My backtests are showing losses of over 5000 pips.Is there preset ea somewhere which matches published performance of machannel on euro/usd and usd/chf
This EA was coded on high/low of the bar and backtesting will not match the forward testing in this case. Besides, the brokers' data are too different concerning high/low of the bar. And because of that the performance of this EA will be different with different brokers.
It is something about this EA coded and different brokers data.
Only forward testing is reliable with this EA. It is not a bug. This EA was coded on hgh/low of the bar and it is this case only.
If some members want to use this EA with different brokers so it is necessary to forward test it with default settings comparing the trades with my trades just to correct the settings.
For example: if you want to use this Ea with IBFX so you need to forward test it with IBFX (default settings, H1 timeframe) and compare the trades with my North Finance trades and after that correct yoyr IBFX settings. It is the only way to use such kind of EAs which were coded on high/low of the bar.
We had one case with this MaChannel EA.
I am testing it with North Finance broker with very very good results. And Codersguru (author) used this EA with last Metaquotes contest. Metaquotes had IBFX data in that time for the contest. So,I am demoing with North Finance for elite section, and Codersguru demoing with Metaquotes/IBFX for the conest. My results were great and very profitable and Codersguru demoing was not profitable and was completely different.
It is only because this EA was coded on high/low of the close bar.
The other EA which can not be backtested or optimized by backtesting is Firebird: it was coded on close bar, open bar and high/low of close bar simultaniously and because of that no one backtesting is reliable with this EA.
The other EA which can not be backtested is Brainwashing all the versons: the level of iTrend indicator is very different for differen brokers and fully depends on the broker's data. Only forward testing can show something. Or trading manually Brainwashing is manual trading system so it is easy to select the setting manually by indicators).
The other one is DayTrading3. No any backtesting is reliable. Because of some specific coding. It is not bug. It was done especially for the better performance but sorry by forward testing only.
The other ones are MTF EAs almost all the versions.
I don't like to backtest EAs. The settings should be suggested by author of EA, or we may select it from same manual trading system in most cases.
I am backtesting EA sometimes just to see that EA works with no error and to have some idea about how it works.
But in some cases backtesting is useless.
In previous post I mentioned about 5 cases (EAs) which we don't need to backtest as backtesting will not match forward testing.
Some EAs which we may backtest to find the better settings:
- SBS all the versions;
- 20pips EA;
- Scalp_net all the versions;
- e-Friday and e-Monday;
and some others.
Some EAs will have very different performance with different brokers:
- all Step versions;
- all Scalp_net versions;
- all MTF EAs (anyone in elite section and public forum).
- Newstrader all the versions.
To use this EAs with different brokers the members need to backtest it with their brokers (not Alpari data), and forward test if comparing the trades with my trades just to correct the settings for their particular broker.
Same case with SBS and TPE all the vertions (comparing the trades to correct the settings) but those cases with SBS and TPE are different: it is something about the time to open pending orders (it is nothing with data feed; it is just to set the time corectly for different brokers).
It is related to all EAs in the world but many people simple don't mention those cases thinking that EA itself is "unificated robot" and will work with every broker with same settings but practically it is not.
For example: Daytrading 3 EA is not working with IBFX broker with any settings (will not be profitable). Because of spikes in data for IBFX. And this EA may work with Alpari broker. But to know the performance it is necessary to forward test it ebacsue we can not make any conclusion using backtesting for this EA (Dayrading3). That is why i am forward testing some EAs for many months here.
no more contest?
Could be interesting a big competition like Metatrader do, but with the spirit of open souce as is this forum.
Is possible to add some rules to make possible also some profit for the winner, like a small subscription just to see the result going.
Small money, many peolple...
What do you thin about newdigital?
Contest should be well prepared. And contest should be big one. Many people should participate.
Otherwise I will win once again with Scalp_net_v1.5 (together with Codersguru and Igorad)
It was some public thread where coders and other people discussed the rules so the main rules were made by members and it is no problem with anything.
Just problem with preparation and because of that we do not have contest yet.
Thank you for the promt answer.
I will be your supporter!
Waiting for news
Newdigital, could help me to find i-balEq indicator or someting like this?
Thank you inadvance for yor help
It is here https://www.mql5.com/en/forum/173442
And some other versions are here https://www.mql5.com/en/forum/177132
Scalp_net_v1.5 the best?
You could be the winner ... I see Scalp_net_v1.5 seems to be one of the best EA in this forum, Profitable stable and not so risky.
I am spending a lot of time to read many and many posts, I have still thousand, but up to now I don't find any like Scalp_net_v1.5.
Have you some info regarding improvements of this EA?
Any your comment is wellcome
I created this EA for our contest (and Igorad helped me to add some options to this EA). And it was the time when we had just few EAs in elite section so I had to code ...
It was two contests.
First contest: 1st place was for GridMACD by pengie, second and 3rd - Codersguru and Igorad.
Second contest: I won 3rd place in the contest. Codersguru won 1st place with MaChannel and Igorad - second place with StepMaExpert_v1.42.
I may be mistaken about who won 2nd and 3rd place because it was long time ago.
It was 4 or 5 participants totally.
Then Codersguru participated with his MaChannel EA for Metaquotes contest. And Igorad improved his 1.42 version up to 1.45 version and then up to StepPortfolio EA. But suddenly Metaquotes decided to use data from IBFX broker for last Metaquotes contest but their EAs were optimized using other data so ...
That is why I said on discussion thread about EAs settings for IBFX data: the settings for IBFX may be different than I posted on files thread.
Scalp_net_v1.5 is not the best. It may be good EA but it is not the best one here.
ND Thank you for info
I am usig InterbankFX and also MIG.
Unfortunatelly MIG has no mini accounts.
I am forward testing many EA from this forum on demo, and now I am starting to put also some EA in real mode (with microlots...!) as I understand datas are different between broker and also between real and demo mode...
I am not so old in trading forex and I am working to make my experience.
I test many martingala based EA, but I understand they havo too many risk inside: big loss or big drowdown, if you are luky. Pheraps exists an maringala based EA with a sort of hedging to prevent these situation, but I still not found them. Reserch continue.
In any case I am looking, styding and working to build a portfolio of EAs working continuosly with reasonable profit and small risk, I am not looking for "magic" EA.
Your help is very precious.
I am not English mother language, as you see, I hope all will be understandable.