Do you think that you can master those areas in just 1 or 1.5 years? Given that amount of time, I strongly believe that a typically good student can only grab those in a recipes style.
I think that is reasonable analysis. To me a "master" is someone in (say) the top 10% of a given profession, a MFE does not do this.
I believe that Dominic is biased and particularly obsessed with C++ programming in quant jobs.
I plead guilty on both counts.
I listen to what hiring managers tell me, and study why people fail to get jobs, and
C++ is up there big time, hence my bias. If they said PDEs or game theory I'd be pushing them. As a student of my posts you will notice also a bias against those who use Monte Carlo rather than understand the problem.
If you think I bang on about
C++ a lot, then I feel I have done my job properly.
However googling on my various IDs with "Excel" will see me tell many people to get that up to speed as well.
A valid criticism of the aggregate nature of my posts is that I complain more than I praise. I talk of things that people don't have which hurt them. I don't ever write "I'm just so happy everyone knows >5 ways of deriving Black Scholes"
It probably takes a computer geek 10-20 years of programming to master the language.
Using the definition I give above you may be right
But at entry level I'm looking for competence, not excellence in
C++. You can do that in three months flat. OK, a
tough 3 months, and you'd need someone to hold your hand.
In my opinion, C++ is important but not as fundamental as Stochastic Calculus, Modeling, Numerical Computations, (and maybe some sense of Finance/Econometrics).
Well, yes I agree, but we are not dealing with quantities amenable to linear addition here, nor is ordering a valid way of deciding how to allocate study time.
A simplistic but useful first model would be to
multiply your rank in the core skills like Stoch, LA, PDE, time series, numerical anaylsis, brainteasers and
C++
Thus being in the Zero percentile at
C++ may often be fatal to your application even if you are wonderful at everything else. Of course
C++ without numerical analysis is a bit futile, and unless you have stochastics and finance theory you are just a programmer.
I entirely agree and commend your position of using opportunity cost to help decide what to learn, where I think we part company is how we evaluate the utility of the portfolio, and the marginal return on an investment of your time.
I guess one implicit assumption in your position is that below a certain critical there is no point learning
C++ because you won't be able to do anything useful. I agree with that
.
My view is that for many students moving from the 98th percentile in stochastics to the 99th adds little to their value, but the same effort can get you from 0 to 20% in
C++.
Ideally, C++ (or programming skills in general) should be required as a preliminary, not as a core part of the FE training.
Agreed, but it ain't gonna happen.
On the CQF, we do
C++ as a post hoc lecture series.
