I'm cross-posting this from the strategy thread. How would one go about reviewing market data to find statistical relationships that "predict" the price movements? What tools would you use to do this? What step by step process would I have to take? Thanks in advance.
Anyone? I'm sure there's some people who do this in order to profit in the markets on a consistent basis. Throw a guy a bone.
What type of analysis do you want to do? I'm a statistical programmer, and I have modeled potential trading systems using SAS. There are other packages that can do it just as well. Do you already have data in hand?
One more thought... You need to be very clear on what exactly you're trying to analyze with your model -- garbage in, garbage out. You should Van K. Tharp's book, Trade Your Way to Financial Freedom. He talks a lot about system design.
It's not a simple answer. I would use OpenTick for their free tick data. Then I would write a program (C# would be my preferred, but you could use Excel, I suppose) that analyzes correlation between different instruments. For example Bonds & ER2 etc... Or you could check some of the sector ETFs and compare with individual stocks. New highs, new lows etc...
Relationships between correlated insturments is at least for me the key to profitability. Most importantly, start simple, just overlay 2 charts of 2 correlated instruments. Try to find relationships, think of divergence, convergence. Sundog
I just read some of your old posts....dude you are an asshole with an entitlement complex. Why did you have your old thread locked? Did all the boring old people that gave you advice get lucky in their predictions? Look at some of the shit you say in your posts. I didn't contribute to this thread, but dammit, I wanted to say that.