Rio Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hrm... I'm not sure ab out using FXDD data. Their live spreads are typically fairly high. Quote Link to comment Share on other sites More sharing options...
hyperdimension Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hrm... I'm not sure ab out using FXDD data. Their live spreads are typically fairly high. That's not relevant. Spreads are stored in your MetaTrader terminal, not in the historical data. The historical data in MetaTrader only contains date & time, open, high, low, close Bid prices and volume. There are no Ask prices or spread values in the data. So the spread that is used will be broker and time-of-day dependent. This is a very important thing to note, epsecially with Alpari (US) because they now have floating spreads. So if you are backtesting over a connected MetaTrader terminal, the Strategy Tester will use the spread at that time. It's a really stupid design, but that's just how it is and we don't currently have any way to change how that works. So if you are testing during an active trading session, you may be testing using a low spread value, so your results are not indicative of the performance you'd get when it actually trades in the proper time window. The same issue exists for other brokers (most brokers?) whose spread widens during the time when MegaDroid is meant to trade. The way you should be testing is to use an offline terminal. You can make an offline terminal by waiting until its within the proper time window during the day (or night), then delete your demo account. The spreads will then stay constant in the terminal regardless of the time of day that you run your tests. You should then note down the spread values of each symbol for your reference. So I think you need to discard your test results so far and start over. Quote Link to comment Share on other sites More sharing options...
Rio Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Thanks for the tip. I'll try that out with the broker I'm actually using Quote Link to comment Share on other sites More sharing options...
Hyglater Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Guys, i'm backtesting ee2x version and results are very suspicious...At mbt and fxcbs (both ecn) the results are horrible. At market makers the result is quite impressive...any thoughts why? :-/ Quote Link to comment Share on other sites More sharing options...
to_ul Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro i still get the stack error on alpari uk. however its running at liteforex and tadawulfx without problems. it appears without backtesting or stuff, just having attached mdpro on all 4 pairs its supporting. so without using eur/usd logic twice at the same time. i dont think the multiple dll is a good alternative cause thats messing the risk settings. :-S Quote Link to comment Share on other sites More sharing options...
fabregas Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro That's the very good point. Anyway, the backtest i did is far from reality, so it was just fun :) Thanks! Hrm... I'm not sure ab out using FXDD data. Their live spreads are typically fairly high. That's not relevant. Spreads are stored in your MetaTrader terminal, not in the historical data. The historical data in MetaTrader only contains date & time, open, high, low, close Bid prices and volume. There are no Ask prices or spread values in the data. So the spread that is used will be broker and time-of-day dependent. This is a very important thing to note, epsecially with Alpari (US) because they now have floating spreads. So if you are backtesting over a connected MetaTrader terminal, the Strategy Tester will use the spread at that time. It's a really stupid design, but that's just how it is and we don't currently have any way to change how that works. So if you are testing during an active trading session, you may be testing using a low spread value, so your results are not indicative of the performance you'd get when it actually trades in the proper time window. The same issue exists for other brokers (most brokers?) whose spread widens during the time when MegaDroid is meant to trade. The way you should be testing is to use an offline terminal. You can make an offline terminal by waiting until its within the proper time window during the day (or night), then delete your demo account. The spreads will then stay constant in the terminal regardless of the time of day that you run your tests. You should then note down the spread values of each symbol for your reference. So I think you need to discard your test results so far and start over. Quote Link to comment Share on other sites More sharing options...
Hacker1713006064 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Was backtesting eur/chf from 2001-2009 with default settings...deposit 500, risk 0.1 and the end result is 19677671129,37 @-) @-) 9999 lots @-) @-) runnin backtest during the weekend is pretty useless.. About all you can really do is verify that the EA trade functions are working.. The backtester relies on too much live data to be of any use during the weeked dead time.. Re-Run your exact same test now that live data is back and see what your results are.. Quote Link to comment Share on other sites More sharing options...
Hacker1713006064 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Well I spend sometime trying to get the EA to use a adjustable time window for trades.. I can get the S1 orders to do whatever I want.. But the S2 trades are always taking place aroung 2100-2300 gmt no matter what.. The Dll is controlling the trade window.. I got around the dll with S1 but have yet to get S2 to behave right.. The best right now for people that want to change the trade window is just change your manual GMT offset to get both trade windows to move.. Quote Link to comment Share on other sites More sharing options...
Hacker1713006064 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hey Don.. Ive noticed while playing with the EA to get a time window working that the only time I could get MT4 to crash is if I have agressive turn on.. While I was trying to get the S2 trades to fall in line with the trade window I made I would move the S1 trades to around 1800 hours and then sure enough the S2 Agressive would get triggered at about 2100 and as soon as the open buy logic was handed off to the dll the crash would happen.. If I turned off agressive and repeated the same backtest the Ea was ok.. Turn back on agressive and as soon as on open order triggered it would lock up.. Something not right with the S2.. At least it seems to me.. Quote Link to comment Share on other sites More sharing options...
mauronio Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Really wonderfull work Mr Don. Thank a lot for this magnificient MDPRO version. Quote Link to comment Share on other sites More sharing options...
hyperdimension Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro The backtester relies on too much live data to be of any use during the weeked dead time. That is incorrect. You should not have MetaTrader connected at all during backtesting, otherwise live quote downloads get mixed in with the historical data (which can cause chart mismatch errors), and the spread may be different each time you run the backtest. Please read my previous post again about properly setting up your MetaTrader terminal for backtesting. Quote Link to comment Share on other sites More sharing options...
Hacker1713006064 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro That is incorrect.. My statement is correct.. You choose to do it a different way does not me incorrect.. Maybe you start a new thread on doing backtest correctly.. That way maybe everyone will read.. Not many reading the MdPro thread all the way thru.. This is a good place to start.... http://www.indo-investasi.com/viewtopic.php?f=6&t=2293 Quote Link to comment Share on other sites More sharing options...
Don Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hacker, all I can say is about the S2 crash problem is that I told to hyperdimension: I have tried every dll and ex4 combinations and it crashed every way. My multi instance-thing seemed to work the best, but it seems it's not resolves the problem completely. We need an "original" customer of the new FapTurbo who also got MDPro and can run the "virgin", unmodified code with the right authentication. I bet s/he will get crashes, too. And I also bet there will be released an official v1.10 sooooon. :D Until that time I would not (and will not) use MDPro on live account. Quote Link to comment Share on other sites More sharing options...
Hacker1713006064 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hacker, all I can say is about the S2 crash problem is that I told to hyperdimension: I have tried every dll and ex4 combinations and it crashed every way. My multi instance-thing seemed to work the best, but it seems it's not resolves the problem completely. We need an "original" customer of the new FapTurbo who also got MDPro and can run the "virgin", unmodified code with the right authentication. I bet s/he will get crashes, too. And I also bet there will be released an official v1.10 sooooon. :D Until that time I would not (and will not) use MDPro on live account. Ok then so whats the big deal with MdPro.. What did it gain.. Seems like if any futher work on MegaDroid that it should take bbrain NO DLL Needed version and go from there... Leave this puppy alone until new dll version or the whatever.. Even if this version was made NON Dll what does it gain over Md11x NonDll?? Can you see anything in this code that is worth it??? I cant.. Although I did pick up a neat function in here called WaitForContext.. They are handling the traffic lane different that I am.. But even then they wrote the code wrong (You never do a loop without a break in it).. But the idea was good.. I fixed the function (and turned it into a bool test ContextBusy) and will use it in my stuff from now on. Quote Link to comment Share on other sites More sharing options...
823 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hi at all! I've downloaded Megadroid EA 3 weeks ago, and i load it on eurusd m15, eurjpy m15 and usdjpy m15; but sice now NO ORDER PLACED!!! Why??? Have i a wrong version? Can you share the right version that I can using on a real account? Thanks Quote Link to comment Share on other sites More sharing options...
Hacker1713006064 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hi at all! I've downloaded Megadroid EA 3 weeks ago, and i load it on eurusd m15, eurjpy m15 and usdjpy m15; but sice now NO ORDER PLACED!!! Why??? Have i a wrong version? Can you share the right version that I can using on a real account? Thanks I dont think your in the right thread.. This is MegaDroid Pro which just came out last week.. Sounds like you have the older MD11 product.. But just for starters did you enter your gmt offset.. OR did you try the AutoGmt thing which didnt work well on the older md11.. When EA working you should be able to run BackTester and get trades.. And please continue discussing on the md11 thread.. This new version of MdPro has shown to crash a lot and be pretty much not usable at the moment.. Quote Link to comment Share on other sites More sharing options...
wizard Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hi at all! I've downloaded Megadroid EA 3 weeks ago, and i load it on eurusd m15, eurjpy m15 and usdjpy m15; but sice now NO ORDER PLACED!!! Why??? Have i a wrong version? Can you share the right version that I can using on a real account? Thanksmd work 1h time frame ;) Quote Link to comment Share on other sites More sharing options...
Iceman777 Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Can someone please look into the code and let me know what the spread limit for EURCHF is when using the default spread control? Just not sure if I should change this to something higher or not as my broker GoMarkets around the time MD trades, spread is usually between 4-5 This pair backtests so well so would like to keep it but dont want to increase spread to much as this will then affect its performance I am assuming I believe EURUSD has always been 3 pip max but dont know if they have increased this for EURCHF as most brokers would have a higher spread on this pair compared to EURUSD Quote Link to comment Share on other sites More sharing options...
⭐ bbrain Posted July 20, 2009 Report Share Posted July 20, 2009 Re: Megadroid Pro Hi at all! I've downloaded Megadroid EA 3 weeks ago, and i load it on eurusd m15, eurjpy m15 and usdjpy m15; but sice now NO ORDER PLACED!!! Why??? Have i a wrong version? Can you share the right version that I can using on a real account? Thanksmd work 1h time frame ;) And on EURUSD only the others will not be handled, except if you have the milti currency version. Br. Quote Link to comment Share on other sites More sharing options...
hyperdimension Posted July 21, 2009 Report Share Posted July 21, 2009 Re: Megadroid Pro That is incorrect.. My statement is correct.. You said:The backtester relies on too much live data to be of any use during the weeked dead time. which is incorrect. The backtester does not rely/depend on or need live data. The only live data that it does use is the spread at the time that you started the test, and this is the reason to why you may get poor results on the weekend. This is one of the quirks of MetaTrader, and is wrong design on the part of MetaQuotes. I encountered this problem myself a while ago when I was doing optimizations and it took me some time to discover the reason for vastly different performance results on the weekend. When your take profit is 10 pips or less, the difference in performance can be huge with a spread of 5 pips versus a spread of 3 pips. I've done plenty of testing in both situations. Actually, the spread on the weekend is probably the same or similar as the spread during the time that MegaDroid is meant to trade (quiet time before Asian session). So if instead you run a test at a time that the spreads are low then you'll get great results because the tester had used a very low spread value. But such results would be misleading if such a low spread value does not exist during the time when MegaDroid is meant to trade. So if you really want results that are closer to reality, you should start the backtest when MegaDroid is meant to trade (e.g. between 21:00 and 23:00 UTC). Or a better alternative is to take the terminal offline by deleting your demo account between 21:00 and 23:00 UTC so that the spread value stored in the terminal can no longer change, and use that for all of your backtesting. Quote Link to comment Share on other sites More sharing options...
crushbeat Posted July 21, 2009 Report Share Posted July 21, 2009 Re: Megadroid Pro Hi, Any updated results today? thanks Quote Link to comment Share on other sites More sharing options...
Ikhnaton Posted July 21, 2009 Report Share Posted July 21, 2009 Re: Megadroid Pro din have any trades on monday..... for me that is Quote Link to comment Share on other sites More sharing options...
samjovanou Posted July 21, 2009 Report Share Posted July 21, 2009 Re: Megadroid Pro din have any trades on monday..... for me that is Same for me. And only one EUR/CHF trade last night. Quote Link to comment Share on other sites More sharing options...
Rio Posted July 21, 2009 Report Share Posted July 21, 2009 Re: Megadroid Pro One very tiny EURCHF trade that closed only 3 pips up. One USDCHF trade that is currently going the wrong way. ...and one GBPUSD trade (logic 2) which went for a tiny profit on FXDD but went for a small loss on another broker. Quote Link to comment Share on other sites More sharing options...
unikin Posted July 21, 2009 Report Share Posted July 21, 2009 Re: Megadroid Pro Monday trades, not bad at all :) 2357271 2009.07.20 22:25 sell 0.10 eurgbp 0.8606 0.8652 0.8556 2009.07.20 22:57 0.8600 0.00 0.00 0.00 9.93 2357598 2009.07.20 23:09 sell 0.10 gbpusd 1.6552 1.6615 1.6492 2009.07.21 00:01 1.6542 0.00 0.00 0.00 10.00 2357727 2009.07.20 23:28 sell 0.10 usdcad 1.1066 1.1110 1.1006 2009.07.21 01:00 1.1062 0.00 0.00 0.00 3.62 2358088 2009.07.20 23:43 buy 0.10 eurgbp 0.8597 0.8534 0.8657 2009.07.21 00:36 0.8603 0.00 0.00 0.00 9.92 2358097 2009.07.20 23:45 sell 0.10 usdchf 1.0682 1.0744 1.0622 2009.07.21 09:59 1.0681 0.00 0.00 0.00 0.94 2358654 2009.07.21 00:34 sell 0.10 eurchf 1.5199 1.5242 1.5139 2009.07.21 01:37 1.5194 0.00 0.00 0.00 4.68 Usdchf was swinging for far too long and could easily ended up in negative. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.