Results 1 to 4 of 4
 0 Attachment(s)    

Thread: Market activity change, EA additive

  1. #1
    Junior Member
    Join Date
    Jan 2009
    Posts
    19

    Default Market activity change, EA additive

    Hi Funyoo

    As you know, all EA need every year new setings, because market change. What you think, is it possible make (indicator maybe?), what check market movement every day, how much and how fast price change his location and how many times hi change direction (zig-zag?) and trend, and base on this data, give us some number. Every month or quarter of a year or year, hi add together all this numbers give us percent market activity (0-100%). Then EA can, base on this precent, change setings by itself, of course setings we give to EA, as an example: 0-10% SL 100, TP 20, indicator setings 5, 61, 20 ; 10-20% SL 55, TP 10 ... and so on.


    We have market history and history go around. 2009 come difference 2008, maybe like 2002. With this system we can maybe defin future.
    Sorry, bad english

  2. #2
    Administrator funyoo's Avatar
    Join Date
    Sep 2008
    Posts
    7,003

    Default

    Quote Originally Posted by viplala View Post
    Hi Funyoo

    As you know, all EA need every year new setings, because market change. What you think, is it possible make (indicator maybe?), what check market movement every day, how much and how fast price change his location and how many times hi change direction (zig-zag?) and trend, and base on this data, give us some number. Every month or quarter of a year or year, hi add together all this numbers give us percent market activity (0-100%). Then EA can, base on this precent, change setings by itself, of course setings we give to EA, as an example: 0-10% SL 100, TP 20, indicator setings 5, 61, 20 ; 10-20% SL 55, TP 10 ... and so on.


    We have market history and history go around. 2009 come difference 2008, maybe like 2002. With this system we can maybe defin future.
    Hi viplala,

    The daily ATR suggested by ixbone is a good way to dertermine if the market is trending or ranging. Then depending on it, you can propose different parameters to the code.

  3. #3
    Junior Member
    Join Date
    Jan 2009
    Posts
    19

    Default

    Hi,

    This is an important line of inquiry.

    One approach has been to use self-adaptive indicators which change their behaviour based on market conditions. TPs & SLs based on ATR for example, expand & contract based on recent range. This behavior allows them to automatically adapt to changes in price action, where fixed TPs & SLs do not. The Kaufman Adaptive Moving Average (KAMA) and Adaptive RSI (ARSI) and Chande's CDMA are some other examples.
    VIDYA/CDMA by Tushar Chande Tradestation Indicators Forex Trading Systems - XEATrade - Metatrader, Metastock, Tradestation, Amibroker - indicators, expert

    KAMA & ARSI use "Kaufman's Efficiency Ratio" to speed up or slow down the MA & RSI. Chande's CDMA uses ADX.

    Mark Jurik used his ADX-like Composite Fractal Behavior (CFB) indicator in a similar way.
    "CFB (Composite Fractal Behavior, Jurik Research) is an index that reveals the market's trending time frame, ideal for creating adaptive window sizes of various technical indicators." Mark Jurik Research Technical Indicators: JMA, VEL, CFB, RSX

    CFB

    Here's a screenshot of an adaptive RSI that uses "r-squared" as the adaptor:
    cartViewItem — Bowfort Technologies Inc.

    I think FRASMA may be considered self adaptive also, as it uses a measure of Fractal Dimension to adaptively control a moving average:
    "The SMA is accelerated during a trend and slowed down during a sideways market, so as to avoid false signals. It's using the fractal dimension as computed by ... fractal_dimension.mq4, and makes use of it to smooth the SMA." FRASMA: Fractally Modified Simple Moving Average - MQL4 Code Base

    ~

    Self adaptive indicators are not the complete ultimate solution however; for example:
    "There are a number of popular indicators that use range in their calculation. One of the problems with using range based indicators is that the optimum lookback period seems to change over time causing this type of indicator to eventually give false signals over that fixed lookback period."
    "To try and avoid the false signals generated by a single lookback period we create a range indicator that uses a variable lookback period."

    The Maximum Likelihood Range System

    ~

    Another approach is to use optimization fairly frequently to determine what the optimum settings of EA input parameters would have been in recent price action. Then use the discovered parameters in the EA until performance drops off again. The EA is periodically 'retuned' to fit recent price action.
    Some debate has raged in the past about this as some (many) kinds of trading systems are susceptible to "overfitting" recent price-action data...that is, they can trade the past perfectly after being tuned by the optimizer, but they may not be able to continue working well on unseen data.
    On the other hand, advocates of this approach say that some kinds of systems are "robust" enough to benefit from this retuning, and that changing market dynamics demand periodic retuning.

    Some interesting discussion about some of this and more here:
    Power Walk Forward Optimizer
    Walk Forward Performance Explorer
    Walk Forward Performance Metric Explorer
    Surface Walk Forward Performance Explorer
    http://www.meyersanalytics.com/publications/optsys.pdf
    http://www.meyersanalytics.com/publications/tricked.pdf

    ~

    There are other approaches as well...such as slicing historical price action data into 'seasonal' batches and looking for commonalities among the batches. This can be done at various resolutions, monthly, trading day of the month, trading day of the week, time of day, etc.
    Last edited by AlisNutrior; 03-11-2009 at 08:55.

  4. #4
    Member
    Join Date
    Jan 2009
    Location
    Austria
    Posts
    325

    Default

    Quote Originally Posted by AlisNutrior View Post
    Hi,

    This is an important line of inquiry.

    One approach has been to use self-adaptive indicators which change their behaviour based on market conditions. TPs & SLs based on ATR for example, expand & contract based on recent range. This behavior allows them to automatically adapt to changes in price action, where fixed TPs & SLs do not. The Kaufman Adaptive Moving Average (KAMA) and Adaptive RSI (ARSI) and Chande's CDMA are some other examples.
    VIDYA/CDMA by Tushar Chande Tradestation Indicators Forex Trading Systems - XEATrade - Metatrader, Metastock, Tradestation, Amibroker - indicators, expert

    KAMA & ARSI use "Kaufman's Efficiency Ratio" to speed up or slow down the MA & RSI. Chande's CDMA uses ADX.

    Mark Jurik used his ADX-like Composite Fractal Behavior (CFB) indicator in a similar way.
    "CFB (Composite Fractal Behavior, Jurik Research) is an index that reveals the market's trending time frame, ideal for creating adaptive window sizes of various technical indicators." Mark Jurik Research Technical Indicators: JMA, VEL, CFB, RSX

    CFB

    Here's a screenshot of an adaptive RSI that uses "r-squared" as the adaptor:
    cartViewItem — Bowfort Technologies Inc.

    I think FRASMA may be considered self adaptive also, as it uses a measure of Fractal Dimension to adaptively control a moving average:
    "The SMA is accelerated during a trend and slowed down during a sideways market, so as to avoid false signals. It's using the fractal dimension as computed by ... fractal_dimension.mq4, and makes use of it to smooth the SMA." FRASMA: Fractally Modified Simple Moving Average - MQL4 Code Base

    ~

    Self adaptive indicators are not the complete ultimate solution however; for example:
    "There are a number of popular indicators that use range in their calculation. One of the problems with using range based indicators is that the optimum lookback period seems to change over time causing this type of indicator to eventually give false signals over that fixed lookback period."
    "To try and avoid the false signals generated by a single lookback period we create a range indicator that uses a variable lookback period."

    The Maximum Likelihood Range System

    ~

    Another approach is to use optimization fairly frequently to determine what the optimum settings of EA input parameters would have been in recent price action. Then use the discovered parameters in the EA until performance drops off again. The EA is periodically 'retuned' to fit recent price action.
    Some debate has raged in the past about this as some (many) kinds of trading systems are susceptible to "overfitting" recent price-action data...that is, they can trade the past perfectly after being tuned by the optimizer, but they may not be able to continue working well on unseen data.
    On the other hand, advocates of this approach say that some kinds of systems are "robust" enough to benefit from this retuning, and that changing market dynamics demand periodic retuning.

    Some interesting discussion about some of this and more here:
    Power Walk Forward Optimizer
    Walk Forward Performance Explorer
    Walk Forward Performance Metric Explorer
    Surface Walk Forward Performance Explorer
    http://www.meyersanalytics.com/publications/optsys.pdf
    http://www.meyersanalytics.com/publications/tricked.pdf

    ~

    There are other approaches as well...such as slicing historical price action data into 'seasonal' batches and looking for commonalities among the batches. This can be done at various resolutions, monthly, trading day of the month, trading day of the week, time of day, etc.
    interesting statement! but for advanced traders only.

    1st) fixed sl tp secures your account, no unwanted gaps can hit your account, especially in trending markets
    2nd) for beginners, its better to become a feeling what is the market doing see 1st.
    3rd) if you know the daily average of a pair you can set always correct fixed sl tp values, example: gbp/chf daily range 500pips, the market is currently at 250 pips, in the middle (room 250 up and 250 down), set the sl to 350 and tp to 200, you catch your profits and be secure with sl (example depends on your tf!) using ts is a option too.
    4th) no indi can give your system adaptive and correct! sl+tp values in real time, they lag always behind, because they calculate on the past-this is a must

    just my personal opinion

    IX

Similar Threads

  1. How to change the swappiness ?
    By fxtester in forum General discussion
    Replies: 2
    Last Post: 04-14-2015, 08:56
  2. req[to change the indiactor into ea]
    By ananthgeethan in forum Ideas for expert advisors
    Replies: 0
    Last Post: 06-23-2012, 12:57
  3. Change direction EA
    By daniel22 in forum Ideas for expert advisors
    Replies: 0
    Last Post: 09-14-2011, 09:44
  4. Help me to change spred in this code
    By aly2000 in forum MQL programming
    Replies: 0
    Last Post: 06-15-2009, 16:00

Tags for this Thread

10%, 100, add, adx, average, change, change direction, expert, forex, forex trading, forex trading systems, historical, index, indicator, indicators, jurik, metatrader, moving average, mq4, mql, price action, real, research, rsi, signals, simple, system, systems, time, trading, trend

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •