NinjaTrader Problems

Discussion in 'Automated Trading' started by ScoobyStoo, May 20, 2009.

  1. Hi everyone,

    I'm currently evaluating the NinjaTrader platform with a view to using it for running fully automated strategies.

    I have to say that after spending a week running loads of data through it and assessing the architecture, I am really disappointed.

    The longest I have managed to keep the thing running is 30 hours. The memory management is so poor that it crashes with a fatal .NET runtime error after this time.

    The architecture is amateur to say the least. It seems to have been written by people who were learning C# whilst they were building it.

    The support staff don't seem to understand the software properly...I am currently engaged in a discussion with one of them about the ridiculous sequencing of events when an indicator object is constructed.

    Anyway, I could go on, but in short I have pretty much made my mind up that I simply can't trust my money to such an amateur piece of software.

    Therefore, I'm asking any serious .NET developers out there to suggest a better alternative.

    I've spent a little bit of time looking at OpenQuant but the documentation is fairly poor and the support looks to be practically non-existant over on their forums.

    I'd ideally like to use the full QuantHouse suite of products but I don't have $100k to throw at the software.

    Are there any other decent .NET platforms out there? I really don't want to build the boilerplate connectivity, order management, execution processing etc stuff myself.
     
  2. Eight

    Eight

    Openquant might be a better choice. A guy was posting on here while he wrote tickzoom, I don't know if anybody uses it yet...

    Ninja works ok for me, I trade manually and shut the whole machine down eod.. and I never can get any indicators that involve time differencing to work right, but I'm not that much of a programmer..
     
  3. Well, we all may hope for ninja 7 ;) Beta out end of next month, iirc - which is pretty soonish ;)

    I am meanwhile working on my own framework, but it is FAR from usable (actually it is just starting), but if you want ;) I am always interested in not working alone.
     
  4. Many thanks to those who have commented on this thread and the many people who have PMed me.

    Seems like, having come to the same conclusion as me, several of you are building your own platforms from scratch. Thanks for your invitations to work on these platforms with you, but unfortunately I just don't have the time...hence my desire to buy a professional platform which handles all the boilerplate stuff, leaving me free to just develop my strategies.

    I shall have a deeper look at OpenQuant.

    Looks like we've definitely identified a big gap in the market though...for a professionally designed, robust, scalable (and most importantly affordable) trading platform which will allow us to compete on a truly level playing field with the hedgies. To that end, I wish all of you 'rolling your own' the very best of luck.
     
  5. Actually, hedgies rely on size, not so much platform. Have you tried Metatrader?
     
  6. I wasn't refering to the size of their positions...but the software platforms the quant hedgies use.

    I've only been exposed to a small number, but those that I have had dealings with were all dependant on fast and reliable platforms. All of them had been custom built by the hedgie themself at great expensive. These guys simply wouldn't trust any code they didn't have complete control over.

    I don't want to spend anywhere near the amount of time and money they did.

    I've had a look at MetaTrader4 previously and it's a pretty good piece of software for what it's designed to do. However, MQL4 is fairly limited in its abilities and is scripted (as far as I remember) so you don't get the performance benefits of compiled code.

    I need the power of a full OO language rather than some Mickey Mouse proprietary scripting language, which for me means either Java, C# or C++. Since in the real world I'm a .NET developer C# is my preference.
     
  7. RedDuke

    RedDuke

    I have spoke with few Ninja guys and they confirmed that 7.0 will have a vast improvement for automation via 6.5 among other things.
     
  8. Have you considered Zen-Fire API?

    rt
     
  9. Zen-Fire is cracking. Truly brilliant. By far the best retail data feed for the instruments it handles.

    I think I'm going to spend some time seriously looking into OpenQuant again whilst I wait for NT7. The NT guys say the beta will be out before end of quarter. That should give me a pretty good idea as to whether they have rebuilt it using a decent extensible architecture.

    That's not to say don't keep the comments coming. I'm always eager to learn from the experiences of others...
     
  10. bob5e

    bob5e Guest

    Yes, Zen-Fire rocks but I'm getting tired of waiting for NT 7.0 . . too many times the Beta date has been delayed and reset. I've gotten tired of waiting for it. Frustrating . .
     
    #10     May 21, 2009