I don't disagree with anything you have said and I will expand on this later when I have some time. Thanks for the lengthy reply, but you did not mention when/if you are tweaking these strats and how often you are doing this? Btw, I do have access to non retail software and hardware/implementation. That is why I am really not sweating the small stuff right now plus the team I am working with is there to work with me and for me not the other way around which makes this nice. Not quite the real big time yet, but I mean better than plugging away in AmiBroker if you know what I mean. Anyway, gotta run
Nice. That's essentially what I'm doing, though not fully automated. Working for me so far, though I'm always striving to do better.
OK. Back. 1. No doubt, I totally agree. I have multiple models - most are 1 minute but I do have a couple 5 minute programs and then also another program that goes both ways (5 min as well - others hopefully off the ground soon enough). I guess that wasn't conveyed in this thread but this thread meandered into something other than what it was meant for. Mainly focused on ES - DAX - TOPIX - KOSPI 2. Agreed. Hopefully it will be benefiting me as well. 3. I know the terms and what they mean, but if you asked me off the top of my head how to calculate the Sharpe Ratio of one of my programs I would definitely not be able to do that. The odd thing is I have never once needed any maths that were any more advanced than high school. Realistically, more like 3-5th grade. I am not sure what all these "quant" guys are doing - either something so magnificently above my level of thinking they just laugh with pity at this thread and move on or they are just good at studying textbooks while concocting the next linear regression model trading program. Most likely did well in AP Lit or Latin in High School as well. Probably both. If this series of recent "HFT" pictures from that guy selling his trading platform and data feed (NxCore) through his picture book is actually legit, did you see that knife pattern? If that was real, and IMO that is a big if, that is something like I would come up with and I would be disappointed as fuck if some "quant" getting paid $250,000/yr + bonus would have to resort to programming the "knife". Now I am probably sounding a bit jaded, I guess I am that asshole 4. IMO creativity is the precursor to the analytical process. But I would have to agree with exactly what you said. I honestly think some of the best ideas I have ever came up with have came from my fractal imagination while my eyes are closed if you can read between the lines. 5. I would have to maybe argue a different way. Lets say you have recently came up with a 15 minute candle "system" with recent data in the year 2010. You found this pattern that could have potentially traded 11 times this year. Are you going to attempt to trade this program in 2011? I know the answer is no, and I do not think you were going there - but yeah. 6. Yes, among other things, I was reading in another thread here I believe that IB was cracking down on some cancellation policy requirement or something... Luckily I do not have to worry about these things. Or else I would most likely have 50 : 1 ++ ratio most days. Multiplied by a handful of programs. Hopefully we get that cancellation tax being propagated by 15% of this website's membership (for some reason) so I can just get it over with and move into my car. Luckily the seat folds down in the back so it's kinda like staying at a Comfort Inn if you have a really good imagination. Your last paragraph has an assload of truth to it. But my research has shown if you are going to be executing these programs that I will be potentially (god help me) running that it is best practices to be continually refining your strategy and it is highly probable this strategy, once used, will never be used again. That was the real point to this thread, but it doesn't really seem like anyone is doing the same thing as I. If someone is they do not like to share. That's cool though, guess I'll have to smoke this bowl of low grade all to myself then. Even took out all the stems...
You are right. There are two challenges. The first one is uncovering the stationary distribution. The second one is trading it. Although I don't use the same method to expose noise, there are similarities in our trading strategies. Options are indeed the way to trade such spreads. Alternatively, when the normalized price is low you could compute the probability of price increasing vs the probability of volatility decreasing. Maybe prices increase more often than volatility decreases and it leaves money on the table for a directional bet. Ninna
I am exploring trading interfaces for automatic trading. What program are you working with to build and backtest, and then which broker is recommended for data and exec