Even with my very limited programming skills, I can see that you are right about QD being more advanced than RE. What I've read on Willmott and Nuclear Phynance, though, is that QD isn't all that great from a true quant perspective, as it's just a bunch of modules thrown together. Keep in mind that these posts are written by some true quants--individuals who would use about 50x the features that I ever would. I am now seeing firsthand why so many ATS traders develop their own systems from scratch. There are just so many different ways to go, and if real dollars are being traded, then it makes sense to get everything just the way you want it. As I a) realize that it would take at least five years of hardcore study to become a truly competent programmer and b) wish to maintain a large degree of control over my project, I sure wish the Wealth-Labs out there could do everything a bult-from-scratch solution could. It seems like I'm trying to rewrite MS Word when MS Word itself is a great product and costs a mere hundred bucks.
Great points...I agree: hate monthly subscription fees...and I can't comprehend why anyone would use TradeBolt, etc. rather than plugging in to IB or REDIPlus directly. An open source initiative would be great...I guess that's how QD started out. I think the market for *robust* ATS platforms is in its absolute infancy. I'd like to say, "Give it time," but I don't want to wait.
Valdis, I don't mean to bash your product or your company. In fact, I think QD appears to be a great product. My point in mentioning the porno ads in your message forum (along with unanswered questions) is to show that you guys just don't provide the best support (at least right now). I'm not saying that SQ is "offering porno"...I'm saying that you guys aren't going through and deleting the porn ads that ad bots are leaving there. I personally think that the retail market has gotten very sophisticated in recent years. Therefore, I don't think that pro products should differ much from retail...maybe just cut down on number of symbols a system can handle in the retail product or something like that. But I think it's a mistake to dumb down the retail product too much. polr
WLD does in fact support futures. Fidelity does not in its Wealth-Lab Pro (from what I've heard). I've said it before, and I'll say it again: WLD is an amazing product. It's easy to use (this coming from a person who just recently learned to program "Hello, World"), it has loads of features, and it has a great, knowledgeable community. What I want is a completely robust version of WLD that can handle a very serious data load along with a thousand trades per hour. If a company could provide this, I'd pay tens of thousands of dollars for it this instant (which is less than I'll end up paying to professional programmers over the next six months for a product that is not as easy to use as WLD). polr
The big difference I've noticed between the two is that SmartQuant has a built in portfolio object. This means that doing actions across different positions (basic position sizing, resizing based on volatility, or position size limits, etc.) can be done very easily. You can write some basic code with just a bit of knowledge. The problem I have with SmartQuant, however, is that they have two types of systems - I may be getting the names wrong here - ATS systems and "other system" (can't remember the name). ATS systems have the advantage of being deployable on their standalone platform for full automation (think: I want to put it on a server that monitors the market and makes automatic trades). The other type is designed just so that you can perform backtesting and easily swap in different techniques for entry, exit and money management (TradingBlox does this very well from what I've seen as well). With ATS systems, you have to have all the code, essentially, contained in a single object. This means that you develop your system in the basic system and have to convert it make it work as an ATS system. I just don't understand why they set it up this way - it is completely counterintuitive to me - maybe it makes sense to others. RE, in contrast, doesn't have a concept of a portfolio object. So it is not as easy to switch to, say, volatility based sizing. Now these guys know this - I've talked about it extensively with them. As they talked about in a forum conversation with them: "For your money management, you can accomplish what you're proposing. Admittedly, I think QD probably has a better system out of the box as far as componentizing money management. If I understand correctly, I think you could get the same results by writing a money management assembly and adding it as a reference to the project. In other words, you code up a vol sensitive money plan or ATR/vol based money plan and in NewSymbolBar() you feed your module the current "situation". Your module would determine based on its rules the position size and feed that back to RightEdge. Keep in mind that you have access to all bar data and indicator values (up to the current bar, no cheating!), so you can feed that to any module you'd like. This is definitely the beauty of using real .NET languages. Ideally, we could write money management plugins that you could select in a drop down and they would get applied automatically. My main concern is that everyone's money management is different and to make this work, there would have to be access to source code or you'd have to implement your own. We'll give this more thought though." For reference: http://www.rightedgesystems.com/forums/Topic275-7-1.aspx So, I guess what I'm saying is this: if you don't need portfolio-level testing, then just about any product can do this. In fact, if I was doing just single-instrument testing, I would use Tradestation - it's fine for that. But if you move beyond that, or you need to plug in more custom code via a C# interface, then QD may be a better choice currently. One thing I want to add about using C# - despite the fact that you are learning a standardized language, every platform uses different objects and concepts - so it's not like you can just port code between them. I know this is obvious, but I think it's just worth mentioning. Make sure you also take a look at NeedforTrade - they are C# based as well. Hope this is helpful....
Well, I think products for backtesting, which is really what we're talking about here, could easily become commodity. They will just move up the chain more and more - so first just backtesting with simple optimization. Then someone will come out with a free genetic algorithm. Then someone will come out with something else...etc. Auto trading systems, however, seem less likely to be commodified. So no real disagreement - I just think it is important to define the products types separately.
Valdis, Thanks for the reply. It's a bit clearer to me now. After reading Derman's "My Life as a Quant," I realize that I am no quant. I have no doubt that QD being a "bunch of modules thrown together" wouldn't affect my strategy. As for the porn ads, I can say that I've never seen one on the Wealth-Lab site, and I've spent a lot longer on that site than on yours. Don't get me wrong--I'm the last guy who is offended by such stuff, but it's tough to take a biz with that in its forums very seriously. I really like QD, and I was a couple days away from forking over $15k to get it. But I would go literally *two weeks* at a time to get replies to very simple questions. For example, I asked why I should purchase DataCenter instead of using SQL Server...and I never did get a reply after asking this question about three times. I repeated other questions in subsequent emails, but they too went unanswered. This really puts potential customers off. The guys at RE reply very quickly to emails, and they go above and beyond in replying to all questions. I figure I can pay a fraction of the cost for their product and then pay them to make RE do everything I want it to do. So the bottom line to QD (as I see it, anyway): great product...a bit overpriced...horrific customer service. I genuinely wish you guys the best, though. I hope your product (and your company, for that matter) proves to be extremely successful. I really want to see professional-level ATS software vendors succeed. That will only serve to improve the software that's out there. polr
You make some good points here. I guess this is one of those areas that would require some consulting by the makers of RE. It would be nice if RE 1.0 had these portfolio-level features. I guess we have to wait until 2.0. It was nice to see these features very clearly marked/defined in QD. It seems like we're all describing very similar features that we want to see in these products--yet there isn't one company that provides it all. QD has probably come the closest...until they effectively exited the market. RE seems to be the one headed most in the right direction, but they're not there yet (but they could be moving there very quickly). Next week, I'm meeting with a programmer who sets up ATSes professionally. While I'd rather buy something off the shelf, is this really the only way to go? (Boy, right about now I sure wish I took more programming classes and fewer worthless finance classes in college!) polr
Well, you should consider TradersStudio. While not a standardized language like C#, it has a lot of the features I was looking for. RE and NeedforTrade are comers. I'm also watching TradeSignal to see if they make a US push. Personally, I want to concentrate on using the tools rather than creating the tools. I like having, in particular, the portfolio stats that I don't want to spend time programming. My own belief is that you can get too wrapped up in creating the tool vs. using the tool - you have to make that decision in terms of how you want to spend your time. Me, I want to be figuring out how to, hopefully, make some money! But do I understand why people create their own tools? Absolutely....
This is so perfectly stated. I think that the traders with a lot of programming experience really enjoy rolling their own. If I knew programming better, I'd probably do the same. I'll never be as good a programmer as those who built RE, QD, WLD, etc., so why should I even try?