Stochastic EA - page 10

To add comments, please log in or register
sfindlay
162
sfindlay  
alpine4133:
Do you still use the Stoch EA?

Which pairs and time frames did you find to work the best?

Do let it close the order or do you close it manually?

I'm using it on both demo and live, just the major pairs at this point, both 15 min and 30 min. The image in my last post gives you my summarised results over 12 days of trading.

The 15 min trades more often but you need to watch the exit point. The 30 minute trades much less but is much higher probability.

I've been playing with the 15 min code trying to increase the probability of its exit but no luck yet...........

sfindlay
162
sfindlay  

Version turning off auto exit option

5

sfindlay:
I have been helped so much by this forum I thought it was time to give something back. THis is my latest EA effort, which is based on the stochastic turning up (or down) after reaching 80 (or 20). It autocloses when the Williams %R hits an extreme point, generally at a profit, although if you want you can set take profit, stop loss and trailing stop loss.

For some reason it trades a lot more in forward testing than in back-testing, and makes a whole lot more money - I have been forward testing for over a month now and it hasn't missed a beat. I would love someone to take it and improve it with money management etc, my skills are not that great.

It doesn't do well in backtesting before 2008 but I figure the personality of the pair has changed anyway and no EA is forever for that reason.

Best results so far are EURUSD 30min, with default settings. I use it on IBFX.

I have uploaded Version 2 with better trading defaults, also i added in the option to turn off the autoexit if you want to exit manually for example. This EA auto closes if the Williams %R is <-3 (close buy) or <-95 (close sell) so if you are manually trading also it will close your manual trades, which is a pain. If anybody has the programming expertise to fix this please go ahead!!!

Thanks

Files:
fxnelson
3
fxnelson  

Close problem - not so hard to fix

sfindlay:
5

I have uploaded Version 2 with better trading defaults, also i added in the option to turn off the autoexit if you want to exit manually for example. This EA auto closes if the Williams %R is <-3 (close buy) or <-95 (close sell) so if you are manually trading also it will close your manual trades, which is a pain. If anybody has the programming expertise to fix this please go ahead!!!

Thanks

It's not too hard to fix the problem. In the code that handles the close commands, also check for the magic number of the currently selected order to be the same number as the MagicNumber parameter of the EA. The number should be something else than 0, though. If you've come this far, you can do it; the magic number is already attached in the order send call to the server, so when you look at existing orders it's there for you to use.

John West
32
John West  

Thanks for uploading this ea, i works great! I've been back testing this ea (haven't tried it on a demo or live account yet). I noticed that the largest losing trades were the ones that were trading against the trend. Although you can turn a profit on a counter trend move, what I have been doing is setting the "long or short" property on the ea to match the trend. In other words, I check the trend on a daily chart, then I set the long or short only property. If I cant identify a trend, I keep the settings default, so far the results are pretty good. Thanks for the ea!

LightampMagic
49
LightampMagic  
johnw213:
Thanks for uploading this ea, i works great! I've been back testing this ea (haven't tried it on a demo or live account yet). I noticed that the largest losing trades were the ones that were trading against the trend. Although you can turn a profit on a counter trend move, what I have been doing is setting the "long or short" property on the ea to match the trend. In other words, I check the trend on a daily chart, then I set the long or short only property. If I cant identify a trend, I keep the settings default, so far the results are pretty good. Thanks for the ea!

I agree with your back test results.Have tested on demo and got gr8 results. And yes setting the trades in the direction of trend is the way to go. An additional parameter could be to close all trades if the floating profit has reached the target amount u want to make in a day.

Sfindlay is that possible ?

John West
32
John West  
Light&Magic:
I agree with your back test results.Have tested on demo and got gr8 results. And yes setting the trades in the direction of trend is the way to go. An additional parameter could be to close all trades if the floating profit has reached the target amount u want to make in a day. Sfindlay is that possible ?

Light&Magic, did you try all the currency pairs? Which ones do you think are more successful?

LightampMagic
49
LightampMagic  

Currency pair

Hi

I have tried it on 16 currency pairs and gold and silver at the same time with great results...then a glitch in the ea blew my ac. This error is now fixed by sfind in the new ea

However i believe for best results test it on EU. This ea has alot of potential

an additional filter when not to trade/when to trade is what we need to work on as a group and of course money management which the ea handles well .

L&M

Geolimber
9
Geolimber  

???Question 2 Autor???

What do you think about no fixed stoploss, stoploss by last Fractal?. And trailing stop by fractals.

bcmfxtrader
47
bcmfxtrader  

anyone having probs with duplicate/multiple consecutive orders?? i.e. it will enter and exit simultaneously for half an hour or so...chewing up equity by losing the spread on every "in and out"... read where someone is try on 16 different pairs, gold and silver... i seem to only have this problem with non-majors. ueor/gbp/chf, and jpy all seem fine. all others either dont trade, or get that multiple order thing going...

anyone else have this problem??

btw ... read back a few pages about "fewer trades but more consistent winners the higher the timeframe" ... thus far with me, that is TRUE. doesnt trade nearly as much on 30m and 4hr timeframes...but very nice equity curve. i've completely dropped the lower 15m timeframe..

let me know if anyone having a prob with the duplicate orders, as well as a possible fix...

thanks !!

sfindlay
162
sfindlay  
bcmfxtrader:
anyone having probs with duplicate/multiple consecutive orders?? i.e. it will enter and exit simultaneously for half an hour or so...chewing up equity by losing the spread on every "in and out"... read where someone is try on 16 different pairs, gold and silver... i seem to only have this problem with non-majors. ueor/gbp/chf, and jpy all seem fine. all others either dont trade, or get that multiple order thing going...

anyone else have this problem??

btw ... read back a few pages about "fewer trades but more consistent winners the higher the timeframe" ... thus far with me, that is TRUE. doesnt trade nearly as much on 30m and 4hr timeframes...but very nice equity curve. i've completely dropped the lower 15m timeframe..

let me know if anyone having a prob with the duplicate orders, as well as a possible fix...

thanks !!

I had the same problem last week, when trading live, the first real significant loss but on USDJPY (15 min) that took out my live "test" account. (I test on real money now since I have found that I get different results parallel testing on demo and live accounts.) I was Not Happy.

I looked at the code, and it shouldn't have happened.

I back tested and couldn't duplicate results, any which way.

I think it is because the code was not looking at the magic number before closing the trade. So, while my USDJPY 15 mins chart had a sell signal, and was going down, down, down, the USDJPY 30 min chart was in a close sell trigger. So, it kept closing my 15 min chart, and then my 15 min chart kept re-opening the trade because it was still on the 15 min candle and in a sell trigger. It didn't take long for slippage and the spread to kill my account.

So I've fixed the code so that it looks for the magic number like fxnelson suggested.

The other way to fix is to turn off "Every Tick". (Turn to False) But this does appear, on backtesting, to reduce the profitability.

If any of you have better coding skills than me, I'd love it to look at the equity 24 hours earlier, and kill the EA if the equity has reduced to a predetermined level.

Files:
To add comments, please log in or register