Net Trap EA - variant backtesting

This is a discussion on Net Trap EA - variant backtesting within the Forex Net Trap forums, part of the T2W Archive category; Hi I'm going to run the EA through the Metatrader Strategy Tester with a selection of the NT variants. The ...

Reply
 
LinkBack Thread Tools Search this Thread
Old Oct 21, 2010, 8:08am   #1
Joined Dec 2006
Net Trap EA - variant backtesting

Hi

I'm going to run the EA through the Metatrader Strategy Tester with a selection of the NT variants. The initial test period will be 01/01/2010 to 19/10/2010. I don't have a lot of faith in the Strategy Tester in terms of absolute results BUT I do think it's a reasonable tool for comparing the effects of the various options.

I will post each test report separately, along with an associated MT parameter "set" file, which allow that variant's parameter settings to be used for the EA.

Cheers,
Greg
SpindriftUK is offline   Reply With Quote
Old Oct 21, 2010, 8:27am   #2
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread And here's one I prepared earlier

Attached are the report and set file for VNT. Just realised I can't post the .set file

Ok, I have attached the set file as vnt.txt. If you want to use it, rename it to .set and place in your experts/presets folder. You can then use the Load button in the EA parameter input dialog to use it.

The little graph is supposed to be embedded in the report but for some reason, it gets lost. Have attached it separately.


Cheers,
Spin
Attached Thumbnails
vnt.gif  
Attached Files
File Type: htm vnt.htm (180.9 KB, 542 views)
File Type: txt vnt.txt (881 Bytes, 253 views)
SpindriftUK is offline   Reply With Quote
Old Oct 21, 2010, 10:12pm   #3
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread Given that RH has just re-opened the doors on Net Trap, I thought I would see how the EA with VNT parameters set would perform over the same period that RH is quoting results for. Not surprisingly, the results are widely different from month to month, with the EA's results for June, especially, being very suspect. I had a lot of trouble getting June's history data and I have a feeling it has probably been constructed on a best guess basis

Anyway, what was surprising was that the EA's final profit for that 16 month period came out very close to RH's published results - $19179 at $10 per pip against RH's £19670 at £10 per pip. I have attached a spreadsheet of the EA's results - read into it what you will!

I did an optimisation exercise for the period 01/07/2010 to 30/09/2010 - 25 passes with stop loss varying between 30 and 70 in steps of 10 and take profit varying between 10 and 50 in steps of 10. I have attached a screenshot of the results (SLPipsS was set the same as SLPipsB for the purpose of the exercise).

Cheers,
Spin
Attached Thumbnails
optimisation-results.png  
Attached Files
File Type: xls VNT backtest 06_2009 to 09_2010.xls (206.5 KB, 239 views)
SpindriftUK is offline   Reply With Quote
Old Oct 22, 2010, 10:14am   #4
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread Going back to the "year to date" tests, I have attached the results for VNT plus RH's advanced tactics: ant1- Safety Net Plus and ant2 - Safety Net Extension. Also attached are the parameter files for the EA for each option.

Cheers,
Spin
Attached Thumbnails
vnt_ant1.gif   vnt_ant2.gif  
Attached Files
File Type: htm vnt_ant1.htm (205.8 KB, 465 views)
File Type: txt vnt_ant1.txt (881 Bytes, 229 views)
File Type: htm vnt_ant2.htm (230.2 KB, 496 views)
File Type: txt vnt_ant2.txt (883 Bytes, 255 views)

Last edited by SpindriftUK; Oct 22, 2010 at 8:38pm.
SpindriftUK is offline   Reply With Quote
Old Oct 22, 2010, 5:52pm   #5
Joined May 2010
Re: Net Trap EA - variant backtesting

Quote:
Originally Posted by SpindriftUK View Post
Going back to the "year to date" tests, I have attached the results for VNT plus RH's advanced tactics: ant1- Safety Net Plus and ant2 - Safety Net Extension. Also attached are the parameter files for the EA for each option.

Cheers,
Spin
Is this risk-adjusted (i.e. 3% at risk divided by number of pips) or at the same lot size?
Tanguero is offline   Reply With Quote
Old Oct 22, 2010, 6:53pm   #6
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread
Quote:
Originally Posted by Tanguero View Post
Is this risk-adjusted (i.e. 3% at risk divided by number of pips) or at the same lot size?
Almost the same lot size I allowed the EA to calculate a percentage but forced it to use the same bank every time. I think the slightly different lot sizes are the result of rounding the 5 decimal point OHLC values prior to adding the 50 pip target during the calculation of open price.

Btw, I notice that I didn't specify the parameters correctly to simulate ANT2. I should have set the lock-in pips to -30, as that would keep the stop loss the same as originally set. As I've got it, stop loss will be moved to break even. I will run it again and hopefully edit the post, although I notice I can only do the edit if the post is less than a day or so old.

Cheers,
Spin
SpindriftUK is offline   Reply With Quote
Old Oct 22, 2010, 7:22pm   #7
Joined May 2010
Re: Net Trap EA - variant backtesting

Quote:
Originally Posted by SpindriftUK View Post
Almost the same lot size I allowed the EA to calculate a percentage but forced it to use the same bank every time. I think the slightly different lot sizes are the result of rounding the 5 decimal point OHLC values prior to adding the 50 pip target during the calculation of open price.

Cheers,
Spin
Thanks Spin though I am afraid I don't quite follow: don't you just divide the bank by the SL so that the candle readings have no bearing on the calcuation of the lot size?
Tanguero is offline   Reply With Quote
Old Oct 22, 2010, 7:28pm   #8
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread
Quote:
Originally Posted by Tanguero View Post
Thanks Spin though I am afraid I don't quite follow: don't you just divide the bank by the SL so that the candle readings have no bearing on the calcuation of the lot size?
You're right, of course. I've obviously been at this too long today

I'll take a look at why there's a slight variation in the lot sizes through the back test.

Time for a beer, methinks!
SpindriftUK is offline   Reply With Quote
Old Oct 23, 2010, 11:44am   #9
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread
Quote:
Originally Posted by Tanguero View Post
Thanks Spin though I am afraid I don't quite follow: don't you just divide the bank by the SL so that the candle readings have no bearing on the calcuation of the lot size?
The reason for the slightly different order volumes throughout the backtest, despite using a fixed bank, is that "convertToGBP is true and the exchange rates differ from day-to-day.

Cheers,
Spin
SpindriftUK is offline   Reply With Quote
Old Oct 23, 2010, 3:29pm   #10
Joined May 2010
Re: Net Trap EA - variant backtesting

Quote:
Originally Posted by SpindriftUK View Post
The reason for the slightly different order volumes throughout the backtest, despite using a fixed bank, is that "convertToGBP is true and the exchange rates differ from day-to-day.

Cheers,
Spin
Sorry, by "lot size" I actually meant the number of lots.
Tanguero is offline   Reply With Quote
Old Oct 23, 2010, 4:57pm   #11
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread
Quote:
Originally Posted by Tanguero View Post
Sorry, by "lot size" I actually meant the number of lots.
I thought that's what you meant. Order volume (number of lots) is multiplied by the exchange rate to convert dollar-based lots for the trade. The EA is not clever here though and if you set convertToGBP to true and your account is in USD, the conversion will still be applied.

Cheers,
Spin
SpindriftUK is offline   Reply With Quote
Old Oct 24, 2010, 12:32am   #12
Joined May 2010
Re: Net Trap EA - variant backtesting

Spin, in your backtest for 1.01.10 to 19.10.10 the result is close to 19430 mark, while for the 16 months' period you say it is 19,179. This would mean that for 7.5 months the system made (250) loss?
Tanguero is offline   Reply With Quote
Old Oct 24, 2010, 12:55pm   #13
Joined Dec 2006
Re: Net Trap EA - variant backtesting

SpindriftUK started this thread
Quote:
Originally Posted by Tanguero View Post
Spin, in your backtest for 1.01.10 to 19.10.10 the result is close to 19430 mark, while for the 16 months' period you say it is 19,179. This would mean that for 7.5 months the system made (250) loss?
You're comparing the profit made on the 16 month RH comparison test, which was 19179, with the final balance of the ytd test, 19492.56.

What is interesting, perhaps, is that the profit on the RH comparison test between June 1st and October 22nd is 8904, whereas the profit on the same period in 2010, taken from the ytd test, is only 4025. And the latter test is using slightly higher order volumes.

I'm afraid my testing so far has been a bit haphazard, which is why, if we're to get any meaningful results, I think we need to come up with some kind of test plan.
SpindriftUK is offline   Reply With Quote
Old Oct 25, 2010, 10:10am   #14
Joined May 2009
Re: Net Trap EA - variant backtesting

Got some interesting findings so far.

Limited backtesting but from 1.8.2010 to date (including Friday/Monday loss today), 3% risk, £3,000 starting bank goes to £4,875 and PF of 1.8 and drawdown only 11.4%...will run on a small live account along with a couple of other settings.
Dubai is offline Trading system vendor   Reply With Quote
Old Oct 25, 2010, 10:57am   #15
 
mark120169's Avatar
Joined Apr 2010
Re: Net Trap EA - variant backtesting

Quote:
Originally Posted by Dubai View Post
Got some interesting findings so far.

Limited backtesting but from 1.8.2010 to date (including Friday/Monday loss today), 3% risk, £3,000 starting bank goes to £4,875 and PF of 1.8 and drawdown only 11.4%...will run on a small live account along with a couple of other settings.
Be careful, you would be able to make most variants "settings" look good from 1/8/10 testing because from 18/8/10 to 17/9/10 there was a huge winning streak with only 3 loss days on variants with 30 pips or lower targets. Run any test from 18/8/10 todate ant the results would be even better due to a losing streak from Mid July until Mid August.

I would be very interested in any settings found for the whole of 2010, then as I have the structure set up I could manualy cross check the MT4 results with forex tester for confirmation.

So any "golden nuggets" found we don't already use please post the S/L & T/P settings and I will manual back test them. Not too many though!! it's hard work!

M
mark120169 is offline   Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Similar Threads
Thread Thread Starter Forum Replies Last Post
Net Trap Variant live results from 17/05/10 mark120169 Forex Net Trap 572 Dec 5, 2011 11:54pm
Net Trap Tools mark120169 Forex Net Trap 10 Dec 11, 2010 10:02pm
Forex Net Trap SA Trader Forex Net Trap 6 Jun 28, 2010 2:44pm
Net Trap Johnnysportster First Steps 5 Mar 30, 2010 10:08am

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)