Like the other professional software developers on this forum I too can smell the amateurish architecture oozing out of NT. From the UI through to the API it just whiffs of a back bedroom project which has grown organically over the years without any real design principles being applied to it.
I have many opinions about what they need to do to improve it, but the only one that really matters is they need stop cramming in new features (it already suffers from cronic feeping creaturitus) and to instead spend the time that frees up implementing professional software development practices. As someone said earlier...
1. Employ test driven development practices (Unit test software is free for God's sake, there is just no excuse for not using it these days)
2. Employ continuous integration (Something like CruiseControl.NET is also free)
No capital outlay required, just time commitment.
But I think the real elephant in the room is that NT's underlying architecture is probably so poor that it needs a complete rewrite. Those of us on this thread that work in the industry can smell it and I suspect both Ray and Dierk know it too. The trouble is, they are such a small operation they couldn't hope to undertake such a big job, the code base has taken them years to build...so they are stuck trying to patch up a piece of software that really and truly needs to be carted out back and unceremoniously tapped between the eyes.
The following 4 users say Thank You to ScoobyStoo for this post:
Yeah, well, I am currently wrestling with NT v18.104.22.168 trying to get the market replay functionality working so I can run a backtest using full transaction and order book data. I needed somewhere to vent my not inconsiderable frustrations.
Someone on this thread also reminded me of the fact that NT only timestamp their data to 1 second granularity. This is a joke because the data feeds themselves timestamp this stuff to millisecond (and in some cases microsecond) precision. Means that when you replay data through NT it first spits out all the L1 data for a particular 1 second period and then all the L2 data for that same 1 second period. In fast markets this means that the L1 and L2 order book data can be out of synch by several ticks. How the hell can their simulated executions be even half way accurate if the L2 order book is lagging by several ticks? There's just no way they can accurately estimate fills.
With regards to your question about alternatives...I spent several months looking around last summer and in the end decided to build my own algo execution platform from scratch. Finished my last project management contract in July 2009 and have been building it ever since. I now have a suite of high performance .NET algos interfacing directly with the Rithmic API. No charts, no UI in fact, nothing unnecessary in there to slow it down. It's been a fairly steep learning curve but I have learnt an incredible amount about how the exchnages operate at a low level. That in itself is now invaluable. Will be running the strat on a box coloed in my FCM's data centre located only a few hundred metres from the Merc and using a direct pipe so should get really low latency.
But back to the topic at hand, yes there are very few if any serious alternatives to NT for those of us in the retail crowd. They all have their faults. I'm lucky in that I have worked in The City myself and still have mates in the banks and hedge funds so I get some exposure to what the institutions are using, but somebody like Sungard aren't going to be interested in dealing with retail clients.
So yeah, for the money NT probably delivers the best bang for the buck. I use the free version for testing and begrudgingly accept its limitations, but wouldn't trust a single penny of my hard earned dosh to it in a production environment.
EDIT: This probably all sounds extremely harsh. I probably shouldn't post on forums whilst I am tearing my hair out.
Last edited by ScoobyStoo; March 25th, 2010 at 11:59 AM.
The following 3 users say Thank You to ScoobyStoo for this post:
I have used NT/Zenfire for several years now and I currently run NT7 for my discretionary trading oder entry/management needs. I have had pretty solid reliability over the years but the past month I have had several small issues here and there. For myself personally, I have decide to start slowly moving accounts and trading operations over to the TradeVec platform through Vcap Futures. I know the guys over at Vcap so I am actually looking forward to the transition away from NT7. It seems to me that NT7 is not going to be the great leap forward that I thought it would be, and there are some features that I just can't wait for anymore.
Ninjatrader has served me well over the years, but in the months ahead I will most likely fully transition away to a new platform/charting arrangement.
I've not seen anything. If it's freezing inside NinjaTrader it's probably due to an indicator on your charts consuming more cycles than your CPU can spare. Do you have something up to monitor your CPU usage? Like with Vista or Seven there is a built-in gadget/sidebar to do it or there are millions of options. You should check to see if there is correlation between your CPU usage and the freeze.
Due to time constraints, please do not PM me if your question can be resolved or answered on the forum.
Need help? 1) Stop changing things. No new indicators, charts, or methods. Be consistent with what is in front of you first. 2) Start a journal and post to it daily with the trades you made to show your strengths and weaknesses. 3) Set goals for yourself to reach daily. Make them about how you trade, not how much money you make. 4) Accept responsibility for your actions. Stop looking elsewhere to explain away poor performance. 5) Where to start as a trader? Watch this webinar and read this thread for hundreds of questions and answers. 6) Help using the forum? Watch this video to learn general tips on using the site.
If you want to support our community, become an Elite Member.
At the same time I think you have to appreciate all the stuff NT has to put up with compared to institutional software. Its basically the same solution to the same problem but with retail software you simply have much less funding and a lot more things to worry about. Institutional software doesn't have to bother with fixing bugs in drawing gann fans or other such nonsense.
NT is good for what it is right now, but considering the scope of the ideal it will probably take until version 10 until everyone is happy.
I think the problem with multicharts at this point is how far off are we from a major tradestation update. The company has 800 million in cash and has had a long time to watch all the "tradestation killers" come along since the last major update and learn from their mistakes.
Ninja does have good bang per buck but only because of the free version bang/0$ = infinity right?
There are some credible alternatives Neoticker being the obvious one if you want tick precise handling of multiple data stream (and a whole lot more). Great architecture but fairly steep learning curve.
InvestorRT seems like an obvious choice if you want to simply access to 'MarketDelta' type studies. (seems like many do).
Multicharts is not bad either. You can get stuff up and running phenomenally quickly in Easy Language plus there is a wealth of public domain indicators. Suffers from similar problems with precision of historical tick data and it is possible to get race conditions in studies processing multiple data streams.
My perfect application Neoticker's architecture, Ensigns looks (can make really pretty charts), maybe Ninjas chart trader with easy language scripting.
I have the skills necessary to 'roll my own' but really don't have the inclination. Having said that I would probably want data visualisation over algorithmic execution. For algorithmic execution roll your own, colo etc. is almost mandatory. What's the point of having the computer execute if you have to sit and watch it?
If one wants to run algo's automatically (and it is not a high frequency/low latency application) Strategy Runner might be an option. Design prototype and test in whatever you like then run the strategy on a brokers Strategy Runner server.