Honestly, regarding the premium support option, it doesn't excite me. My thinking currently is that once I really am beyond the initial learning phase I won't need more support than what is already available... I'm in a "slow down to speed up mode", just plugging along and reading the forums and coding stuff and the need for support is growing remote... I'll vote for it however as it would be a nice insurance policy....
Definitely see what you mean. I kinda know what you mean, right when you begin you must ask ?s in order to figure things out, but then after you really get to know the platform you really can figure out anything
Maxpi and Sky123987, I found the answer to my question here : http://www.smartquant.com/forums/viewtopic.php?p=17831#17831 ...OpenQuant simulates the entire FIX message chain when you submit or cancel an order. This guarantees that the order management system works identically in the simulation and live trading modes and simulations are pretty close to reality. The price we have to pay for this is simulation speed. It's not that bad if you have a few orders in your strategy but if you sumbit / cancel orders on every bar or tick, then the order management slows down simulations a lot. ....
They are working on a 2.0 release that is supposed to address the speed problem to some extent, personally I am not too unhappy with the speed. Optimizations can be done iteratively, do a test with bigger steps to find the good zones then repeat it with a smaller parameter space and smaller steps to zoom in on it... break the test into unrelated parameter sets and test those sets individually, etc. Probably everybody in the world that is doing backtests is used to doing all of those things already. I had to do the same thing with Tradestation 2000 and 8. If they make it faster, fine, but speed is not a show stopper at this point and my processor is a couple years old too... I'm just now realizing how tightly integrated OpenQuant code can be with TWS, it has the event driven methods to update things and the code to do it... that was what made Tradestation unacceptable to me, your strategy could get out of synch with the brokerage and there was no way to handle the situation, same for Multicharts currently, they keep promising to release a version with full ordering capabilities, I don't care what they do anymore, all those years with Easy language made me more than a little crazy essentially I'm in hopes that OpenQuant will restore me to full functionality, make me a productive member of society again.....
****Limit entry simulations***** I have a system generating limit entry orders (using OnBar). I'm trying to get more realistic backtesting results by not entering at the High/Low of the next bar. For instance, if I send an entry buy limit order, I don't want the simulator open a position at the low of the next bar if that low of the bar is equal to my buy limit entry price. In reality, I will most likely not get executed at that price. In other words, for backtesting purposes, how can I access NEXT bar's high and low values? Thanks.
Hello Sky, Thanks for the answer to my question. It works. Yes, I also ask questions on the OQ forum. Unfortunately, it usually takes quite long to get any response from OQ forum. It seems to me OQ does not have good documentation at all. I have another question for you. How do I write cross(close, ATR(atrlength)) in OpenQuant? I only know sma1.Crosses(sma2, bar) from the OQ example for 2 moving averages cross. Here, in the cross(close, ATR(atrlength)), the value of ATR(atrlength) keep changing and it is not a SMA. I don't know how OQ handles this. I am new to OQ and only used it for about 1 month. I have used Amibroker for a few years and found it way easier to use because it has much better documentation and way more responsive technical support. Currently I have not yet found any big advantange for OQ over Amibroker. Thanks, Gundam
Gun, Not really sure as I really don't use that kinda stuff. On a basic level all the ATS out there do a fine job. However once you start to add an element of complexity to your strategy coding in the others is a nightmare. Especially when you are dealing with limit orders that are filled / partially filled / if it is partially filled and you want to add a profit target, if you send an order and a black box that is co'loed right near the exchange sends an order to counteract your order and it's processed quicker than the latency for the message to get back to your computer, and many others ---> all these can lead into very complex and harry situtions. IMO the openquant framework enable you to write code to get around these situations better than any others