zifx25000
83
zifx25000  
DerkWehler:
Answers inline below...

thanks for answer

kr

jerome

sinecure
324
sinecure  

No trades

I have had no trades on 4 pairs with the new settings. Should settings be changed?

autumn

sinecure
324
sinecure  

Settings changed

On a whim, I switched MoneyManagement and DualWinMM to true and set a SL of 40 and immediately triggered long entries on USDJPY and GBPJPY 19:00 hours GMT.

autumn

Michael
733
Michael  

I'm your test and code helper dude ... person

DerkWehler:
Autumn:

In your posts, sometimes I think you are talking about my Ind/EA, sometimes yours. If you are speaking of mine, could you mention me in there so I know? Otherwise I will assume you are talking about yours.

And BTW, some sort of concerted testing effort is always a good idea, but orchestrating is the hard part, and I don't have time to oversee that. If someone else wants to contribute by getting folks going, I will be happy to assist.

Thanks,

-Derk

Derk,

Just tell me what you want tested (back or forth) what settings and platform.

I can backtest on my laptop (AMD Turion 64x2 2GB Ram Vista Ultimate x64 - getting back data from MoneyTec(?). It's StrategyBuilder FX (StrategyBuilderFX MT4 Trading Platform). It's based on MT4). I like back testing using these guys because I get the most reliable data (outside of alpari) from them.

I've also got a server that can run a few virtual machines (I'm thinking XP SP2 w/ 512 MB RAM and the ONLY thing they'd be running would be the OS and Metatrader(s); plural if we use different brokers to test on). I could put multiple traders on each OS.

My only requirement in testing is NO MoneyManagement. I've had a few EA's that when that was turned on, i'd get stopped out of testing because the demo account got drained or it 'profited out' due to limitations in the tester.

NOTE: and this is to EVERYONE, I see a lot of you have been having problems forward testing. Here's your answers: 1. Pick a time frame for your chart and stay there. They guy on the Synergy webinar even says so. When you've got indies that draw based on historical data, you always run that chance of them re-drawing or not drawing at all when you switch time frames. 2. When forward testing, only have 1000 bars or less in your charts (Tools, Options, Charts). No indicator in this list or EA should need anymore than that. And if you've got a trade expanding more than 1000 bars, that's a different issue.

BTW, I've used this as my lot calculations per postion. Risk is amount risked on equity PER TRADE.

lot=NormalizeDouble((AccountFreeMargin()*(Risk/100))*AccountLeverage()/MarketInfo(Symbol(),MODE_LOTSIZE),2);
Derk Wehler
345
Derk Wehler  

A Brief Chat On Money Management

Autumn:

How odd...

This makes me believe that something was wrong before with the amount of lots it was attemting to trade. By default that was 0.1 lots. Any reason to think your broker will not do that (who is your broker)?

Did you look in the experts & journal tabs to see if it was giving you any messages? If you switch it back and watch the experts tab, I expect you will see that it is trying to take a trade but failing for some reason...

Also, I should have told everyone: One of the things you can optionally change is of course the money management. By default it was set to NOT use it and to trade 0.1 lots. But you may wish to change your lots value or enable one of the MoneyManagement strategies. For now, there is regular money management and the DualWin MM. The DualWin is the same as the Binary MM system,for those of you familiar. The system is too long to get into in this post, but basically it seeks two winning trades in a row to make back any previous losses. If you research "Binary" and "d'Alembert", you may find something on it.

And as for the regular MM, if you have SL set to zero, it will calculate lots in the standard fashion*, but if you do not, it will use that SL value to calculate the lots to bet such that if you lose, you actually lose that percentage of your account. I realise now that there is one more change I need to make: If you are using the PAC as your stoploss (which I currently am), then it SL value is indeed known at the time the trade is placed,; therefore I can use that info to calculate the lots.

*What I call "standard fashion" is where you use your account leverage/margin to calculate the lots. Ex: You have a $10,000 account and you set it to risk 5%, and you have 200:1 leverage. It calcualtes 5% of 10K to be $500. At 200:1, the margin on one lot is $500 so that is what it trades.

My other method, which requires the max SL to be known ahead of time, works like this example: You have the same account as the example above, and still risk 5%. It calculates the 5% to be $500 and best the amount of lots where if you lose, your account will be down that $500.

-Derk

autumnleaves:
On a whim, I switched MoneyManagement and DualWinMM to true and set a SL of 40 and immediately triggered long entries on USDJPY and GBPJPY 19:00 hours GMT. autumn
Derk Wehler
345
Derk Wehler  

Backtesting

Non:

Excellent...thanks. I'd hoped someone would step up with assistance. Perhaps we can work a bit more closely, and you can post any relevant results back to this group...?

Do you use skype? You can catch me on there, same username. Or the same username at gmail for email.

I tend to think it is best to test w/o MM as well, then if you get good results you know you will get even better ones running with it turned on. My lot calcs are similar to your posted, at least for the basic one. The one I prefer is outlined in my last post. The only flaw is it requires you to know your initial SL, which is simply not possible with some strategies.

nondisclosure007:
Derk,

Just tell me what you want tested (back or forth) what settings and platform.

I can backtest on my laptop (AMD Turion 64x2 2GB Ram Vista Ultimate x64 - getting back data from MoneyTec(?). It's StrategyBuilder FX (StrategyBuilderFX MT4 Trading Platform). It's based on MT4). I like back testing using these guys because I get the most reliable data (outside of alpari) from them.

I've also got a server that can run a few virtual machines (I'm thinking XP SP2 w/ 512 MB RAM and the ONLY thing they'd be running would be the OS and Metatrader(s); plural if we use different brokers to test on). I could put multiple traders on each OS.

My only requirement in testing is NO MoneyManagement. I've had a few EA's that when that was turned on, i'd get stopped out of testing because the demo account got drained or it 'profited out' due to limitations in the tester.

NOTE: and this is to EVERYONE, I see a lot of you have been having problems forward testing. Here's your answers: 1. Pick a time frame for your chart and stay there. They guy on the Synergy webinar even says so. When you've got indies that draw based on historical data, you always run that chance of them re-drawing or not drawing at all when you switch time frames. 2. When forward testing, only have 1000 bars or less in your charts (Tools, Options, Charts). No indicator in this list or EA should need anymore than that. And if you've got a trade expanding more than 1000 bars, that's a different issue.

BTW, I've used this as my lot calculations per postion. Risk is amount risked on equity PER TRADE.

lot=NormalizeDouble((AccountFreeMargin()*(Risk/100))*AccountLeverage()/MarketInfo(Symbol(),MODE_LOTSIZE),2);
Derk Wehler
345
Derk Wehler  

Resutls of the week

Hello my little droogies:

Just to let you know, I ran the EA on GBPJPY (Default settings) this last week and it ended up 33% ahead. It should be noted however, that there was a bug in the code for the first couple trades; it opened a second set of trades where it should not have. Then it closed all 4 at the same time. Since the second set should not have been placed, and were good for $130, if you subtract those out, I'm at 20% for the week.

That is on a $1000 demo account on Velocity4x, trading 0.1 lots per trade, no MM.

sinecure
324
sinecure  

Logs and Backtesting

Hi Derk,

Thanks for the comments on the intricacies of Synergy trading. If it is of some help to you I attach copies of the Experts and Journal logs captured today. Please let me know if I can do anything else in the meantime.

For the moment I am immersed in redesigning the signals in the Chimera simple version so that it can be forward tested efficiently, and this involves a fair amount of backtesting. Once I begin to get consistent results I will upload a functional version and at that point I can turn my attention to backtesting SynergyEA.

I'll keep watching the thread for developments. It's great that others are pitching in on backtesting SynergyEA, as this should speed up development considerably.

Derk Wehler
345
Derk Wehler  

Autumn:

Please use this Ind instead; it has the Print statement removed (the ones that cause these lines:

SynergyInd USDJPY,H1: PAC EXIT: i = 0; MA_Hi_1 + Spread = 108.18479711

Other than that, I can see there was an error toward the bottom of the file:

2007.11.23 12:37:53 LibOrderReliable USDJPY,H1: OrderSendReliable v1_1_3:non-retryable error: 131:invalid trade volume

2007.11.23 12:37:53 LibOrderReliable USDJPY,H1: OrderSendReliable v1_1_3: attempted OP_BUY 0.05000000 lots @108.31000000 sl:0.00000000 tp:0.00000000

Any idea why it was attempting to trade 108.31 lots?

Turn both MoneyManagement settings off, set Lots = 1.0, use a $10000 demo account (or larger); use this new indicator with the output turned off so it does not confuse the issue, and if you see it fail to take another trade, paste the contents of the experts tab over to me again.

Thank you,

-Derk

autumnleaves:
Hi Derk,

Thanks for the comments on the intricacies of Synergy trading. If it is of some help to you I attach copies of the Experts and Journal logs captured today. Please let me know if I can do anything else in the meantime.

For the moment I am immersed in redesigning the signals in the Chimera simple version so that it can be forward tested efficiently, and this involves a fair amount of backtesting. Once I begin to get consistent results I will upload a functional version and at that point I can turn my attention to backtesting SynergyEA.

I'll keep watching the thread for developments. It's great that others are pitching in on backtesting SynergyEA, as this should speed up development considerably.
Files:
sinecure
324
sinecure  

New Indy

Got it, but I'm not sure how to turn off outputs. Is that in the indy?

autumn