Old Dog with New Tricks | Page 49 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Old Dog's Taming of the Beast? 34 replies

Old Dog Fib Trading - EA Development 18 replies

Old Dog Fib Trading - EURUSD 399 replies

5 Tricks of Trend Trading, by Rob Booker 25 replies

How To Work Around A Market Maker's Tricks 0 replies

Attachments: Old Dog with New Tricks
Exit Attachments

Old Dog with New Tricks

  • #961
  • Sep 22, 2010 7:56am Sep 22, 2010 7:56am
I have to more than 600.000 bars on charts.
  • #962
  • Sep 22, 2010 8:11am Sep 22, 2010 8:11am
Your data then goes back to approximately July 2009 on your charts?

Can you post a picture showing the losses?

If the losses are spread out, then my guess would be spread spikes.

If the losses are close together, you may have hit a mini-swan.

Overall, the backtester inside the EA is as accurate as the data it gets. Barring the win rate, did the trades meet the rest of the criteria? If so, leave the EA running. What you may have come across was a statistically unlikely situation. This can go both ways:

 

  1. Long string of losses
  2. Long string of wins

  • #963
  • Sep 22, 2010 8:46am Sep 22, 2010 8:46am
Hy ther guys.I have one of topic for you RONALD RAYGUN.Is there posible to run EA on offline renko chart.I look for somone to code simple grid ea for offline renko chart.Thnx
  • #964
  • Sep 22, 2010 11:33am Sep 22, 2010 11:33am
Quoting Ronald Raygun
Your data then goes back to approximately July 2009 on your charts?

Can you post a picture showing the losses?
I can go back to Nov 2008 on chart
Please look at the attached screen
Attached Image (click to enlarge)
Click to Enlarge Name: hilo.png Size: 84 KB
  • #965
  • Sep 22, 2010 3:53pm Sep 22, 2010 3:53pm
Quoting LeW
I can go back to Nov 2008 on chart
Please look at the attached screen
Looks like you hit a huge news spike.
  • #966
  • Sep 23, 2010 3:19am Sep 23, 2010 3:19am
Yes, the FOMC Statement caused high volatility on the markets and losses
Ron, is the EA opened trades during this news on yours account?
What about my settings? Is it less "strict" then yours? Which parameters?

Regards
  • #967
  • Sep 24, 2010 4:26am Sep 24, 2010 4:26am
Hi,

After a while doing some experiments with this EA, I need, if possible, some clarifications:

Somewhere in this thread, Ronald said that sometimes he gets SS above 120 that means for example a WR-60 and RR-2 or WR-80 and RR-1.5 and somewhere before he mention to start with WR-60 and RR-1 and try to improve results.

What I'm trying to ask is which parameter is more important, WR or RR because I've made a backtest with 1.5 years data and WR-80, RR-1.5, SL/TP-110 on USDJPY and the results where exactly the opposite - only 20% wins.

Can anyone help on this?

Thanks,

Pedro
  • #968
  • Sep 26, 2010 10:05am Sep 26, 2010 10:05am
Quoting Pgcoimbra
Hi,

After a while doing some experiments with this EA, I need, if possible, some clarifications:

Somewhere in this thread, Ronald said that sometimes he gets SS above 120 that means for example a WR-60 and RR-2 or WR-80 and RR-1.5 and somewhere before he mention to start with WR-60 and RR-1 and try to improve results.

What I'm trying to ask is which parameter is more important, WR or RR because I've made a backtest with 1.5 years data and WR-80, RR-1.5, SL/TP-110 on USDJPY and the results where exactly the opposite - only 20% wins.

Can anyone...
I think your Sl and TP is too tight. The EA built-in tester does not include spread value during optimization so if your target is 110, the 20 pips spread may have a big influence on live results...

Has anyone found the profitable settings? If yes, please post some advise how to configure this EA?

Thanks
  • #969
  • Sep 27, 2010 2:21pm Sep 27, 2010 2:21pm
Ron, is it true that your SS is set to about 1,2 ??
  • #970
  • Sep 27, 2010 5:40pm Sep 27, 2010 5:40pm
Not even close
  • #971
  • Sep 28, 2010 8:04am Sep 28, 2010 8:04am
Quoting LeW
Ron, is it true that your SS is set to about 1,2 ??

I've just reviewed all Ron Posts and really He didn't mention anytime SS of 120. Probably I dreamed with it.
Sorry people for this misleading info in my previous post.

Pedro
  • #972
  • Oct 3, 2010 5:27pm Oct 3, 2010 5:27pm
A lot of people here seem to ask for the best settings. The reality is that there are different profitable settings for each pair.

So here is a tip for you all.

1.) Set the bot to whatever settings you want to test, and allow it fully optimize itself.

2.)Open your mq4 and search for

Inserted Code
BarCount = SelfOptimize(SymbolUsed);
and you change it to

Inserted Code
//BarCount = SelfOptimize(SymbolUsed);
3.) Copy the optimized files from expert/files and paste them into tester/files.

4.) Run the back-tester in mt4 to see how you would have done.

Things to remember:

Commenting that line out prevents the bot from re-optimizing and forces it to read the files from the last optimization, so if you want to re-optimize you need to change that line back.

If you change the name of the bot, you need to change the name of the files too or it wont use them.

After you set your settings in the initial optimization and copy the files, your back-tester settings don't really matter. It uses the files to know what to do.

If you want to test different settings or with a different pair, you need to re-optimize and copy the files again.

If your modeling quality is "N/A" then you need to recompute your data in the history center.

Hope this helps everyone.
  • #973
  • Edited 7:27am Oct 4, 2010 6:44am | Edited 7:27am
I would like to present a chart of potentiality as well as a little challenge. This is EU tested from 01/01/10 using 5% risk money management settings, starting at $10k.

Can anyone top it?
Attached Image (click to enlarge)
Click to Enlarge Name: EU 01-01-10.jpg Size: 80 KB
  • #974
  • Oct 4, 2010 9:52am Oct 4, 2010 9:52am
Hi rfking2

I was doing the other way around, I mean, I have a 2 years FXT 1min data and i'm doing several tests with different WR/RR/Nr. Bars but as you can imagine... takes ages every iteration.

So, I ask, when you change the code, the bot does not optimize himself during the backtest? and, the bot with that modification can work in forward test?

Thanks,

Pedro
  • #975
  • Oct 4, 2010 3:42pm Oct 4, 2010 3:42pm
rfking2

You made a backtest on the same period as you previously made optimization?
If yes, not surprising that backtest statement is profitable
  • #976
  • Oct 4, 2010 5:31pm Oct 4, 2010 5:31pm
Quoting Pgcoimbra
Hi rfking2

I was doing the other way around, I mean, I have a 2 years FXT 1min data and i'm doing several tests with different WR/RR/Nr. Bars but as you can imagine... takes ages every iteration.

So, I ask, when you change the code, the bot does not optimize himself during the backtest? and, the bot with that modification can work in forward test?

Thanks,

Pedro

Yes, if you are happy with your bots strategy then you can comment this line out and it will work in forward test without re-optimizing. This can be useful to save computer resources. But if you want to change any setting, you must allow it to re-optimize.
  • #977
  • Oct 4, 2010 5:33pm Oct 4, 2010 5:33pm
Quoting LeW
rfking2

You made a backtest on the same period as you previously made optimization?
If yes, not surprising that backtest statement is profitable
No, I did not use future data with that back-test.
  • #978
  • Oct 5, 2010 4:05am Oct 5, 2010 4:05am
Ron

1/ What the WinRate value you set?

2/ What's your real WR (in forward test) [win trades/all trades]

I have set WR = 50, but my real (in forward test) is lower

Waiting for your reply

Regards
  • #979
  • Oct 7, 2010 2:47am Oct 7, 2010 2:47am
Hi Ron et al,

I've noticed that the EA after a few hours working "freeze" in the platform. And this symptom his metatrader broker or version (225 and 226) independent. Also I run it in two different machines (Win Vista and Win 7) and the same result.
I've tried to increase the sleeptime from 100 ms to 500 ms and also using the EA with the modification indicated by rfking2 (commenting on the optimizing at day start line code) and using only 5000 bars.
And still the same final result. The EA freezes after a few hours and the last tick and cycle count numbers are stopped and the computer processor is "working hard".

Anyone has noticed this?

Thanks,

Pedro
  • #980
  • Oct 7, 2010 4:39am Oct 7, 2010 4:39am
Quoting Pgcoimbra
Hi Ron et al,

I've noticed that the EA after a few hours working "freeze" in the platform. And this symptom his metatrader broker or version (225 and 226) independent. Also I run it in two different machines (Win Vista and Win 7) and the same result.
I've tried to increase the sleeptime from 100 ms to 500 ms and also using the EA with the modification indicated by rfking2 (commenting on the optimizing at day start line code) and using only 5000 bars.
And still the same final result. The EA freezes after a few hours and the last tick and cycle...
Pedro

I have had the same problem here. Like you I have tried many different settings etc but still the ea eventually hung.

Paul
1 trader viewing now