stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

Stretchy Slope. Blame Bunker for this one.
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=716		 Page 10 of 14
Author:  ianp [ Sun Oct 07, 2012 8:48 am ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Empty4 61012.gif
Here is a update for the last week, equity at close on Friday 28th Sept was $10532 , at close this Friday 5th Oct was $10711 a nice increase for a EA left to do its own thing !!!
Author:  Bunker [ Sun Oct 07, 2012 11:21 am ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
feucht wrote:Ok week over and the tally is...

11 winners for 770 pips and 2 losers for 260, net 510 pips, 5 trades still open currently about 210 down, so a 300 pip week.

I made a change to my development version to implement the pivot stuff. It took 1 trade straight away last night, a sell on AUDUSD which would need a need a new 2 or 3 month low to hit TP, but it did just before the close. :D I will tidy it up and try and post before the open on Sunday so we can have a play.
This is incredible!

The work and progress that was done with this strategy is just amazing. Talk about pushing an idea to its limit. Although my attention has been on the “Big Mac” strategy lately, I occasionally drop-by here to see how “Stretchy” evolves.

Feucht, your work is absolutely remarkable. We are all benefiting from the dedication and intelligence you are devoting into this. A big THANK YOU to you sir. We’ll have to change the topic subject for:
“Stretchy Slope. Blame Feucht for this one” ;)

Cheers!

Bunker
Author:  Radar [ Sun Oct 07, 2012 11:47 am ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Hey Pipley,
Pipley wrote:Ya, I am real embarassed that I didn't realize that BEFORE I sent the post!!

I am real excited about the possibilites with this. I am wondering, (if it turns out that your information is correct) if the same theory could be stepped down to H4 / M15 stretch. I am just thinking that if a D1 / H1 (24 to 1) ratio works that well, then maybe a 16 to one ration might work also. It's obviously not going to be as good as the original Stretch, but it might be worth checking into for a few pips during ranging days. (?)
Just thinking out loud.....
I modded DWMPivots a few weeks back, so it would show lower TF pivots, (down to the 5M pivot on 1M charts).

Throw it on a chart, and see if it helps with this system.

Have fun!

Radar =8^)
Author:  HamFX [ Sun Oct 07, 2012 6:35 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Hi,

It's nice to see this one going somewhere. I had a similar thought way-back-when but eclipsed me with his fantastic elastic band analogy which explains the concept so well.

top marks to feucht for keeping this going - storing ticks

I notice you are relying on jumping/trailing stops etc for closure and have another suggestion:-

If we are using the "stretch" for opening trades, then why not also use it for closing them?

Easily implemented in LookForTradeClosure() with a couple of one-liners and a new external to define the closure level. 0.1 seems to be a good value to start off with and works better for me than the JS.

Code: Select all

    if (HtfSlopeVal - LtfSlopeVal < StretchCloseDistance) CloseThisTrade=true;
Hope this helps, it must be worth experimenting with.
Author:  feucht [ Sun Oct 07, 2012 6:56 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Bunker wrote:
feucht wrote:Ok week over and the tally is...

11 winners for 770 pips and 2 losers for 260, net 510 pips, 5 trades still open currently about 210 down, so a 300 pip week.

I made a change to my development version to implement the pivot stuff. It took 1 trade straight away last night, a sell on AUDUSD which would need a need a new 2 or 3 month low to hit TP, but it did just before the close. :D I will tidy it up and try and post before the open on Sunday so we can have a play.
This is incredible!

The work and progress that was done with this strategy is just amazing. Talk about pushing an idea to its limit. Although my attention has been on the “Big Mac” strategy lately, I occasionally drop-by here to see how “Stretchy” evolves.

Feucht, your work is absolutely remarkable. We are all benefiting from the dedication and intelligence you are devoting into this. A big THANK YOU to you sir. We’ll have to change the topic subject for:
“Stretchy Slope. Blame Feucht for this one” ;)

Cheers!

Bunker
Thanks Bunker, you really are too kind :oops: I still take my hat off to you for the idea of the stretch. I'm glad to see people are getting some nice results out of this. SS has always had some great weeks, and some awful ones, so we will just have to see how the modifications can swing the balance between them, and realistically it may be months before we can see how this works long term (despite me now having just short of 40 million ticks worth of price and slope data to paw over :shock: ).

It is hard not to end up backfitting the strategy to match the data, so I'm glad that the pivot stuff was a genuine bolt from the blue that seems to have a good effect on results. Attached are updated results up to Friday. Although we had a positive week on new trades generally (not by much), the overall totals are down a bit, which confused the hell out of me for a while until I realised there would have been some trades running since week 37 that finally hit SL on Friday! Inevitably we will always have trades hanging, and I am in 2 minds as to whether to include the current pip count or not. My record of no trades not hitting 90 pips TP has well and truly gone, but I'm glad as that is just cloud cuckoo land stuff!

With the new change I am also wondering if some of the other trade filters are still of any use or not. As slope and stretch and pivot positions are all interconnected, some filters may be doing much the same thing as each other, though perhaps inefficiently. I don't open trades by default with a D1 slope above 0.5, but have been modelling results with this opened up to 1 with the pivot filter in place. I get about another 20%, and interestingly trades taken at a D1 above 0.8 are very few and far between, I imagine it is difficult for D1 to reach that level without the weekly pivot getting pulled above the monthly. I will be defaulting to 1/-1 on HtfMaxBuyLevel/HtfMinSellLevel, TP of 80 and SL of 130.

I will add a parameter to control whether to use the pivot filter or not for those who don't like that idea. I still hope to get a new version out before the open, hangover permitting...
Author:  feucht [ Sun Oct 07, 2012 7:02 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Thanks Ham for the input, this is an interesting idea, and if it is easy to add to the EA, it will be easy to model as well. I have given up on breakeven targets and locks and jumping stops and settled now for simple TPs and SLs (which again are easier/quicker to model). It does make logical sense that what puts you in the trade should take you out again.

Were you thinking of combining the close on a contracted stretch with a TP/SL or it being the only determinant of when to get out?
Author:  HamFX [ Sun Oct 07, 2012 8:18 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
It seems to stand up quite well on its own ( just backtesting using Empty4 strategy tester on EU), adding an SL/TP didn't seem to improve.

It seems to me that when using SL/TP you are more likely to need different values for each pair rather than global values - this approach is more general it seems.

By the way - so far have you notices any pairs that seem to work particularly well, or particularly badly.
Author:  feucht [ Sun Oct 07, 2012 8:37 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Attached is the new EA, haven't been able to run this at all yet, so extra warnings, though changes today are quite small, new param called TradeOnlyOnMisalignedWMPivots, which defaults to true which turns on the filter to make the EA only take trades when the weekly pivot is below the monthly pivot for buy trades and vice versa.

Ham, I have added a breakdown of pair by pair performance (lots of runs added together, so ignore the absolute values). I have started coding up some routines to look at the effect of closing on the contraction of the stretch, firstly on a what hits first basis (TP, SL or stretch hits x), other variants should be straightforward.

Have a good week everyone :D
Author:  Pipley [ Mon Oct 08, 2012 12:08 am ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Radar wrote:Hey Pipley,

I modded DWMPivots a few weeks back, so it would show lower TF pivots, (down to the 5M pivot on 1M charts).

Throw it on a chart, and see if it helps with this system.

Have fun!

Radar =8^)
[/quote]

I will definitely have fun with this! Thanks!
Author:  effluvium [ Mon Oct 08, 2012 1:59 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
feucht wrote:Attached is the new EA, haven't been able to run this at all yet, so extra warnings, though changes today are quite small, new param called TradeOnlyOnMisalignedWMPivots, which defaults to true which turns on the filter to make the EA only take trades when the weekly pivot is below the monthly pivot for buy trades and vice versa.
Hi Feucht,

Just curious: From you regular posts and detail results, you seem have done extensive tests with this EA. Are you using a VPS to let the EA run all days of the week, or do you let your computer running with no interruption ?

How do you manage your Empty4 platform? Are you attaching each set to a different chart , or do you have a single set for each Empty4 platform ?
Could I ask you to post your different sets with your different parameter values you are using actually ? Thanks very much.
All times are UTC Page 10 of 14

Powered by phpBB® Forum Software © phpBB Limited