Trading setup....can anyone give me a brief (or not so brief) overview

[ppcassidy]

Hi,

I was wondering if anyone can give me a brief overview of the trading software they are using right now. I will be more specific.

I would like to incorporate some machine learning into my trading setup. I am currently using mt4 on a demo account, but I will probably make the switch to mt5 shortly just because it opens up more possibilities in relation to connecting to other software packages and it allows you to work with multiple pair EAs.

I've read how some people are using matlab, r, weka, Neuroshell, Theano etc and a whole host of other math/NN packages.

I would like to know

(a) What setup are you using?

(b) Why are you using that setup - what strategies does that particular setup give you access to?

c) Is it working - are you seeing concrete live trading advantages to using that setup?

I have a computer science background, though I have no knowledge of any of the software packages mentioned above. But I'm not afraid to get my feet wet and put the time in.

I've been reading a lot of academic papers about the merits of using an BPNN system and hybridized systems but I would like to get some read world evidence that this stuff works. I've read a few threads on this site and on others but they all seem to be at least a couple of years old, so some up to date info would be great!

Please feel free to share as much as you want, throw in links, the whole 9 yards

Thanks.

P

 

[Shakone]

YOu might find this blog worth a read

Intelligent Trading: Practical Implementation of Neural Network based time series (stock) prediction - PART 2

 

[NVP]

you certainly have the math dude ....so you have an enormous advantage over most "grunt" traders already ......

but I still suggest you spending some time in the trenches reading and researching all the many simple methodologies out there alongside learning all the MM, RR and phycology , Zen , mindset etc etc........the works

you have all the time in the world to fuse the Programming into what and how you chose to trade........so take the time to find the best system for you to apply

most programme based traders lose because they dont get down and dirty enough first (or ever) and realise the world is not perfect........ and never will be ....!

N

 

[ppcassidy]

Thanks to both of you for your input....appreciate it!

NVP I agree that this is something I have to do in parallel with getting my hands dirty and doing what needs to be done. I am hoping that if I become knowledgeable in what I admit will be a largely academic endeavor for possibly a long period of time (supervised learning models etc) somewhere down the line I will be proficient enough to tie it all together.

I think that a lot of people see NNs as being a plug and play solution to the trading issue and there are so many parameters involved in the meaningful and successful implementation of these solutions, that you have to attempt to gain a decent amount of knowledge about them and their setups. The problem is that there are so many solutions out there and each one is so proprietary that if you invest time and effort into one (and money), you may realize that you've wasted a lot of resources with something that could never work. Hence this thread.

But thank you for your input. I do agree with you. I am going to start working with MT5 soon - just because it should open up more doors in terms of it's connectivity to other software packages.

P.

 

[NVP]

Thanks to both of you for your input....appreciate it!

NVP I agree that this is something I have to do in parallel with getting my hands dirty and doing what needs to be done. I am hoping that if I become knowledgeable in what I admit will be a largely academic endeavor for possibly a long period of time (supervised learning models etc) somewhere down the line I will be proficient enough to tie it all together.

I think that a lot of people see NNs as being a plug and play solution to the trading issue and there are so many parameters involved in the meaningful and successful implementation of these solutions, that you have to attempt to gain a decent amount of knowledge about them and their setups. The problem is that there are so many solutions out there and each one is so proprietary that if you invest time and effort into one (and money), you may realize that you've wasted a lot of resources with something that could never work. Hence this thread.

But thank you for your input. I do agree with you. I am going to start working with MT5 soon - just because it should open up more doors in terms of it's connectivity to other software packages.

P.

welcome to trading .......ive now spent over 10 years developing one indicator and method ..........we wont talk about the other 20 years or the many alleys i've wandered that could easily create a city !

N