If the two stochastic processes are positively correlated and we use a superimposed subtraction of those processes the resulting process will have its variance that is less than the variances of either of the two original processes. That is exactly what arbitrage is based on.
What maximum ROR return on risk capital do you believe are achievable?. Not your actual returns or drawdowns, but what your current thinking on what's realistic with a fully automated robot.
I share the same "annoyance" towards many economists myself There are a lot more of pseudo scientists around than the real ones. I do not pretend to be a "hard-core" scientist; I am just doing my best to create new behavioral models and test them thoroughly. Cheers, MAESTRO
You might not like my answer and I might come across as a prick, however, if you are asking about my honest opinion and you promise not ask for the actual results I can tell you that 8 - 10 % a month is more than achievable. I can also tell you that the Sharpe ration of above 2.86 is also very real. Now, remember what you have promised!
That's pretty damn good for a computer. Being a real prick, I completely understand your posture, and I will not tell you what I know is possible for a human!. Damn 8-10% monthly is a real money machine, if it is scalable!. Corrected your typo above too. Cheers.
Scale-ability is an issue, of course! Not all the models allow us to go pass a few M. However, some of them (with much lower return ratio, like 2- 4% a month) can go to a B.
Hats off to you MAESTRO. Stash some of the money out... just in case the models break down, or the markets get shut!. Nice of you to offer all this info. A machine-trader!. What's the world coming to?.
it's not just me. it's the whole crew of very talented scientists, programmers, mathematicians that work in our company. I am just an "idea guy". They make it work. You should see what our robots look like! They are incredible. I would have never be able to do it on my own!
It is quite an accomplishment to have a fully objective automated robot using trader's principles quantified into math, like the arcsin and meander laws. I gave it a go, but have since given up. I got sidetracked with other ideas, and now I only trade my own money. Maybe I'll tackle the beast once again after reading that you have done it and try to follow in your steps. Or at least I will try to put someone on it. I'm not too comfortable talking about it too much anymore now that you say that it is doable... mainly because it is very simple but hard to implement. There are many more implementers than big thinkers. So hats off once again. Semi-automated on the other hand is quite easier, I would imagine.