I thinking about to make my backtesting in Matlab. I need to decide if to code it myself with "plain" Matlab matrixes or to use time-series toobox? Time-series toolbox already have code for TA indicators, so it will be less coding, but as you already said it's non-flexible, since you can have only 1-minute bars (this is less important for me, since my data is in 1min bars). Any pros and cons? Thanks in advance.
Nitro, Presently, the Matlab Datafeed Toolbox connects Matlab to 4 different real-time intraday data vendors -- Bloomberg, Factset, Hyperfeed, and FTInteractive Data. These are all high-priced vendors, $2000 and up per month. Please ask them if they intend to make the Datafeed Toolbox compatible with a lower-priced vendor, such as eSignal or Qcharts. Nobody I've spoken to at Matlab seems to know. Also, automated trading -- do they intend to offer anything along these lines. Something that will connect Matlab directly to a broker, similar to the Gloriosia product that connects to Interactive Brokers, but perhaps with more broker alternatives. Thanks!
HLB I am a long way from having a complete solution myself. I have thrown together a couple functions which allow me to import 15 second data, but I have not done much as far as backtesting, etc. Others are probably in a better position to advise you. My current plan is to avoid the "financial time-series objects" due to their limitations, but I might change my mind.
Experiment with both. Try to code it yourself, and fall back on the built in functions if you need to. Consider using http://www.gloriosia.com/ for the datafeed API, and for example indicator code.
Maybe we can persuade Jerry Medved to write a sockets-based interface to export real time QuoteTracker data into Matlab? That would be extremely useful.
FYI - We now have a Socket based API for tick by tick data from QT: http://www.quotetracker.com/help/QTStreamingAPI.shtml I am not familiar with Matlab, nor do I have the software. Since our API is public and fully documented and quite simple, it would be far easier for Matlab to put together the needed connector, thus in affect adding support for all the datafeeds QuoteTracker supports. The Streaming API works for registered users only, but we would be more than willing to take care of that if someone decided to write the needed connector
Incredible! I have always been impressed with QT's richness of features. I developed my original system ideas while trading with QT doing the charting. As much as I would like to provide everyone with a Matlab connector, I don't have the time currently to write a proper, efficient implementation. I do all my real time system analysis using compiled Matlab code and custom sockets code written in C. None of this is useful for getting real time data into the Matlab interpretive environment. Maybe The Mathworks or the Trading with Matlab people would be interested in doing this?
Anybody here from Los Angeles using Matlab to develop and research trading systems? If you are coding in Matlab drop me a note. Interested in exhanging thoughts on implementation. Thx.