Results 1 to 10 of 10
 2 Attachment(s)    

Thread: Funyoo. Do you really get 90% quallity?

  1. #1

    Default Funyoo. Do you really get 90% quallity?

    Hi Funyoo.

    I need to ask you if you are 100% sure that you get 90% quality when backtesting? The reason i ask is that when downloading history from metaquotes which is the case for most brokers, including Alpari UK, you will recive huge gaps in the data. The most recent gap is on EURUSD between 2010 05 07 - 2010 06 24.

    I have attached 2 pictures. Both on XBars. I have used the 1.03 version and the set file you provided in the first post.

    The first picture shows backtesting on Alpari 2007-01-01 to 2010-06-25. Here i have even asked Alpari to provide me with the missing tick data, and they did.
    A great result
    How ever... Alpari and every other broker for that matter isn't going to sit down and look for the gap for you. I you will most defently not notice every missing tick.

    The second picture shows the exact same period of testing and set file. Only here i have recived tick data from a friend of mine that have downloaded it from his broker which provide this data for it's customers.

    Huge difference.

    And yes i am aware of that the result will differ from broker to broker. But hardly not with over 100 000 USD.

    So my question to you Funyoo is: What tick data do you use and how reliable is it?

    Regards / JoLi
    Attached Images Attached Images Funyoo. Do you really get 90% quallity?-xbars_alpari_ticks-gif Funyoo. Do you really get 90% quallity?-xbars_correct_tick-gif 
    Regards / JoLi

  2. #2

    Default

    hi Joli,
    if you want 90% modeling quality with Alpari, go to Alpari russia in the weekend. Download the mt4 platform open an account, login and go to history centre. Download whatever you want. Do not use Alpari uk, for then you get the standard datadownload you mention.
    Then test an ea. You will have without any doubt 90% mod. quality. Do not open this mt4 platform during the week, but only on sunday to update with the history centre and download again whatever pair but always use 1min. to download. If you want to test on another timeframe then use the script period converter to convert 1min data to whatever timeframe you want.
    That's it.

  3. #3

    Default

    Quote Originally Posted by roodstaart View Post
    hi Joli,
    if you want 90% modeling quality with Alpari, go to Alpari russia in the weekend. Download the mt4 platform open an account, login and go to history centre. Download whatever you want. Do not use Alpari uk, for then you get the standard datadownload you mention.
    Then test an ea. You will have without any doubt 90% mod. quality. Do not open this mt4 platform during the week, but only on sunday to update with the history centre and download again whatever pair but always use 1min. to download. If you want to test on another timeframe then use the script period converter to convert 1min data to whatever timeframe you want.
    That's it.
    Been there... done that.

    try it yourself. You will have aprox. the same result as pic.#2. At least that's what i get when using Alpari RU. I have tryed it on Alpari RU both offline and online... i have imported Alpari Ru's history to Alpari UK and tryed it on UK both offline and online. No difference to mention.
    As said... try it your self and you will see.

    So... of course i can agree that Alpari RU have more reliable tick data when downloading. But what does that mean then? Does it mean that evey Expert Funyoo have optimiazed is bad? Well i don't know... i just know that when using Alpari RU (or the one i describe in post# 1) i have approx. 40% plus from 2007-01-01. Meaning that an account started with 10k ends on 14k and not 118k.

    Regards / JoLi
    Last edited by jolimoney; 06-28-2010 at 21:04.
    Regards / JoLi

  4. #4

    Default

    Quote Originally Posted by jolimoney View Post
    Been there... done that.

    try it yourself. You will have aprox. the same result as pic.#2. At least that's what i get when using Alpari RU. I have tryed it on Alpari RU both offline and online... i have imported Alpari Ru's history to Alpari UK and tryed it on UK both offline and online. No difference to mention.
    As said... try it your self and you will see.

    So... of course i can agree that Alpari RU have more reliable tick data when downloading. But what does that mean then? Does it mean that evey Expert Funyoo have optimiazed is bad? Well i don't know... i just know that when using Alpari RU (or the one i describe in post# 1) i have approx. 40% plus from 2007-01-01. Meaning that an account started with 10k ends on 14k and not 118k.

    Regards / JoLi
    I get your point Joi, but I myself have stopped using backtesting. I only do forewardtesting, for backtesting, even on 90% m.q. tells me nothing for one can optimize for a certain period and then let it run a backtest on that same period. The result only tells that on that specific period how the ea should have functioned. It tells nothing about the future. So a good ea has to prove itself on livetest, that is what I think.

  5. #5

    Default

    Quote Originally Posted by roodstaart View Post
    I get your point Joi, but I myself have stopped using backtesting. I only do forewardtesting, for backtesting, even on 90% m.q. tells me nothing for one can optimize for a certain period and then let it run a backtest on that same period. The result only tells that on that specific period how the ea should have functioned. It tells nothing about the future. So a good ea has to prove itself on livetest, that is what I think.
    In fact i agree. Backtesting only shows what could have been.... nothing about what can be.
    But backtesting get's a view of if the system has something to offer or not. If it's worth fwd-testing.
    As a TA-trader i rely on that the history will repeat itself. At least i hope so.... otherwise i need to become a FA-trader and i don't like it.

    The problem remains however.... how to really get 90% m.q? Do Funyoo even get it? If he only use Alpari UK i doubt it.

    Regards / JoLI
    Regards / JoLi

  6. #6

    Default

    With some tools it is possible to get 98 - 99% m.q.
    You need to catch all ticks and convert that to the timeframe you wish.
    Here is a possibility: Parnu Rivertrading, Parnu River Trading Inc.

  7. #7

    Default

    Quote Originally Posted by roodstaart View Post
    With some tools it is possible to get 98 - 99% m.q.
    You need to catch all ticks and convert that to the timeframe you wish.
    Here is a possibility: Parnu Rivertrading, Parnu River Trading Inc.
    Tnx for the advice. Perheps i will try it later. But my maine task right now is to get 90% m.q first. No gaps etc.

    Regards / JoLi
    Regards / JoLi

  8. #8

    Default

    Quote Originally Posted by jolimoney View Post
    Tnx for the advice. Perheps i will try it later. But my maine task right now is to get 90% m.q first. No gaps etc.

    Regards / JoLi
    Hi JoLi,

    Here is a free solution that solves most common issues with backtesting quality.

  9. #9
    Administrator funyoo's Avatar
    Join Date
    Sep 2008
    Posts
    7,003

    Default

    Quote Originally Posted by jolimoney View Post
    Hi Funyoo.

    I need to ask you if you are 100% sure that you get 90% quality when backtesting? The reason i ask is that when downloading history from metaquotes which is the case for most brokers, including Alpari UK, you will recive huge gaps in the data. The most recent gap is on EURUSD between 2010 05 07 - 2010 06 24.

    I have attached 2 pictures. Both on XBars. I have used the 1.03 version and the set file you provided in the first post.

    The first picture shows backtesting on Alpari 2007-01-01 to 2010-06-25. Here i have even asked Alpari to provide me with the missing tick data, and they did.
    A great result
    How ever... Alpari and every other broker for that matter isn't going to sit down and look for the gap for you. I you will most defently not notice every missing tick.

    The second picture shows the exact same period of testing and set file. Only here i have recived tick data from a friend of mine that have downloaded it from his broker which provide this data for it's customers.

    Huge difference.

    And yes i am aware of that the result will differ from broker to broker. But hardly not with over 100 000 USD.

    So my question to you Funyoo is: What tick data do you use and how reliable is it?

    Regards / JoLi
    Hi jolimoney,

    I use metaquotes history. You are right, there is a big gap between 2010 05 07 - 2010 06 24. Currently I only optimize up to 2010 05 06. They will probably resolve this in the future. Well, backtesting is relative, as you say it only shows a potential. What you get with metaquotes should theoretically be reproduced with each broker's tick data. Unfortunately, the brokers don't offer the possibility to download their history as in the past with Alpari UK. Yes you get a real 90% of modelling quality as long as you don't do your test over this gap.

    Note that I have stopped to work with Alpari UK, I have only got a lot of disconnection issues with them during the last weeks in demo and real... it was a nightmare. I have also lost a lot of time with my demos, they began to disable them after few days and I had hundreds of demo accounts... They probably want to do some savings with their servers. I will continue with InstaForex, they have fixed spreads, their minilot is equal to nanolot, you can do more diversification, the duration of their demo account has no limit, they are serious in real mode (good tick data, no disconnection, good support).

  10. #10

    Default

    Quote Originally Posted by funyoo View Post
    Hi jolimoney,

    I use metaquotes history. You are right, there is a big gap between 2010 05 07 - 2010 06 24. Currently I only optimize up to 2010 05 06. They will probably resolve this in the future. Well, backtesting is relative, as you say it only shows a potential. What you get with metaquotes should theoretically be reproduced with each broker's tick data. Unfortunately, the brokers don't offer the possibility to download their history as in the past with Alpari UK. Yes you get a real 90% of modelling quality as long as you don't do your test over this gap.

    Note that I have stopped to work with Alpari UK, I have only got a lot of disconnection issues with them during the last weeks in demo and real... it was a nightmare. I have also lost a lot of time with my demos, they began to disable them after few days and I had hundreds of demo accounts... They probably want to do some savings with their servers. I will continue with InstaForex, they have fixed spreads, their minilot is equal to nanolot, you can do more diversification, the duration of their demo account has no limit, they are serious in real mode (good tick data, no disconnection, good support).
    Hi Funyoo.

    That's not the only gap. I have also found gaps at this periods:

    20100312 to 20100325
    20100507 to 20100610

    What do you think about Alpari RU then? It doesn't seem as they have this gap problem.

    Regards / JoLi
    Regards / JoLi

Similar Threads

  1. Funyoo HELP
    By paciola in forum MQL programming
    Replies: 0
    Last Post: 06-03-2009, 16:36
  2. Funyoo Please help! 5 34 5 EA!
    By generous forex in forum Ideas for expert advisors
    Replies: 5
    Last Post: 05-19-2009, 18:19
  3. hi funyoo
    By magidkaramy in forum Ideas for expert advisors
    Replies: 0
    Last Post: 02-03-2009, 19:18

Tags for this Thread

100, alpari, alpari uk, backtesting, broker, brokers, demo, eurusd, expert, free, funyoo, how to, instaforex, live, livetest, mt4, offer, online, real, support, system, systems, test, time, trader, xbars

Posting Permissions

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