I trade derivative relationships. I do not do directional systematic directional outright backtest optimize kind of things.. If someone defends their liking of a programming language, what stake do you have in getting involved in an argument or discussion about it.. I'm just being helpful... If there is a better language more suited for an application then i would say have at it.. Your expression of " you prefer +__ how so?" seems like a advance for a debate... I'm not interested sorry.. maybe you can find some R defenders interested.. Good luck
Large sorting in R
- Thread starter R1234
- Start date
I trade derivative relationships. I do not do directional systematic directional outright backtest optimize kind of things.. If someone defends their liking of a programming language, what stake do you have in getting involved in an argument or discussion about it.. I'm just being helpful... If there is a better language more suited for an application then i would say have at it.. Your expression of " you prefer +__ how so?" seems like a advance for a debate... I'm not interested sorry.. maybe you can find some R defenders interested.. Good luck
Hmm , just as I described. As soon as someone dares to question R being the best tool for everything God given out come the poster boys and sharp shoot. R by the way is not a programing language and for anything but vectorized computations is crawling slow. You are the one who suggested an R package for strategy development and/or backtesting. I simply expreased my curiosity why you suggested R for such task when several (apparently quite informed) users just laid out in detail why R is possibly the worst choice of them all for a strategy backtest harness. Don't take yourself too serious please, it was just an honest question.
Last edited:
" Good.. I thought so" what did you watch some gangster movie just now or something... like maybe even Bruce lee.. Practicing that round house mang... I'm not exactly submissive either so you might look somewhere else for your type of bottom boi
I just expanded and can only again suggest for you to relax and ease up, nobody is trying to attack you. I simply did not understand why you recommend R to be a suitable test harness when others walked in detail through valid reasons why R is potentially the worst choice. That's all. Peace out mate.
i'm messing around.. I'm chill no worries.. I know some of the guys that are the main contributors of the main packages for systematic trading use portfolio analytics, quantstrat, TTR and a few other packages.. most of which they have contributed themselves. I do believe your probably right about other languages being more efficent in getting large amounts of processing done more quickly but again i don't evne know i'm not dealing in that area..
All of the links you provided use marketing gibberish to promote their own products. Anyone can easily spin up multiple R sessions and distribute workloads in parallel. I have done it multiple times. It all however does not change the fact that any algorithm that cannot be vectorized should never be run with R scripts because it is way too slow. Op's problem clearly is not a vectorized computation and hence spinning up several tasks in even C# will blow away any comparable R script. Of course you can reply that users could peruse RCPP and call C++ routines to which I would reply why not dropping then R and do it in C++ directly. R has its strengths no question. But certainly it is not non vectorizable operations.
