which school do you think will accept me?

[ironmonkeyz]

So i graduated from University of Toronto with a BBA in Finance, and am currently working as a trader at a prop shop. Im thinking of applying for MFE as I am very interested in the topics covered and wish to learn more quantitative tools to add to my trading strategies. However, as a BBA in Finance I have had little exposure to math.

Assume I go back to take non degree courses in math and stats, and score fairly well on GRE, what do you guys think are my chances of getting accepted to a mfe program and which schools do you think will most likely accept me?

I'm feeling a huge lack of confidence regarding getting admissions since I'm from a non-quantitative background. Please let me know what you guys think. Thanks a bunch

 

[GoIllini]

Hi ironmonkeyz.

You have a strong professional background for an MFE admit- many of these programs would like to take proprietary traders. However, here is what I typically recommend in terms of academic background for an MFE:

1.) Calc I (derivatives)
2.) Calc II (integral calculus)
3.) Calc III (multivariable calculus)
4.) Linear Algebra
-> You should be able to remember a few details of the invertible matrix theorem, remember whether you can multiply a 1xn matrix with an nx1, and maybe have a hazy memory of eigenvalues and eigenvectors.
5.) Calculus-based probability, or at least some stats.
-> If I tell you that you have a uniform distribution from 0 to 10 and I want you to calculate the expectation of X^3, you should know (or be able to quickly remember) how to do that. If I give you a joint probability density function of two variables X and Y, you should be able to remember how to calculate the covariance.
6.) A basic programming course- something like Python, Java, or C++.
7.) Maaybe differential equations, but you can survive without it for the most part.

Schools might take you if you have less of these classes than what I recommend, but not if you have waaay less than this. The good news is that any state school for a big city (EG: UIC, Baruch, UWM) is good enough for most of these schools in terms of showing you know your stuff on the math front.

So how many of these courses do you have, and is there a local school that can teach the remainder in the evening or online?

 

[ironmonkeyz]

Hey Golllini, thanks so much for your reply. I have not yet started taking these courses yet, working at a prop firm is extremely time consuming (12 hours a day in the office). Im thinking of taking these math courses mid 2014 at University of Toronto full time so I can get most of them finished in time for the Dec 2014 application deadline. I have already done a online c++ course (the one offered at berkeley).

Do you think math courses from a Canadian school will be recognized? Thanks

 

[GoIllini]

Of course. I think Princeton has actually taken a few U of T students over the years. We have like 4 Canadian second year students in the program.

 

[systemtrader]

Hey Gollllini how about doing a pre-MFE program at one of the Universities that offers them. Would completing a set of pre-programs be enough to replace taking individual math courses? I'm currently self-studying for the math requirements because it would take too long to take all the required math courses one by one. Do you think you NEED to have college credits for all those courses you mentioned or doing good in pre-program courses be enough?

 

[GoIllini]

Hmmmm.

I'd give UChicago a call. See what they think.

We are talking about 18-20 credit hours of math courses- equivalent to greater than the maximum a FT student can take in a single semester without permission from a Dean. The summer program will be much shorter.

I think UChicago's summer program provides good signaling value. (There may be other summer programs that do as well.)

I am not sure you'll be ready for UChicago's summer program unless you start going over stuff on khanacademy in your spare time.

 

[systemtrader]

Hmmmm.

I'd give UChicago a call. See what they think.

We are talking about 18-20 credit hours of math courses- equivalent to greater than the maximum a FT student can take in a single semester without permission from a Dean. The summer program will be much shorter.

I think UChicago's summer program provides good signaling value. (There may be other summer programs that do as well.)

I am not sure you'll be ready for UChicago's summer program unless you start going over stuff on khanacademy in your spare time.

oh i see. thanks~

 

[Daniel Duffy]

Linear Algebra
-> You should be able to remember a few details of the invertible matrix theorem, remember whether you can multiply a 1xn matrix with an nx1, and maybe have a hazy memory of eigenvalues and eigenvectors.

Wow!

1Xn and nX1 matrices are called vectors.. and the operation is called inner/dot/scalar product.

Multiply nXm matrix by a pXq matrix.

 

[GoIllini]

Linear Algebra
-> You should be able to remember a few details of the invertible matrix theorem, remember whether you can multiply a 1xn matrix with an nx1, and maybe have a hazy memory of eigenvalues and eigenvectors.

Wow!

1Xn and nX1 matrices are called vectors.. and the operation is called inner/dot/scalar product.

Multiply nXm matrix by a pXq matrix.

Uhh, you can have a 1xn matrix or an nx1 matrix. The matrix multiplication happens the same way as for two vectors having a dot product taken, subject to the orientation of the matrix, of course.

Fighting about mathematical semantics is about as foolish as arguing about grammar. If you want to argue about how to represent stuff in linear algebra, you should seek an MS Mathematics degree, or maybe a JD.

 

[Daniel Duffy]

Uhh, you can have a 1xn matrix or an nx1 matrix. The matrix multiplication happens the same way as for two vectors having a dot product taken, subject to the orientation of the matrix, of course.

Fighting about mathematical semantics is about as foolish as arguing about grammar. If you want to argue about how to represent stuff in linear algebra, you should seek an MS Mathematics degree, or maybe a JD.

The nice thing about mathematics (and grammar for that matter) is that something is either right or wrong.

Saying that a column or row vector is a a matrix is not even wrong. But no mathematician thinks like that.


For the record, mathematics is not something to 'remember' ; you either know it or you don't.

BTW what is 'JD'?

 

[GoIllini]

...only though most quants are engineers or physics folk with a smattering of CS backgrounds. The fact that people play such a fast and loose game with mathematical notation and then defend the results as "but I got the right price" must really irk you.

 

[Daniel Duffy]

...only though most quants are engineers or physics folk with a smattering of CS backgrounds. The fact that people play such a fast and loose game with mathematical notation and then defend the results as "but I got the right price" must really irk you.

Why should it irk me? I don't see why. Good for you if you can get the right price without maths.

Nothing wrong with engineering. Things become interesting when you have to come up with a new product. Then more is needed.

It's always nice to know _why_ you got the right price and it was not just a remembered formula.

 

[bigbadwolf]

BTW what is 'JD'?

Juris Doctor (law degree).

 

[Daniel Duffy]

Juris Doctor (law degree).

In a sense, Gollini is correct; Law is very precise, like maths.

 

[GoIllini]

It's always nice to know _why_ you got the right price and it was not just a remembered formula.

Or you just do a ghetto montecarlo and get the same answer you'd get analytically.

 

[Daniel Duffy]

Or you just do a ghetto montecarlo and get the same answer you'd get analytically.

If you are not too worried about accuracy and speed, it's fine.

 

[GoIllini]

Well hopefully you'll already have it in an in-memory cache before-hand so you don't need to solve analytically. Getting the standard normal cdf, for instance, isn't a cheap call if you're doing it on the fly.