Do BBs recruit MS in Math?

[Four]

I'm thinking of switching from a MS in computer science degree to a MS in math, where the coursework is easier and more interesting (for me).

Do BBs recruit graduates with a MS in Math degree? My school is a target school.

 

[systemtrader]

Did you major in Math as an undergrad? I would like to know how to take a MS in Applied Math if you did not major in Math as an undergrad. Do schools provide pre-program courses or do you have to prepare by yourself.

 

[Andy Nguyen]

MS in CS will open more door for you career wise. You most likely need some solid programming skills these days anyway.

 

[muttiah]

MS in CS will open more door for you career wise. You most likely need some solid programming skills these days anyway.

I don't think this is true.

In my experience the general perception is :
(Pure) Math degree => This guy must be really smart.
Engineer degree => This guy knows how to solve problems.
CS degree => Not as smart/good as the math or engineer guy.

In general from the same university, the math and engineer degrees are stronger degrees (generally require more work and harder to do well in). Programming skills can be more easily acquired through self study (beyond basic intro to CS course and algorithms 1 course). For developer jobs the CS degree helps some.

Also to OP,
where did you get the idea that a MS in math is easier than MS in computer science??

 

[IntoDarkness]

I don't think this is true.

In my experience the general perception is :
(Pure) Math degree => This guy must be really smart.
Engineer degree => This guy knows how to solve problems.
CS degree => Not as smart/good as the math or engineer guy.

In general from the same university, the math and engineer degrees are stronger degrees (generally require more work and harder to do well in). Programming skills can be more easily acquired through self study (beyond basic intro to CS course and algorithms 1 course). For developer jobs the CS degree helps some.

Also to OP,
where did you get the idea that a MS in math is easier than MS in computer science??

I don't think what you said is true...

 

[Ian Kaplan]

I don't think this is true.

In my experience the general perception is :
(Pure) Math degree => This guy must be really smart.
Engineer degree => This guy knows how to solve problems.
CS degree => Not as smart/good as the math or engineer guy.

In general from the same university, the math and engineer degrees are stronger degrees (generally require more work and harder to do well in). Programming skills can be more easily acquired through self study (beyond basic intro to CS course and algorithms 1 course). For developer jobs the CS degree helps some.

Also to OP,
where did you get the idea that a MS in math is easier than MS in computer science??

There are not a lot of option pricing jobs out there these days. Unless you work in options pricing, where raw brilliance and math skills are critical, programming skills cannot be emphasized enough.

Broadly, I think that its safe to say that jobs in finance can be broken down into three areas:
- Mathematical modeling
- Empirically based modeling (which uses data and statistical analysis)
- Infrastructure

Only the first of these do not involve heavy programming skills. And these skills are not something that you can just "pick up". They take years to develop. People who think they can just "pick up" programming do not respect software engineering and write terrible software that is unmaintainable and unreliable.

I'll give a concrete example: I am working on portfolio models. I am using the Compustat data from the Warton Research Data Service. So far all of the data sources on WRDS are "dirty". That is, they are missing critical data values, like shares outstanding or even close prices. These values must be filled in before you can use the data. Then once the data is in shape to be used, you must evaluate how to build a model that will make money. All of this takes programming and statistics. People tend to greatly underestimate the amount of programming required to prepare data.

There's something really important to remember. There is one and only one thing that matters in finance: can you make money. You can have all the theory in the world and your hard core math skills, but if you can't deploy them to make money and build reliable software that does this, then you have little future in finance. While there are a few genius math modelers who can depend on someone else to build their software models, most "quants" will be required to build their own models.

There is also one other thing to think about: I would say that there's a pretty strong consensus right now that the job market is far better for software engineers than it is for "financial engineers". Also the salaries for FEs are not much better than the salaries for software engineers, especially when you adjust for geography. The world has changed and the days of high salaries and bonuses are gone, at least for now.

 

[muttiah]

Fwiw I'm a software engineer. I studied computer engineering and math in school. I acquired programming skills primarily through self study (years of study). My points are
1) Its much harder to learn programming through self study, compared to abstract math
2) Many people with non CS degrees do programming for a significant percentage of their work hours (quants, engineers, data scientists etc)

In terms of developer job prospects there isn't a huge difference in degree (math, CS, applied math, stats, computer engineer etc). What's most important is actually being able to solve problems and implement solutions in code.

There are not a lot of option pricing jobs out there these days. Unless you work in options pricing, where raw brilliance and math skills are critical, programming skills cannot be emphasized enough.

Broadly, I think that its safe to say that jobs in finance can be broken down into three areas:
- Mathematical modeling
- Empirically based modeling (which uses data and statistical analysis)
- Infrastructure

Only the first of these do not involve heavy programming skills. And these skills are not something that you can just "pick up". They take years to develop. People who think they can just "pick up" programming do not respect software engineering and write terrible software that is unmaintainable and unreliable.

I'll give a concrete example: I am working on portfolio models. I am using the Compustat data from the Warton Research Data Service. So far all of the data sources on WRDS are "dirty". That is, they are missing critical data values, like shares outstanding or even close prices. These values must be filled in before you can use the data. Then once the data is in shape to be used, you must evaluate how to build a model that will make money. All of this takes programming and statistics. People tend to greatly underestimate the amount of programming required to prepare data.

There's something really important to remember. There is one and only one thing that matters in finance: can you make money. You can have all the theory in the world and your hard core math skills, but if you can't deploy them to make money and build reliable software that does this, then you have little future in finance. While there are a few genius math modelers who can depend on someone else to build their software models, most "quants" will be required to build their own models.

There is also one other thing to think about: I would say that there's a pretty strong consensus right now that the job market is far better for software engineers than it is for "financial engineers". Also the salaries for FEs are not much better than the salaries for software engineers, especially when you adjust for geography. The world has changed and the days of high salaries and bonuses are gone, at least for now.

 

[Four]

Broadly, I think that its safe to say that jobs in finance can be broken down into three areas:
- Mathematical modeling
- Empirically based modeling (which uses data and statistical analysis)
- Infrastructure

Are there a lot of mathematical modeling jobs available? I can code and get good at it if need be, but would like to spend more time on math and less time on programming

 

[Ian Kaplan]

There are, of course, jobs doing many things in finance. But if you're talking about a job in quantitative finance, I don't think that there are jobs where there is not a lot of software development in R, Matlab, C++ or Java. In my view, this is a bad field to work in if you don't like to code. For one thing, you'll never be any good at something you don't like to do.

There are probably a few people who design options models and who are geniuses who don't program. But these people generally also have finance experience. And they are really, really smart. It's worth noting that there are no known analytical solutions to many types of options, including American options. They are priced by simulation (lattice or Monte Carlo). There is an analytical solution for some Asian options, but the solution is complex enough that people still use models for pricing.

This is not theoretical physics. As I noted, its about making money. Areas like market impact all have mathematical models, but in the end it comes down to application.

After Knight Capital lost $400 million in forty minutes there may be increased focus on software engineering, software quality and software risk management. This suggests that it may be increasingly difficult to work in finance unless you take software design and engineering seriously.

And I think that I mentioned that the job market and the salaries suck right now.

 

[Alexandre Chretien]

Are there a lot of mathematical modeling jobs available? I can code and get good at it if need be, but would like to spend more time on math and less time on programming

It takes time to become good at coding. I try to keep a balance between these two things.

 

[Four]

I don't think this is true.
Also to OP,
where did you get the idea that a MS in math is easier than MS in computer science??

I probably should have worded my original post differently. I know the MS Math is not easier but it will likely be more enjoyable. I don't like programming, don't like reading code, and like visualizing things (geometry, Penrose tilings) and understanding and solving puzzles (Tower of Hanoi, Birthday and Barbershop Paradox, etc).

 

[Phil]

I so love these kinds of discussions.

My take is that the math required to reach the top 85% of quant traders is harder than the computer science required to at least get into quant trading.

To really beat the market, you need a combination of modelling and intuition rigorously framed in mathematics. Those without proper training in abstract math (analysis) just don't have the necessary tools to proceed. They'll eventually get stuck to a point where the only way forward is this higher plane of thinking.

Those new to computer science can follow a standard path to a stage of building trading systems - data structures, algorithms and software design. Understanding usually comes after logical thinking. But those without math training will find it harder to accept abstract concepts in analysis as I feel you need logical thinking and time to "get used to it".

I always remembered this saying (I might have read it in Quantnet): "You may know the mathematics but can you generate the mathematics." The top 85% of quant traders are the ones who can generate the mathematics, an ability that takes years to develop.

 

[TraderJoe]

You will get a lot of free time to study Math indepth if you get a job as a waiter. There will be no need to ever do any coding.

 

[Daniel Duffy]

You will get a lot of free time to study Math indepth if you get a job as a waiter. There will be no need to ever do any coding.

With washing up plates you learn Stack (LIFO) ADT?

And (priority) queue with Poisson arrival and exponential service..