bigdir

Hi all,

 

There was a thread going on in another forum trying to replicate the performance/strategy of Million Dollar Pips. Unfortunately the thread seems to be going walkabout on occasion, so I'd thought I'd start a new one here with the hope it can find a new life :)

 

Below is a link to the code and a description of the MDP code - full credit goes to Capella for developing & maintaining the code. From what I understand, this is an entirely open source effort and does not use the original source of MDP, merely the strategy is attempted to have been replicated.

 

This said, I haven't been able to get some decent tick back-tests out of the code. I'm wondering whether the original MDP back-tested well? The other notable difference is that I believe MDP uses a number of "channels" which presumably are different timeframes which the Bollinger Bands are breached (which is how it scalps). This clone uses a combination of moving averages and bbands instead.

 

Really interested to see whether we can make an effort to enhance the system...

 

 

http://www.multiupload.com/4GMJEKBEJR

Hey guys - does anyone have the latest and greatest of this? Even a re-post of the last link would be very helpful!

Hey all - the data import works fine. Just have a single column in your input file (numeric/floating point).

 

Regarding the modelling side of things, not sure if anyone (esp. progressapama's friend) can offer some high-level tips of how a profitable model can be designed?

 

For eg:

- What time-frames are typically used?

- Are the profitable strategies scalping ones?

- How many data samples are typically used as inputs?

- Do you use any custom functions to determine best fit?

 

Thanks in advance for any feedback!

Hi guys - did a tick-data back-test with the "pure" (demo) version you can download from the EA-Boss website. The win rate is good, but it regularly started hitting stop-losses in the last few months which saw it in the negative over 6 months. Doesn't seem to match up with what's being shown on their website!

It's shocking that they seem to be releasing new versions with apparently few changes in the strategy. Hitting the stop-loss once takes you back to square one (and even before that). Has anyone managed to get this sucker to actually work?

I have reviewed the code between v48 & 52 and can't see a big difference except for an introduction of the EURUSD pair for scalping. Not sure how they are supposedly getting the results they show on their site...

Thanks for the share, guys.

 

Not sure whether I'm doing something wrong with it? I'm testing it on Dukascopy tick data for EURGBP and the equity graph heads pretty much down all the way. I get a few "OrderModify 1" errors, however I think this could be due to the EA not checking if the modified price is the same as current - don't _think_ it's affecting the results.

 

Anyone can confirm that they're getting some decent results of of this? Is it back-testable or do you need to forward test?

Umm - perhaps we can keep objective in this discussion, otherwise I could also question your motives (and it really wouldn't make me lose or win an extra dollar, so why should I care? This IS a free EA we're talking about.)

 

Backtest EURUSD M15 - cacusmaccimov2 - Dukascopy tick data - jan '10-Jun '10 - 100% wins. Maximal drawdown - 9%. 30% ROI.

 

Whether or not it works in live mode is an exercise for all the fine folk out there.

arjunna - this is normal. The EA is just displaying information about it's price analysis.

Hi cacus, I felt compelled to reply to hopefully make you see the events as something positive. If it wasn't for forexlegend's appearance, I would not know anything about you or your good work (and the forum you've been frequenting). As they say, imitation is the best form of flattery and in terms of respect I think numerous others on this forum would concur that sharing EA Maccimo really puts you amongst valued contributors.

 

If you want to use the Linux analogy of sharing, then I think that ForexLegend is simply the RedHat of your work - making it accessible to those who want something shrink wrapped. You, however, are the Linus. Problem is, however, that FL are going to have to somehow sustain some sales in the face of other people soon realizing they can get this work for free from a guy named Cacus.

 

As one member suggested, maybe this is an opportunity to get some income off the back of your efforts (if that's of interest to you). Anything coming from 'cacus' is going to have much more credibility than something coming from ForexLegend.

 

Regards.

cacus - would you be able to kindly describe how your MACCIMO strategy works in a few words or point us to some information on it (I've had a read of the forexmt4.com but it doesn't describe the strategy)? From what I can see, it is much more performant then most of the commercial stuff that's out there. I'm just wondering also whether there are some weak points where it could blow an account. Back-tests I have done on the EURUSD shows excellent results.

 

Thanks!

Hi Benz - Looks like a little playing around with the GMT (to -1 for Dukascopy) seemed to yield better results for EURCHF. Thanks for the tip. Might have to do a few backtests to find out more. Does anyone know of a version of this that eliminates the DLL? I always like to see what makes thing tick ;)

 

http://i29.tinypic.com/dfvdb4.jpg

Hi Benz - Interestingly, the EA seems to hang on one of the MT installations I have but not another. Looking through the DLL, it seems to be searching for an open MetaTrader window. Tried closing only the one I wanted to use, however did no good.

 

Anyway, I have backtested this with default settings on EURGBP/15m and the results seem quite erratic - certainly nothing like fx4_ever's graph! 26% draw-down although a good win-rate of 70-80%. Equity graph makes a nice run followed by hefty losses that gradually whittle down the account.

Benz - thanks for your file. Not sure if anyone else is having the problem, but whenever I try backtest either v1.04 and 1.05, the backtest just hangs. Last message in the journal is:

 

2010.08.02 10:49:49 EAKAIN SCALPER PRO 2010 v1.05-bnz inputs: GMT_Offset=0; Risk=10; FixedLots=0; MaxUsedMargin=100; MagicNumberS0=1; MagicNumberS1=2; MagicNumberS2=3; Slippage=2; EURCHF_maxspread=4; EURGBP_maxspread=4; EURUSD_maxspread=3; USDCHF_maxspread=4; USDCAD_maxspread=5; USDJPY_maxspread=3; GBPCHF_maxspread=8;

 

On the 'Settings' status bar of the historical tester, not one notch progress is ever registered.

Further investigation seems to indicate that the first call to CheckLic call is hanging.

 

Appreciate any tips (dlls are enabled - other EAs using DLLs are working fine).

Hi abdulisback - I have done a few backtests on EURGBP (v3052) and I must say that the equity curve is rather unstable. That said, with 5% risk, it seems to be able to hold a steady break even until it finally comes across favorable market conditions. This is when it makes a few gains. All-in-all, from what I've seen so far I would tend to agree with you that there's nothing here to write home about..

Might be worthwhile checking out some of Birt's forward tests here:

 

http://www.myfxbook.com/members/birt/euronis-4/12393

 

And his great review here (of v4160):

 

http://eareview.net/archives/140

 

His backtests make it quite clear that the system is spread sensitive which means it'll probably take a bit of analysis to determine which brokers & pairs are suitable. Regarding "blowing" the account, Birt also comments on the fact that there is a setting to limit risk & number of open trades. When he uses a reduced risk the drawdown is much more manageable.

 

If you have an EA that doesn't need this type of analysis before going live with it, please let us know ;)

Thanks iwjw for your response.

I started a new thread here with Euronis 3052, although I haven't yet tested it:

 

http://indo-investasi.com/showthread.php/7964-Euronis-3052

 

Quick link to archive: http://www.2shared.com/file/bOIAg4Nb/3052.html

This has been mentioned in another thread as being quite good. I managed to get a Russian version of the archive - it should be educated although I haven't tried it yet.

I have translated some of the directory names into English, however the manual is still in Russian...

 

www.2shared.com/file/bOIAg4Nb/3052.html

Thanks cacus.

 

If the ForexLegend EURCHF scalper is working well - is anyone able to comment whether the cacus' work is the entirely the same? If so, surely that can be used instead of Forex Legend?

cacus - great work mate! What is the yahoo group you've been frequenting? Sounds like it's a good place to learn.

Guys - I am not sure how you're seeing EA Kain to be profitable long term. My back-tests with Dukascopy data show that while you do get a high-percentage win rate, inevitably the strategy hits a comparatively large stop-loss (10:1). The equity curve takes some time to recover and hit a few more of these and it's downhill from there. I have tested this on EURUSD & EURGBP with aggressive mode on and off. While they perform great for a few months at some stage their luck runs out, losing much of their gains.

I concur! If someone has the original version I can try to educate...

Hi there - According to their website, BlueWave already have created an automated strategy for their system (http://www.bluewavetrading.com/tradingsystems.cfm). Does anyone have access to this? I can possibly educate it..