Page 10 of 14 FirstFirst ... 89101112 ... LastLast
Results 91 to 100 of 138
  54 54 Attachment(s)    

Thread: Stat_Euclidean_Metric

  1. #91
    Junior Member
    Join Date
    Dec 2008
    Posts
    6

    Default problem with historical trades

    Hi,

    I am new int htis thread and wanted to understand how this EA works. So I had a look at the code and I am in doubt wether there is a big problem/missunderstanding in the deinit - function:

    - this function walks through complete history and uses these trades to calculate some MAs (and writing them to file). In some other forum, they told, that during backtesting the historical trades are the "real" historical trades of the account you are backtesting with. If so, you would get very strange and unpredictable results in your .dat-files. I also tested this EA with same settings and same currencies as other people and got totally different results !

    What do others know about this ?

  2. #92
    Junior Member Forexpal's Avatar
    Join Date
    Mar 2009
    Posts
    24

    Smile To obtain similar results

    Quote Originally Posted by JStein View Post
    Hi,

    I am new int htis thread and wanted to understand how this EA works. So I had a look at the code and I am in doubt wether there is a big problem/missunderstanding in the deinit - function:

    - this function walks through complete history and uses these trades to calculate some MAs (and writing them to file). In some other forum, they told, that during backtesting the historical trades are the "real" historical trades of the account you are backtesting with. If so, you would get very strange and unpredictable results in your .dat-files. I also tested this EA with same settings and same currencies as other people and got totally different results !

    What do others know about this ?
    To obtain similar results, it is first needed to set the “base” to “true”, run a simulation, and then set to false again.

    EA needs live demo testing to confirm, as said in recent posts.

  3. #93
    Junior Member
    Join Date
    Dec 2008
    Posts
    6

    Default

    Quote Originally Posted by Forexpal View Post
    To obtain similar results, it is first needed to set the “base” to “true”, run a simulation, and then set to false again.

    EA needs live demo testing to confirm, as said in recent posts.
    Yes, I did it just this way but I think its the wrong way because of the mentioned problem with historical trades.

    what do you think about this ?

  4. #94
    Junior Member Forexpal's Avatar
    Join Date
    Mar 2009
    Posts
    24

    Default

    The potential from the EA comes from the opening of many trades that happen to be on the correct side; a fact confirmed in back and forward testing.

    The base feature and the Euclidean metrics logic, are quite sophisticated. Hard to figure out what EA does by just looking at the code.

    Forward testing will anyway confirm how it behaves in time. Trying all the settings, and then comparing them.

  5. #95
    Junior Member
    Join Date
    Dec 2008
    Posts
    6

    Default

    Quote Originally Posted by Forexpal View Post
    The potential from the EA comes from the opening of many trades that happen to be on the correct side; a fact confirmed in back and forward testing.

    The base feature and the Euclidean metrics logic, are quite sophisticated. Hard to figure out what EA does by just looking at the code.

    Forward testing will anyway confirm how it behaves in time. Trying all the settings, and then comparing them.

    Mhhh ... I think, the idea with this EA is, to train it by backtesting over some period of time, store the training results in *.dat files and use it for trading decisions in live trading. But if the training results are not correct and consistant there will be no stable live trading result.

    Funnyoo, what do you think about this way (using historical trades in backtest) to train the EA ?? I did some
    tests with all the same settings and only changing the "account history" in my MT4-Terminal (by change the custom period so that it shows different results) and i got different results in the live test ( setting base to false ) !!

  6. #96
    Administrator funyoo's Avatar
    Join Date
    Sep 2008
    Posts
    7,003

    Default

    Quote Originally Posted by JStein View Post
    Funnyoo, what do you think about this way (using historical trades in backtest) to train the EA ?? I did some
    tests with all the same settings and only changing the "account history" in my MT4-Terminal (by change the custom period so that it shows different results) and i got different results in the live test ( setting base to false ) !!
    Hi JStein,

    As Forexpal said, you need an account with some trades history with losses otherwise you will not get interesting results.

    I have only made few tests but I have not yet applied the EA on an account with such conditions.

  7. #97
    Member
    Join Date
    Mar 2009
    Location
    Florida, USA
    Posts
    278

    Default

    Here are my back testing results. Not good at all.
    Attached Files Attached Files
    Think of adversity as that tiny piece of sand that gets trapped in the oyster's shell. Although it's a pain in the oyster's arse, he eventually makes a pearl out of it.

    http://hft.serialcoder.com/ -- Quantitative Market Reversal Levels --- The only indicator you'll ever need.

  8. #98
    Junior Member
    Join Date
    Dec 2008
    Posts
    4

    Default Stat_Euclidean_Metric_EA.

    I attach the statements of almost one month of this particular EA that, indeed, after a negative departure, has not only recovered but has gotten notable results.

    For how much some trades have manually been closed to avoid regression, the EA, in the version "Stat Euclidean-No Martingales (Magic + Trailing). ex4" it seems to work in the sense that operates not in the sense that is without of risks while I have not been succeeding in having only a trade neither from the version "Stat_Euclidean_Metric (Trailing Step). ex4" neither from the version "Stat Euclidean-no Martingale2 (Magic+ Trailing +ADX). ex4." that perhaps surer are.

    Which can motive be?

    For how much I make them turn on the same base they have magic different numers however.

    I also attach the set of the attached statement specifying that I have used only three crosses: GBPJPY, GBPUSD AND USDJPY.

    Thank you and best regards.

    Frankalde
    Attached Images Attached Images Stat_Euclidean_Metric-detailedstatement_01mag09-gif 
    Attached Files Attached Files

  9. #99
    Junior Member
    Join Date
    Feb 2009
    Posts
    25

    Default

    I tried another timeframe and the result is not great like 1H timeframe. Is that meaning the expert dose not cheating in historical Data and it use a real Strategy on the backtest?
    because if this expert cheating in historical data the result of other timeframes and currencies should be great too(It will records orders lost and won then improve the backtest)

  10. #100
    Junior Member
    Join Date
    Dec 2008
    Posts
    6

    Default more than one currency

    Quote Originally Posted by Frankalde View Post
    I attach the statements of almost one month of this particular EA that, indeed, after a negative departure, has not only recovered but has gotten notable results.

    For how much some trades have manually been closed to avoid regression, the EA, in the version "Stat Euclidean-No Martingales (Magic + Trailing). ex4" it seems to work in the sense that operates not in the sense that is without of risks while I have not been succeeding in having only a trade neither from the version "Stat_Euclidean_Metric (Trailing Step). ex4" neither from the version "Stat Euclidean-no Martingale2 (Magic+ Trailing +ADX). ex4." that perhaps surer are.

    Which can motive be?

    For how much I make them turn on the same base they have magic different numers however.

    I also attach the set of the attached statement specifying that I have used only three crosses: GBPJPY, GBPUSD AND USDJPY.

    Thank you and best regards.

    Frankalde
    Hi Frankalde,
    you used EA with multiple currencies at the same time. EA is not build to do this because it uses the same *.dat files. But these files are different form currency to currency. You should change the EA so that filenames contain the currency name (for example Buy_PositionsGBPJPY.dat and so on).

Similar Threads

  1. Stat_Euclidean_Metric
    By funyoo in forum Expert advisors live statements
    Replies: 2
    Last Post: 01-31-2009, 19:22

Tags for this Thread

100, 2011, advisor, adx, alpari, alpari uk, average, backtest, backtesting, breakout, broker, brokers, change, channel, check it out, closing, coding, color, copy, demo, donchian, ea works, eas, eurusd, ex4, expert, expert advisor, experts, forex, free, fxopen, gbpjpy, gbpusd, hedge, hedging, historical, how to, indicator, indicators, learn, live, live account, macd, magic, magic number, manual, martingale, metatrader, mq4, mql, mql4, mt4, offer, online, ordersend error 4051, profit, profitable, programmer, programming, real, request, robot, security, sells, server, signals, simple, step by step, stop, stoploss, strategy, system, systems, test, thank you, time, trader, trading, trailing, training, 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
  •