Fully automated Portfolio trading system looking for partner.

Alright, let's discuss the trading aspects then.

Someone claims they have a valuable strategy. It's simple to evaluate a strategy: does it have an edge or not? All the jibberjabber about a billion parameters and trader intuition and sentimental analysis is just dust in the eye to cover the fact that there's no edge whatsoever. No way they're having an edge, they do not have an edge, never edged, never had an edge, never will have an edge, no edge! (Some might recognize the Seinfeld reference )

Since many of you are familiar with software development, it's the same situation as with programmers: "There are two methods in software design. One is to make the program so simple, there are obviously no errors. The other is to make it so complicated, there are no obvious errors."

Rephrasing for snake oil trading systems: "There are two methods in trading strategy design. One is to make the system so simple, there is an obvious edge. The other is to make it so complicated, that it's not obvious it has no edge".

I rest my case

 

Your First Logical, relevant and excellent post on this thread. You asked excellent question and more importantly you summarized it beautifully in words.

If you want, I can give you my take on this. Only if you delete your other posts on this thread AND/OR apologize for your insulting and borderline racist behavior on this thread.

Sarcasm, comedy, Criticism, Peer review, Skepticism etc all are important aspects of seeking of any kind. But there is a thin line that if you cross, will take you in the domain of insulting, arrogance etc. I don't want to spend more words on explaining why and what is wrong in your previous posts. I want only to stay only with logic and relevant topic here.

Choice is yours.

 

If you narrow down the holdings to a more manageable number, you might be able to trade the system yourself. For example, in pseudocode,

Code:

set_of_symbols_to_trade = empty set.

sort universe of symbols high-to-low liquidity.
for each symbol symbolToTest in sorted_universe
do
    for each symbol symbolToTrade in set_of_symbols_to_trade
    do
        if correlation(symbolToTest, symbolToTrade) >= user_defined_threshold
        then
            skip to next symbolToTest in sorted_universe
        endif
    done
    add symbolToTest to set_of_symbols_to_trade
done

The result would be each symbol in set_of_symbols_to_trade is correlated to each other symbol in set_of_symbols_to_trade by less than user_defined_threshold.

 

Thanks for your constructive reply here.
the approach is not applicable. Because the assumption here is all edges are equal. It is just not about the threshold of edge, but type of edge too. (and so many other things that can not be explained here).

BTW..I am getting more and more clarity on how to proceed further, and trading with own real money *exactly* with same system is inevitable. I will soon close this thread (i don't know how or if i can ) with closing comments.

 

Well sorry if I offended you, it wasn't my intention. Mind though, finding an edge is next to impossible, which is not to say an edge is a prerequisite to making money - plenty of successful examples otherwise. Perhaps the real genius is making money with no edge. Like no way they're providing an edge, they do not provide an edge, never edged, never had provided an edge, never will provide an edge, no edge whatsoever and yet they get valued at billions of dollars (for example, Robinhood).

 

Correct.
to give another reference from movie Fight Club. "On a long enough timeline, the survival rate for everyone drops to zero."
Even hardcore successful quantitative Funds for years like Rentech (Dr. Simons) says similar thing. Thing are working for now and it seems like a edge, but we don't know. There is possibly a wolf out there staring in our eyes and will get us. Can we avoid that wolf in our time that is the only challenge.

Theoretically, Logically, mathematically there can not be edge.
But try to understand, Market is not a real(natural) thing like Gravity. You can apply all the mathematics and code on a natural thigs and study it, understand it better and solve it if you can.
One can not solve market. Because there is nothing to solve. Market is product of Mind and so is our Religions, Gods, cultures, Nations, Currency, Love etc etc.
Any product of mind will have it's own mind and nobody will know how it will behave.

Humans have created these myths and it helped humans to achieve big things. e.g. Natural limitations on other species to collaborate (herd, team work, Pack) is like few hundreds to thousands. Humans had that limit of (say few hundred in a Tribe)) few thousand years back and then we discovered some Myths. some Myths stick and we were able to unite on a much larger numbers and create much larger impacts. China wall, Science discoveries, Space etc.

So some of the myths mentioned above become so big that we started to treat them like Realities because they started to impact us like Realities. (Trust me, IRS affects me much more than Gravity).

But at the end of they they are not realities. 500 years from now, will there me Market as we know now? I am sure there will be gravity. Nobody will remember names of people who fond edge(?) in market in 2100 era.

Another way of looking at Market is Time. Market is actually Time. Can you win against time? Nobody can. The only practical question is, can we play better in whatever time we are given? Can we make it count? and that is what should be the aim of any Quant/system developer/trader out there.

I know it all sounds like woo woo and for many of you instant reaction will be to call it snake oil. To summarize, I agree that theoretically there can not be edge. Eventually Apple like successful company will also die. But can i still make money buying appl stock low and then selling for profit in my lifetime? That is the only practical question.

and from that practical point of view, edge should exist.

 

Correct. And that is a Million dollar observations. One can explore further on that idea alone towards building a good automation. The practical truth is somewhere in between the 2 realities you mentioned. Somebody else mentioned on this thread (criusecontrol?) that he/she automate all that is possible and Ignore totally many other factors. That is a practical approach. Too much complicated system will naturally absorbs too much of energy and can not be efficient.
I can not comment more on this for practical reasons.

 

Thank you all who contributed, commented, suggested, wrote Pseudo code, moderated and at times Insulted on this thread. It helped me get a clarity on my though process. I hope I have given something back via my thoughts that might be useful for somebody someday.

What I have realized is there are 3 paths to take this forward.

Path A: Share Flow chart and pseudo code, entire code, strategies for peer review.
Path B: Share just enough at higher level and try to seek partner.
Path C: Actually Trade system with real money and generate results that can be submitted for review.

Now pathA, I will not take. Any quant worth any salt will not sell his code before he sells his kids. And no, I am not selling my kids.

PathB, the probability of that success finding a match is less than 1 percent. and worst is, this path sounds like a snake oil sales. More than I love my code, I hate the snake oil salesmen comparison. So no that is not a path for me.

that leaves PathC and I will figure that out now, as i have more clarity on my choices and things that i need to do. That is the only logical path.

I will not be checking this thread as often now. I have things on my plate now.
Thanks everybody.