function of where you are, but I guess the ranking is :
4
3
2
1
5

5 is firmly last. Matlab is hugely common in finance, but it doesn't really help you get a job, and since it's easier to teach than C++ I fear it is >1/3 of the probability.

 

[Flex]

Hi,

@ bigbadwolf: Last year the coding was done in C.
There is an introductory course to C a few weeks before this course starts. However, it's the same time when the Baruch C++ course will take place. Do you think it's a good idea to learn C and C++ at the same time or will I get confused?

I checked the contents of course 5 again and it seems that Dominic is right. Most years they have done little or no C/C++ at all.

@ Dominic: Thanks for your ranking. How is the distance between the courses?
I don't think that I will manage to do the stochastic calculus course (which I will definitely take) and course 2 or 3, because they are all very time-consuming. So, would the first course which you ranked next-to-last still be a great benefit for Finance or not?

How much does the material in stochastic calculus that most MFE programs features go beyond the topics of course 4 mentioned above? From what I have seen, most MFE programs don't see it as a prerequisite to have the knowledge of this course so I'm wondering how exclusive this knowledge will be , compared to other undergraduates.

 

[bigbadwolf]

There is an introductory course to C a few weeks before this course starts. However, it's the same time when the Baruch C++ course will take place. Do you think it's a good idea to learn C and C++ at the same time or will I get confused?

There are some differences in syntax but I don't think you will get confused. If you don't have prior exposure to things like functions, arrays, and pointers, then the prior course in C might be useful. The differences will be that C++ will also cover objects (and all that that entails) as well as using the STL (e.g., using vectors instead of arrays).

 

[Flex]

Hi bigbadwolf.
Thanks for your answer.

I think that I will take the prior course in C although I am yet unsure if I will take the Algo-math course. How important would you rate the topics covered?

I'm thinking about just taking the course on stochastic calculus and fully concentrate on this course.

 

[bigbadwolf]

I think that I will take the prior course in C although I am yet unsure if I will take the Algo-math course. How important would you rate the topics covered?

It's not a priority for a budding financial engineer. It's a "nice to have": study of algorithms takes your understanding of programming deeper. You understand why some algorithms can be orders of magnitude faster or more efficient in resources than others. Your understanding of real code is enhanced. My own two cents is that understanding of programming and algorithms will be of more enduring value than knowledge of stochastic calculus. Others may not agree.

 

[DominiConnor]

I agree with BBW that algorithms and a good understanding of programming is mor critical than the vagaries of any single language.

If you can't learn basic C all by yourself then frankly you are not cut out for this line of work.

But algorithms by their very nature require you to think in a way that's different to your default. Quicksort is a good example of this of course, but there are many more. It follows that algos need to be taught.
For instance the obvious method to raise a complex number to a given integer power is basically
for ( i= 0 ; i != N ; i++)
{ ...}
That takes N steps, can you do it in about log (N) ?