most would agree that amibroker is the bettter platform.. save your time, and pay the fee.. you can use ninja.. but i don't recommend it..
j2ee, - what's your discretionary trading experience? - what's your automated systems design experience? Your first (and largest) hurdle when starting to develop automated systems isn't going to be live execution, but simply getting to your first tradeable strategy. This is really the tough nut to crack, and until you have something you are absolutely convinced you can trade live (meaning, a few thousand trades in backtest minimum, plus 3-12 months of forward testing), you won't make the difference between any 2 platforms. I would recommend NinjaTrader, because its C# and free until you want to trade live with it, or MultiChart Starter C# (also free, but hot off the press, might not be quite stable yet), TradeLink might also be a candidate BUT Ninja support team is really a great plus. You absolutely need quality historical data for your research and development, visit TickData, and get 10+ years for your markets of interest. Ninja has a number of bugs and limitations (all products in the same boat!), but to date I have always been able to do what I want with it (although, some workarounds have taken me a lot of work), and I am now running strategies always in the market, 24/7, 100% automated (including rollovers).
your right about all this.. i really have nothing against ninja really.. but the guy has not given us crap to go on as far as what he is trying to do .
NinjaTrader is good because it is free until I really want to trade? This is really funny comment! That is exactly how NinjaTrader gets some people get used to use it then pay for them!
For C#, I can just use tradelink. Base on this link info: http://www.ninjatrader.com/about I need to at least paying $50 per month for ninjatrader. There is active community forum for tradelink and I cannot imagine ninjatrader worth $50 more per month than that.
Sorry I really don't see the humor in that. Given that there is 90% chances you won't get to a profitable strategy that can be traded live, this is saving you a mere $995 for Ninja, and you still get the benefit of their customer support. And if you hit that 10% where you actually can go live with your strategy, you might find that $995 a fair price.
That level is pretty low on the floor, though. The reason you do not see any competitor close is likely that most more professional setups have their own infrastructure that is non public. For example one hugh problem with backtests in Ninja /and tradelink is quality (ninja) and scalability (both). Quality in ninja because hey, when a strategy gives 3 different results in 3 optimizations over the same timeframe you start to wonder where the random factor comes from (not genetic, sorry). happens it is data downloads on different machines that - fail without telling you. Then you get into execution quality. Sucks as ninja is not able to simulate the event feed of the exchange - so bid/ask are not available and not used in backtests. THat also makes certain thigns not usable in backtsts (any indicator using for example the order book). Then you get bigger. Make a backtest over 5 years of data. Have fun on your computer - NInja happily does it on a 100gb machine - it keeps all in memory. But memory is not your main problem, the thing takes ages. To make it better you must run it on a grid - which neither Ninja nor Tradelink have out of the box. Yes, with TradeLink you can do it - but then a lot has to be done. We got our own (c#) based infrastructure with some partner traders (no, not able to share it) and for example we do backtests and optimizations in one week steps (sunday-saturday) as we are always closed on the weekend. A 5 year job gets split into 1 week jobs - and gets distributed over a grid of computers. Reesults go into a database, no "in memory and if something happens it disappears". You can go back to a (marked as keep it) simulation in a year and compare it. I assume many setups have something similar. But nothing gets productized. The costs would be high, on both sides. To make it a product it needs a LOT more testing. And then someone has to pay it - no 995 USD deal here, we have 1 person working on that half time, and another full time position unfilled. Purely for infrastructure.
i would bet the guy is not in the position to code up an entire trading/backtesting/optimzation platform... but again.. we don't know.. he doesn't tell us shit......
Thanks for your detailed answer. Why don't you guys use C++ instead of C# when performance matters that much with huge data amount?
Why do you think Ninja $995 customer support is better than Tradelink free community? I cannot find out anything Ninja itself is better than Tradelink. Ninja has its own "script" language which is waste of time for people to learn. Why not just use real C# in Tradelink with the most flexible? Using Ninja script is just a way to let Ninja to earn money.