Designing a System from scratch

beginner_av

Well-Known Member
#71
very interesting stuff simple. at this point all i can do is echo CV.
but let me ask you why are you doing this? if it is for learning, even if it is public, then my advice is have a plan. put it down here. if you are a beginner, write out the concepts here for yours and everyone's benefits. for example, you can start from the conceptual framework - what is volatility? it may help you understand if there is any relationship between volatility and inside day at all, and if so in what way. why did you choose inside day? what is your timeframe?

thats all i can say. keep up the good work and enlighten us.

one more thing. your first step incorporates many of the modules. go one by one. dont go into set up/entry/exit all at one go. why already decide on the filter before deciding on the type of volatility system that you want to address? take it step by step and you will be clearer.


and ultimately - Let the BOSS be with us
 
Last edited:

kkseal

Well-Known Member
#72
One thing here regarding Volatility Cycles is that even though it may be cyclic its difficult to distinguish where one cycle ends and other one begins bcos amplitude and phase are constantly changing.
CV, could you please elaborate a bit on the phase & amplitude aspect (would be wonderful if some 'Auto adjustment' feature can be built into indicators; but for that one needs to understand the stuff better. Requesting a 'tutorial' from CV, Raja & all else who may have knowledge of the same)

Regards,
Kalyan.
 
U

uasish

Guest
#74
Simple,
I am on the job for collecting all type of codes on volatility ,whether in Ami or MS.Actually my personal process of understanding is always going through the codes,becoz in a code it is all writen ;what the code writer has understood by going through lots of books ( i am bit lazy).
After that i will make a workable one public in this thread,then we will incorporate the 'Amplitude' & 'Phase' part .(few half done work is already there in my work) .
To-day morning was studying few approaches to Volatility ,there are already few established methodology practised by Exchanges world over for determining the S.P.A.N.(Standard Portfolio Analysis of Risk ) margin the daily VAR (value at risk) the Brokers follow these exchange parameters .
Plz continue i will chip in whenever i get something worth to share here.
Asish
 

beginner_av

Well-Known Member
#75
sorry for downgrading a great thread of CV. but I have to go back to the basics here. in this era of software, we have a tendency to talk everything in terms of code, and due to easy availability, often we start with the code itself that we collected from a magazine, forum or other places. whereby its perfectly sound to apply systems analysis and design concepts in developing trading systems, still the ideal point will be to start with a pencil and paper.
can we put the whole system in plain english on paper? because if we cant, then the code will forever remain incomplete or inaccurate. only after describing the system in plain language can we go ahead and translate it in any programming language that we like. computers and code are here to add horsepower, scan through hundreds of instruments, look for combinations, optimizations etc that would be almost impossible by hand. but to think that the code is the start and the end of the system is absolutely incorrect. CV had also put the same thing, but instead of plain language, he used pseudocode.
 
U

uasish

Guest
#76
Absolutely right bav ,that part is Rajesh taking care of .He has also given a step by step method for the Module.I am prepering in my way ,to chip in when or if reqd by all of you.You & Ajay will be more better equipped to assist Rajesh in building the basic foundation.
 
Last edited by a moderator:
C

CreditViolet

Guest
#77
can we put the whole system in plain english on paper? because if we cant, then the code will forever remain incomplete or inaccurate. only after describing the system in plain language can we go ahead and translate it in any programming language that we like.
Indeed.Very correct.Thats the conceptualization part of strategy development which very few people write down or give much thought about.Having a global template for other parts of the strategy like Sizing, MM, Scaling In/Out helps to stay streamlined as well.

Fredrick Brooks estimates that it takes nine times the effort to create a complete, reliable system as opposed to an initial program that starts to do the job. From The Mythical Man-Month
 

oxusmorouz

Well-Known Member
#79
My 1st neural net

Attached is the excel sheet of how, in my view, not to proceed to create a neural network. It took me 3 hours to train a net for 500bars of input data in 4 columns and some 591,000 odd epochs to build 1 predictive model and bring the standard error down to 4.17%. The reason of course is my half cooked knowledge and 2nd grade tools (I wonder if they would have been 1st grade had I much knowledge in NN based models?). Anyway, I hope this acts as some foundation to higher learning. I have extended the prediction of close by 1 day, which is shown in the E4055 of the excel sheet.
Looking for comments from experts.

http://rapidshare.com/files/44565410/Nifty1.xls

Sincerely yours.

P.S: CV, am I proceeding in the right direction? If no, please rectify my mistake. Thanks.
 
#80
Re: My 1st neural net

Attached is the excel sheet of how, in my view, not to proceed to create a neural network. It took me 3 hours to train a net for 500bars of input data in 4 columns and some 591,000 odd epochs to build 1 predictive model and bring the standard error down to 4.17%. The reason of course is my half cooked knowledge and 2nd grade tools (I wonder if they would have been 1st grade had I much knowledge in NN based models?).
.
.
.
Could you elaborate on the tools being used in this spreadsheet?
 

Similar threads