Actually our other product, QuantDeveloper, features a "component" strategy development model where you can build a strategy from Entry, Exit, MoneyManager, etc. components. In this model you can develop a strategy playing a lego game, i.e. combining f.ex. SMA entry with Stop exit and ATR money manager. Although this can sound very attractive for a non-programmer, there are problems with this approach: 1. a user still need to program new components (the same in TB) 2. the model is not good for automated strategies where you need to trace order state change between components Thus I think OpenQuant offers the most simple ATS development model where a lot of things happen behind the scene, so that actual strategy programming is very simple. As for you marketing advise... I read it as "add a couple of zeroes to product price and position it as a hedge fund product since masses are not able to use it anyway"... This is a good point to consider
Seriousely speaking I think you you do not take into account that OpenQuant is a product for automated trategy development and TB and other products are designed for backtesting. It's pretty easy to introduce a rule builder or strategy development wizard in the case of backtesting since you do not need to take reality into consideration. It's easy to generate if A crosses B then Buy else Sell and call it visual startegy development but then think that a simple Buy in backtesting world can turn into a variety of different execution scenarios in the ATS world. You order can be rejected, stay pending for a long time, partially filled, partially filled and then cancelled, network can go down, etc. You should take all this into account when you develop an ATS. I agree that it's a very good marketing move to have a visual strategy development wizard but to put it simple ... we don't know yet how to develop it right, so that a user can develop a fully featuring ATS with such a wizard and we don't want to offer something obvious but very limited (like RightEdge startegy builder).
While it is evident the RightEdge visual development facility has limitations, it also on the other hand has some interesting uses. For preliminary investigations of ideas using simulation and optimization it can be convenient and efficient. More usefully, the objects on the visual surface can be referenced in code. The mixed mode can be convenient for tinkering with aspects of a strategy that is in the process of being refined. For the non-programmer, since fully functioning strategies, albeit simple, can be constructed, it offers a window onto gaining some practical experience with what strategy development is all about. A visual facility, as just one of the ingredients in the tool chest, has its uses it seems to me.
We have about 20 strategies from Trade Like a Hedge Fund book in OpenQuant install. We tried to understand if we can develop these strategies using a tool similar to RE or NT 6 rule builders. The answer is no way so far... but we are thinking Another point is that it doesn't work both ways, i.e. you can generate some strategy code from visual designer but you can not generate visual layout from the code (think of VisualStudio design mode, where you can change it both ways). Thus it doesn't look like a mixed mode but rather like a "locked" mode ala WealthLab wizard.
OK Does anyone have experience with the NinjaTrader 6 app? From my understanding, it can also be coded using C#.
We are working with a programmer currently on NT6.0 and will let you know the detail in 2-3 weeks once the initial work comes back. The .net framework should cover the needs of most. We will see in time but Version 1 will be buggy I suspect but can improve from there. Finding that one program that allows for system dev, testing and than execution for the trader past TradeStation should do well in this niche market.
As the starter of this thread, I suppose I should give my conclusion. I liked RightEdge quite a bit, and I found the company to be quite responsive to my questions. In the end, though, I wanted a product that would help me get my ATS up and running in a robust and professional manner as quickly as possible. This turned out to be the SmartQuant suite of products. It has advanced my project by several months and ended up saving me tens of thousands of dollars in development costs. While I was terrified at the prospect of learning C# to operate QuantDeveloper, I have a) found the language to be actually rather fun to learn and b) realized that QuantDeveloper has much in it for the non-programmer. So I'm a very happy client of SmartQuant and user of QuantDeveloper (the sibling to the newly-released OpenQuant). It's funny how sometimes really difficult decisions seem so simple in hindsight. polr
I think it is imperfectly stated. Once you know what your system needs to be, not once in my life have I seen a canned product be able to handle the kind of objects and computation on those objects that are needed for these types of systems. Most of these packages are built for people that want to test systems built out of canned indicators. They are essentially worthless. Programmers that build tools that "never" get used don't even know the game they are going after. They lack domain knowledge and are hopelessly misdirected. That is why the tools never get used. Otherwise, the effort to build a custom tool would be so concentrated with the result so clear in ones mind, that each day you would see your system that much closer to going live with real money. The difference is, that a professional programmer, when he is building a tool for a specific system, knows how to program for generic systems, and he has in mind how all the pieces fit together, so he programs for the generic case. nitro
Polr, You bought a QD license from Quanthouse or you are using OpenQuant? QD is quite expensive for retail traders -- $000's/month.
Use NinjaTrader. The soon to be released v6 can be programmed in C#. For $50/mo you can roll your own indicators and strategies, backtest them and then go real-time.