Should System Developers Have Multiple Systems
Page 1 of 612 12 LastLast
Results 1 to 10 of 11

Thread: Should System Developers Have Multiple Systems

  1. #1
    Hello,

    From my research and studying, I have read that one trading systems do not work for all market conditions.

    My principal focus is day trading systems (average one transaction per day less then 60 min timeframes). I am not interested in swing trading systems at which hold transactions for X period of days.

    So far I have been creating one system what the believing that the machine will continue forever and be more profitable forever. From my research, I may be led down the road path.

    Queries:
    1. If a machine trader have numerous trading systems for different market conditions?
    2. If yes, do you recommend system traders maintain each of their systems running in sim way to constantly assess the operation and turn best performing system as required?
    3. If a system trader develop skill for obtaining when market conditions are changing and make one system adaptive for all market requirements.

    The questions are from readings that no one system remain constant profitable forever. Until the systems is highly elastic to market conditions.
    I hope I feel.

    Thanks for your remarks.

  2. #2
    Quote Originally Posted by ;
    Check this out blokes perspectives --very different egy http://www.longtermforex.com/
    Thank you Iamfx for responding to me. I will read the link now.

  3. #3
    1 Attachment(s)
    Quote Originally Posted by ;
    Hello, In my research and studying, I've read that one trading egies do not work for many market conditions. Questions: 1 ). Should a system trader have numerous trading egies for different market conditions? 2. If yes, do you recommend system traders maintain all their systems operating in sim mode to constantly assess the performance and turn on best performing system as required? 3. Should a system trader develop ability for accessing when market conditions are changing and also make 1 system adaptive for many market requirements. The queries are from...
    1) yes, develop three systems and trade with the exact huge picture of the market.

    2) yes, follow one easy trading methodology however exchange that in various timeframes - day trading, swing trading and position trading.

    3) yes, most important is to keep current with all the changing market conditions which specify the market type to exchange.

    This old article may help,

    https://www.axitrader.com/uk/market-...x-market-types

    there is a trend in every timeframe we see on our charts, understand the differences in each timeframe and understand the entries and exits...

  4. #4
    Quote Originally Posted by ;
    Hi, From my study and reading, I have read that one trading systems don't work for many market conditions. Questions: 1 ). Should a machine trader have multiple trading systems for different market conditions? 2. If so, do you advoe system traders maintain all their systems operating in sim mode to continuously assess the operation and turn best performing system as needed? 3. Should a system trader build ability for accessing if market conditions are changing and also make one system flexible for many market requirements. The questions are from...
    Hello goodoboy

    Excellent questions partner.

    System diversifiion is a great way to handle the correlation between return flows. There are many various ways that you may diversify to attain improved risk-weighted returns including:
    a) tool diversifiion....ideally across asset classes whereby you start looking for uncorrelated tools that enable you to attain higher risk-weighted yields than what is achievable with one instrument;
    b) timeframe diversifiion that also permits you to deploy uncorrelated return flows to optimise risk-return relationships; and
    c) system diversifiion where you deliberately construct your return flows to provide uncorrelated relationships.

    Of the different methods explained above system diversifiion offers additional benefits to that delivered by a) and b) in that one can actually design your uncorrelated return distributions to stay constant instead of leaving matters to prevailing market conditions. Where a) and b) become more debatable is that given that the non-stationery requirements of a non-gaussian market, there's absolutely not any guarantee that the correlation between return flows stay constant. What may be uncorrelated at one point in time, may become associated in other periods when market conditions vary.

    The way I look at systems designed to attack specific market conditions is no single system can be profitable across all market conditions; nonetheless a blend of uncorrelated systems has the capacity to tackle a suite of market requirements.

    Now either you can deploy market state filters to turn'off' or'on' your various systems to capitalise on requirements that respond to the advantages of each system....or rather you may develop quite robust systems that outperform when circumstances are favorable and handle capital stringently when circumstances are unfavorable....and therefore maintain each system turned on all the time. I prefer the latter technique as the installation of market state filters are inevitably lagging in nature and when you flip in your system, you may realize that market conditions immediately turn negative and there is a lag until you flip them off.

    Should you prefer the former'market state filter' egy, then one method to accomplish this is to run live walk ahead systems in demo mode in tandem with your live progr and when the demo accounts are entering drawdown, then turn off your live system trader until we see a change of drawdown fortunes where then you flip the live system on again....but despite the theory you may be a bit disappointed with the outcome (given the lags explained above).

    In designing your portfolio of systems it can be very helpful to consider in terms of the following:
    1. Classify your systems into two broad classes....whether they are divergent systems (capitalise on trending intervals ), or if they are mean reverting systems. Obviously these two wide system types are uncorrelated or anti inflammatory connected to each other. When one system is performing nicely another is floundering and vice versa.
    2. Determine which system you will use as your major bread-winner. This should be the machine with the best risk-weighted performance metrics and also have positive expectation. I have a tendency to discover the safest bet is to ensure the major bread winner is a divergent system with large positive skew (eg. Many tiny declines with occasional large wins). This makes sure your total portfolio when assembled also is dominated by positive skew which aids in robustness tests.
    3. Select a system from the alternate class that has positive expectancy that's un-correlated or anti-correlated together with the primary system and adopt a lower alloion weighting to make certain that this extra system supplements instead of dominates the mix.

    This post gives an notion about what I am about.

    Cheers

    C

  5. #5
    Thank you for your answer.

    I update my first post with more opinions on my view of the way I see the markets and trade.

    Quote Originally Posted by ;
    quote 1) yes, create three systems and trade with the same huge picture of the market. 2) yes, follow one easy trading methodology but exchange that in various timeframes - day trading, swing trading and position trading.
    I only develop systems for day trading. Does your remarks still apply for systems that are distinct if programmer on seek edges in day trading?


    Quote Originally Posted by ;
    3) yes, most important would be to keep current with all the changing market conditions which define the market kind to exchange.
    Simply to make sure I understand: 1) its important to keep up with and define market conditions even if day trading two ) equally important to match a trading platform with the ideal market state?

    Otherwise, trade one system and suffer drawdown once the border of the system doesn't match the market state that best match the system?

    Thank you for response.

  6. #6
    Thank you for responding to me.

    My first plan was to develop a trading system (1) that would produce income forever. But as I study and read more, develop and rear test more, I see, there's times when the edge produces money in rear test and times as it doesn't. And overall, program and back many (20 so far) intra-day systems from 2006-2011 (in sample) and none of them are profitable unless really over optimize all the parameters. I am just thinking, is it something I'm missing or not thinking clearly.

    I will respond to your comments below and appreciate your reply.

    Quote Originally Posted by ;
    No single system could be profitable across all market conditions; however a mix of uncorrelated systems has the capacity to address a package of market conditions
    what's ununcorrelated systems imply?

    Is a good example of uncorrelated systems is System A that conducts seeking to trade market trading in a range (counter trend). System B which runs seeking to trade trending (instrument tendencies in 1 direction throughout the afternoon ) market days. System C which runs seeking to trade momentum setups.


    Quote Originally Posted by ;
    it is possible to develop quite powerful systems which outperform when circumstances are favorable and manage capital stringently when circumstances are adverse....and therefore maintain each system flipped on all the time.
    I favor these kind of systems like I feel a system must run perform similar how it done in rear test. And the odds in my favor if I choose every trade. In the rear test, I must know when my system will do good and when it will perform awful. I rather take every trade.

    Quote Originally Posted by ;
    Classify your egies into two broad classes
    Could an illuion of two broad egories be: Systems for trending market conditions and systems for range bound market requirements?

    Thank you

  7. #7
    Quote Originally Posted by ;
    quote Yep mate I hear you. What is essential is the connection between how your system performs over a broad variety of different market requirements. Any kind of trending egy is likely to work well when the market is trending....but likely to flounder and input drawdown phase when market conditions are non-trending. Non trending markets can be broadly egorised into markets which do not diverge away from a mean but rather converge towards an expression. For instance what some call counter-trend egies may fall into the egory or people...
    Thank you so much for detailed reply,

    I'll respond tomorrow . Thank you

  8. #8
    Quote Originally Posted by ;
    quote In analyzing your system under an array of distinct market conditions....then the longer the test the better. There's no fixed number of assuredness. The best way to assess the strengths and weakness of your own system is to plot the way your equity curve (or system reunite distribution) works under different market conditions.
    ,

    Thank you so much for the effort and response and time and patience. Your comments really helped me.

    I am still studying your article in detail and understanding the full area of the article as it uncommon such considerate knowledge is exposed online. Thank you. I will respond with more questions or comments.

    I have a burning question cycling in my head for the past days, now that I know 1 system does not work for all market conditions and the key is know when your system is profitable and what market condition fit your system.

    My experience is only with Intraday (or day trading or positions closed through the day) trading. I'd like to only build systems for intraday trading. I study 30 to 5 minutes charts.

    With day trading their is a market condition changing frequently.

    Does your writings employ to designing and analyzing day or intraday systems as well for different market conditions? As an example, there are days when price breaks out, trade in range, trends in 1 direction, or even reversals.

    Thank you,

  9. #9
    Quote Originally Posted by ;
    quote, Thank you so much for your effort and answer and time and patience. Your comments helped me. I'm still studying your post in detail and understanding the entire extent of the post as it uncommon such thoughtful understanding is subjected online. Thank you. I will respond to comments or more questions. I have a burning question cycling in my mind for the previous days, now that I know 1 system does not function for all market conditions and also the key is know whenever your system is profitable and what market condition match your system. My experience...
    No probs G

    I've a thread devoted to your line of enquiry that gives info on the best way best to test and plan to get a systematic portfolio that manages the gamut of different potential market requirements. . .timeframes and systems deployed aren't critical to this subject, yet I do use a favorite divergent and convergent system at the excuse. The'secret sauce' is in how to combine your systems....not within their individual performance. The ribbon does justice to this notion. . .yet it is called EDTT (which most think is the important ingredient). If you are patient and have a read....hopefully you will get the gist.

    Do not worry about the long winded read at first. Start looking at articles about mid-way through the thread and outside that begin talking portfolio management and production. If it is possible to manage general nonsense such as music clips....then there is lot's of information in their own and also articles from the professional fund management distance on the topic.

    If you want some clarifiion, then just leave a post in this thread and I will do my best to describe.

    Cheers

    C

  10. #10
    Quote Originally Posted by ;
    quote No probs G I've a thread devoted to a line of enquiry that provides info on the best way to examine and plan for a systematic portfolio that handles the gamut of distinct possible market requirements. . .timeframes and systems deployed aren't essential to the subject, yet I do use a preferred divergent and convergent system in the explanation. The'secret sauce' is in how to combine your systems....not within their personal functionality. The thread does...
    Hello and invite you for invite to other thread.

    I will leave this thread now and go join the other thread for review and reading. I'll ask questions there after doing some reading if necessary.

    Thank you kindly.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
This website uses cookies
We use cookies to store session information to facilitate remembering your login information, to allow you to save website preferences, to personalise content and ads, to provide social media features and to analyse our traffic. We also share information about your use of our site with our social media, advertising and analytics partners.