Live Algo Trading
That was a great overview and exactly what I needed to find out. I saw very quickly that R was powerful and could perform my tasks, but I also know how interpretive and scripted platforms perform when in production. I will stick to AQ and keep control of the internals implementation. I am just realizing how much work it will be!...
Jeff Ryan wrote:
Using R as a live trading platform is really on a case by case basis. For 1 minute bar trading, with limited computational overhead, R is a perfectly workable solution. Inside of that time-frame, things may get difficult. A multitude of factors should be considered. First and foremost would be R proficiency. quantmod isn't up to the plug and play task of automated trading (yet?!). So you'd need to do quite a bit of work to get things up to speed. It is doable, but you'll either be writing a lot of code yourself to fill in some infrastructure bits, hiring someone to do it for you, or most likely a combination of the two. And it will take time. All worthwhile things do. The IBrokers package is a decent example of what you can do. Run across multiple sessions on a multicore platform, you can manage to process as much data as IB will let you. You'll not be able to take a raw feed from a real data source of course (think millions of messages a second). Preprocessed (limited symbols, aggregation, many cores, capture-engine intermediary, etc), and you are getting closer to 'real-time' reality. The advantage to myself and those I know who *do use* the above approach, is that you can think in "R". If you are doing backtesting, post-analysis, etc in R, it is a quasi-natural fit to move the execution into R. Another approach taken is to keep the execution stuff outside of R (C++ or Java for example), and simply make calls to R when needed. Of course if your logic relies on the R code, you are still imposing the same potential limit on the total process. Simple take away: it is possible, but not easy. You gain an incredible amount of flexibility and speed of development (if R-versed already), but the trade-off is in raw processing capacity. A strategy relying on ticks or orderbook data with only R would likely be suicide. 15s to end-of-day style, very doable. Best, Jeff On Wed, Dec 23, 2009 at 10:55 PM, tradetree <stock at ancientsaturn.com> wrote:
I wanted to see if there are people using R and quantmod to trade live, or only as an off-line tool for algorithm development? ?I am evaluating ActiveQuant and R to decide what to use for a live trading platform. ?I am coming from Tradestation and right now considering AQ and the IB interface. It is very hard to find an overview of approaches and pros/cons for automated trading. ?Some people use AQ for the trading engine but do algo development in R and quantmod. ?Is this a recommended direction? -- View this message in context: http://n4.nabble.com/Live-Algo-Trading-tp978300p978300.html Sent from the Rmetrics mailing list archive at Nabble.com.
_______________________________________________ R-SIG-Finance at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-finance -- Subscriber-posting only. -- If you want to post, subscribe first.
-- Jeffrey Ryan jeffrey.ryan at insightalgo.com ia: insight algorithmics www.insightalgo.com
_______________________________________________ R-SIG-Finance at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-finance -- Subscriber-posting only. -- If you want to post, subscribe first.
View this message in context: http://n4.nabble.com/Live-Algo-Trading-tp978300p978313.html Sent from the Rmetrics mailing list archive at Nabble.com.