Right now I use Excel/VBA to do all my backtesting. I'm looking to expand one of my systems (100+ trades per day) and I'm thinking either R or MATLAB would be a better tool for the job (mostly research/backtesting, possibly execution down the road). Does anyone have experience with BOTH of these options, and a preference for either? I realize R is free and MATLAB costs 2k+, which isn't really a concern. I'm just looking to get started with a solution that I'll stick with.

R About only advantage to matlab is some signal processing support-- but, R has huge support and development behind it as well as financial libraries available.

If you want a IDE environment like Matlab and are willing to spend money, I can also advise you to look into Mathematica. I find its symbolic math, programming and graphics (plotting) possibilities way better than Matlab. Matlab is however known to be a little quicker on numbers, you can google for benchmarks. As another interpreted language you can also try Python. It might be the easiest, yet very powerful language. There are excellent libraries available. There's no real difference in possibilities anyways. It all depends on your preferences and habits. I'm biased towards both Python and Mathematica because I know they integrate quite easy with C++/C#. But this might also be true for R and Matlab.

I found some insight on this topic on nuclearphynance, although some of the R critiques were many years old. Curious if things are different here in 2012. I guess the main factors I'm interested in are: - computational speed - development time - integration with Excel/VBA

R has XTS, Quantmod, Quantstrat, Portfolio Analytics, numerous varieties of contegration, etc. Nothing else matches these imho.

do you trade through r? who's your broker (ib)? i run on hundreds of stocks live via excel (people who harp on speed simply don't understand their own speed or how to optimize excel). i'm debating moving my strategies to python or r to add functionality and prototyping speed --- can't decide, and it's a big investment so want to get it right. i guess i could always use rpy and do both. suggestions?

Unless you want to use MATLAB's compiler, distributed computing, or parallel computing toolboxes, I'd consider using Octave (free) over MATLAB.

I do not trade through R directly and do not use IB. If you do, there is the R IBrokers package. http://cran.r-project.org/web/packages/IBrokers/IBrokers.pdf However I myself think R is best for analysis and I would actually execute using other software.