Welcome back Forex fans. Here's a quick weekly update on the robots i'm following.Megadroid had a pretty decent week and picked up 20 pips on both the live and demo accounts. This week's Megadroid trades were exactly the same in the live and demo accounts.
Another interesting thing to point out - I backtested Megadroid for 2010 year to date using the same risk setting I use in my live account (0.05). The results were very close to what I got in my live account, +7.5%, although the trades were not identical. I'm not sure what this means, but I thought it was worth pointing out since it may address some concerns that back-tests results will differ substantially from live results.In other action, God's Gift GBP/USD dropped 54 pips and is down about -5.9% on the year. God's Gift EUR/USD picked up 9 pips but remains down over 10% for the year. The developer indicates these robots trade flat to downish most of the time and typically make their yearly gain in just a few big winning trades. So it will be interesting to see how it all works out.
Anyway, check back later because I have some good stuff coming up in a post over the weekend.
The results were very close to what I got in my live account, +7.5%, although the trades were not identical. I'm not sure what this means, but I thought it was worth pointing out since it may address some concerns that back-tests results will differ substantially from live results.
ReplyDeleteChris, wouldn't this be due to the interpolation error on the 1 minute bars that occurs in backtesting? I would expect that many trades wouldn't initiate at the same point on the bar, nor end on the same point. It would be another matter altogether if you weren't getting simulated trades or exits at roughly the same time as the real trades, if the data sources were the same. I'm sure there are some interesting complications lurking in the backtesting vs live trading issue. Hopefully this work you are doing can help explain the discrepancies.
Hey Chris,
ReplyDeleteGood to see you are doing well with Megadroid. I still can't wait to see your results in 6 months. Not that I want you to experience draw down, but I would like to know how Megadroid behaves in a draw down situation on a live account. I think you had one in your demo right?
I am pretty close to just ditching my blog. I just disabled anonymous comments after the recent negative one I got on there. These people out there probably just want me to stop as I don't have a lot of good to say about Zulu, and being close to the top of a search for Zulutrade on google with not much good to say, does not bode well for the signal providers/promoters on there. This last guy was not too happy about my comment on Pro-B5 while responding to another comment...
At any rate, Tiger Tail had another successful trade this week. My code was a bit tight on the trade though. I labeled the currency pair as being more risky and the code only calculated 1 mini lot when I probably should have swung 2 mini lots. I may have a EUR/USD trade next week, and I hope it works out.
Redford-
ReplyDeleteI did some analysis on the intra-bar interpolation issue by exporting the EUR/USD M1 bars to a spreadsheet for the month of January. I subtracted the low from the high and found that most of time when Megadroid trades (9PM-12AM EST) the range of the bar is well under 10 pips which seems to be the default take profit for Megadroid.
In my limited understanding, this means that the intra-bar interpolation errors are not as likely to occur since one bar's action is not likely to trigger an exit in back testing that would not occur in a live or demo execution.
I believe that the issue occurs in situations where the take-profit is less than the range of the 1 bar because you can't be sure if the take profit would occur during the bar simulation in the same order as it would in real trading.
So IMHO the issue seems to be overstated and under represented in reality. But nothing is forever, particular in Forex so time will tell.
Correction - Megadroid trading windows is 9PM - 12AM GMT..
ReplyDeleteJT-
ReplyDeleteRight - it will be interesting to see how I make out with Megadroid. I know i'm overdue for some drawdown with 18 wins and 1 loss. In the last 2 years, there's been up to 45% drawdown so i'm not under the delusion that I won't go into drawdown. That's why i'm keeping the $ amounts small.
Also, i'm working on getting some other systems into the pipeline and i'm just getting started there.
As for Zulu, you know where I stand, so i'll leave well enough alone. Cheers,
Chris
I wanted to ask you. How are you running all the demos and such? You must have at least 2 PC's? I've been thinking about getting a netbook where I can just let it run without using too much power for my live account when I get enough data from the demo account.
ReplyDeleteI am actually pretty interested in Megadroid, I'm just a bit worried about the draw down periods as I back tested the Megadroid Pro and noted those periods.
Sorry about saying Zulu on your blog. It is quickly becoming a swear word for me too. :)
JT
JT -
ReplyDeleteI have 2 year old Dell computer running 6 different instances of Metatrader on it. Its the same computer we use for family e-mail, etc so I have to train my family to just minimize MT4 if they see it, and so far I haven't had any incidents.
I think i'm close the limit on my main computer and I think i'm going to start up a secondary to test a few more EA's.
I'm going to move everything to a VPS once I get a broker arrangement where they will host it for free. I'll get there later this year. The only thing I would say is to use a separate account and instance of MT4 for each EA, its much easier to keep track of that way,
Chris
Here's a fine article on how the spread can impact your backtests, specifically for Megadroid.
ReplyDeletehttp://trading-gurus.com/the-metatrader-4-backtesting-blues-track-1/
Okay... I was wonder how people were doing multiple instances on one PC. I Found that you can have separate installs so to speak. Basically start up a different terminal.exe. I was hoping to do this on a low power consumption PC, like a netbook. I may have to go with a regular PC.
ReplyDeleteJT
Redford -
ReplyDeleteThanks once again for the info.
The key take-away from the article is that the back-tester uses the spread displayed on the terminal at the time you press the 'Start' button on the back-tester for the entire length of the back-test. Most of the time the spread on EUR/USD for Forex.com is 2 pips, but it can move to 4 pips during the scalpers trading window - which can definitely make a difference.
Chris
JT -
ReplyDeleteThe key is to install each instance of MT4 in a different folder - which you get prompted for during setup.
From there you have a distinct instance of MT4 with its own experts folder, etc. Of course, you can run multiple EA's attached to different charts inside a single MT4, but you can only be logged into 1 account at a time. So that's why it makes sense to use a completely different instance of MT4 for each EA,
Chris
Goog blog!!
ReplyDeleteThank you.
Please hold out in the future.
...(I'd be pleased if you exchange reciprocal link with me.)
forex chart analysis and a cat