Actually... Here's an open source project that looks interesting. Didn't spot this the first time around. http://www.marketcetera.com Anyone have any experience with these guys?
I fail to see marketcetera as open source. Look at the price to pay for real use. Really. Yes, I may get the source, but it is not "Open Source". That said, if ninja 7 does not do what I want I will step in. Moving my small development shop over to trading slowly, so I can as well put in the additional funds to have a PRODUCT, not something for my own use But that means 4-6 months for early beta, sadly ;( And I am not generous, I just fail to see where I loose on that.
Yep. Having spent the last half hour reading up on them this is a classic example of where commercial open source ends up costing more than proprietary software. Also, for various reasons I prefer C# as a language and really want a native .NET app...not some bastardised .NET/Java hybrid. That's just another point of potential failure and performance degradation as far as I'm concerned. Looking forward to seeing what you decide to release to the outside world NetTecture. Best of luck in your efforts...
Ninja is good for basic charting, but other than that I am hoping ver 7 is more for someone who wants to test ideas rather than learn C# and the ins-and-outs of Ninja-specifics. One thing I found strange is how Ninja marks it's bars. If I have a 5 min chart and a 30 min chart and I look at the bar open price at 9:30, I get differing prices for the Bar-Open. You have to look 1 bar back on the 5 min chart for the 30min open price to match up. They say this is because they don't know when a bar closes until the 1st tick of the next bar, but how do eSignal, TradeStation, and everyone else sync up the pricing on the diff charts then? Their support is often "we don't support answering that question" too. But, again, for basic stuff, it's fine.
Are you looking for performance? It's hard to believe anybody can beat tickzoom if you care about performance. I tried it with 100 million ticks and it finished in just about 5 minutes and never took more than about 80 meg of memory. Sorry, I didn't keep the exact results. When you do like 1 year or 2 years of tick data it only takes seconds. It's scary how fast it is. And given all the other attempts by people to build platforms and failure. The odds are against anyone else getting it that fast. But tickzoom has other issues. So weigh your options. Speed/Performance isn't the only thing, of course. But, wow. It's really nice. I must say.
Well, I'm not fussed about busting a gut to shave milliseconds off because my trading strategies aren't that time critical. But I'm always going to avoid performance penalties if I can and will always go for the most optimal solution which meets the business requirements. I really couldn't give a toss about backtesting either as I think it's pretty much useless. Thankfully my educational background in physics brought the horrors of curve fitting to my attention many years ago.
What about RightEdge? Seems they have a decent interface and are constantly improving. Haven't used it yet but am downloading the trial version today. They also get great reviews in the customer service department on these forums.
Hm, for me they are too focused on systems. I miss the "ad hoc" component, the order entry. At least my impression so far - they focus totally on automated. Besides that they look good Better so far than ninja.
I actually looked at RightEdge before getting to NinjaTrader and liked the demo a lot. The developers have obviously planned it properly and the API is much cleaner than NinjaTrader's mess. The environment has also been modelled on Visual Studio in the same way OpenQuant have modelled theirs on VS. (NinjaTrader, please take note for NT7!) However, it just doesn't feel finished. And I can't put my finger on why. However, if it avoids chewing up all your memory and then falling over like a drunk off a bar stool, then it's lightyears ahead of NT6.5.
And I remember now... I also couldn't figure out how to programmatically get access to the level II datastream. And that's when I stopped evaluating it.