Backtesting/Optimization - page 19
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Any EA which frequently closes trades with less than 20 pips of profit or loss is likely to show a very different result using the 90% modelling quality that comes from fractal interpolation.
The interpolation method made on everytick simulation (the most accurate) simulates a tick based on an entire minute. But anything can happen in a 1M candle.
The fractal interpolation algorithm only guesses by simply using the 1M Open/High/Low/Close data, and clearly this may not be anything like what really happened.
If your EA has any possibility of entering and exiting a trade within 1 minute, then fractal interpolation will give highly inaccurate results. And remember that 1M candles frequently have a range of 50 pips during news announcements.
Never rely on 90% backtest, unless you use them for EA logic tests, not performance tests.
Don't lie to yourself.
Accurate alternative to MT Backtester ?
Hi everyone,
I think Metatrader is a great tool and its brilliant to automate strategy to remove emotions...however the more I look into metatrader - and I've been studying it for a while now !! the less faith I have in the backtester - it seems that every new version or update of Metatrader gives different results and sometimes when the same test is rerun different results again !! I know forward testing is the most accurate way but that takes ages, surely with all that historic data out there there must be a way to accurately backtest... so does anyone have links to backtesters which can run experts and give accurate results ? I've tried testing offline too but am still not confident in backtests produced. IF accurate backtests could be achieved then we could develop much better backtests !! I have been using 90% quality.....
I think the Metatrader backtester engine is fine. The issue is data. Do you have quality tick data? If you have, then your backtest results should be pretty reliable. If you do not, still there is hope. If your system uses completed bars for entry and exit, the results from "quality" minute level data should be good.
Hope this helps.
Hi Maji, thanks for your input, yes I am using minute data (90% quality) from Alpari databank - so backtests are only accurate if refencing a closed bar ? that could explain some problems, I have tested offline data also. Just seems the backtester is buggy giving different results from time to time.
There are quite a few alternatives available. TradeStation, Wealthlab, and Amibroker being three that have very good reputations when it comes to backtesting. What I have asked several times and have yet to receive a satisfactory answer to is, if Metatraders backtester is not at fault, why has it such a poor reputation? The only other platform I have experience with is Wealthlab's who's backtesting is universally respected. It is however more suited to stocks and EOD timeframes and this appears to be where the bulk of the development has been concentrated. It can of course work with futures/forex at lower timeframes as well but personally, I found it 'messy'. Also, what is hard to reconcile is the almost complete lack of profitable stategies available, at least according to the MT backtester. Download the Wealthlab demo and you will find a wealth of profitable 'wealthscripts' freely available. Granted, these are mostly EOD stock strategies.... Anyway, something to think about.
90% is NOT RELIABLE
okay...
It has been proved over and over that 90% backtests are good only to make testers who didn't invest yet to feel comfortable.
The Noobie Saga (I did it, and think hundreds of beginners may still do):
1. control points backtest (since it is the standard): you get millions, but the reliability is ZERO. You feel the luckiest guy on earth and think you found the holy grail.
2. You discover control points is crap and then pass to everytick (interpolation): you work on the EA's parameters until make some millions. You feel okay again.
3. The modeling quality is low and then you discover you should make it get 90%. Depression comes back. You download alpari data or Metaquotes data and have the 90% backtest.
4. You think it is reliable and after multiple backtests you start a forward test or account all excited with the upcoming fortune. The results are CRAP. You lose money. Depression.
Conclusion:
The only way to have reliable results, is based on 99% backtests gained from the TICK DATA achieved from the Broker Server you will start your account with.
Join us on making 99% modelling quality a reality for all backtesters. Lets turn Potential EAs on REAL PROFITABLE AUTOMATED SYSTEMS!
http://www.cubesteak.net/mt4-multi-broker-tick-and-m1-data-project/
Join us on making 99% modelling quality a reality for all backtesters. Lets turn Potential EAs on REAL PROFITABLE AUTOMATED SYSTEMS! http://www.cubesteak.net/mt4-multi-broker-tick-and-m1-data-project/
I went to the website and noticed that there is no way to download any tick data. Where do you go to download the tick data to get 99% backtests?
Thanks.
Tick data testing
I know that we can get hold of tick data for couple of years from different sources but i dnt think we can import tick data to mt4 even if we make it fxt file. we have to convert that data to 1m and then do testing.
My question is -- is there a way that we can run statergy testing straight form tick data this way aur testing will be nearly 100% correct.
Tick data testing
I know that we can get hold of tick data for couple of years from different sources but i dnt think we can import tick data to mt4 even if we make it fxt file. we have to convert that data to 1m and then do testing.
My question is -- is there a way that we can run statergy testing straight form tick data this way aur testing will be nearly 100% correct.
I was wanting to know if there was a website that someone is upkeeping that is collecting Tick Data for Metatrader 4.
- - - -
Anybody know of a place that is allowing downloads and uploads of tick data?Hi,
Check out http://www.cubesteak.net/mt4-multi-broker-tick-and-m1-data-project/