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 13 of 14
Author:  feucht [ Thu Oct 25, 2012 8:18 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
And duh, I will also need H4 slope values, I'm hard of thinking this week :shock:
Author:  HamFX [ Thu Oct 25, 2012 10:51 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
feucht wrote:There may be a version of TMASlope that gives you what it was at the time, rather than the standard one that can change a bit (repaint) after the fact. I'm sure that there was some discussion about it in the early 10.2 days. If I could get "real" M15 slope values even at a M1 or M5 resolution (a single value per M1 or M5 bar), that might be enough to get some simulations rolling...
feucht, you can find what you are looking for here, in Baluda's TMASlopeMTF indicator http://www.stevehopwoodforex.com/phpBB3 ... f=45&t=629

Look for CalcTmaTrue.
Author:  Pipley [ Thu Oct 25, 2012 10:53 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Great news, Feucht! Thanks!

I had three A/N trades fire in the last three days. The first resulted in a gain of around $50 (using .1 lots), the second lost a little more than $50, and the third is currently up about $25. Other than that, I've only had a G/A fire which is currently down $40.
So, since I added Feucht5, I have had 5 trades fire. All of them had AUS on one side. I am assuming that it is the market, but I wanted to point that out in case there is more to it than just the market.

Thanks again for your hard work!!
Author:  feucht [ Sun Oct 28, 2012 9:39 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Another week about to kick off, I've finished analysing results up to Friday. I'm quite struck by the idea of splitting the targets for contraction of the stretch for when in profit vs in drawdown. With the attached updated results for this, you can see quite a strong pattern here. If we compare first of all week on week at -0.4 close for both (almost the default for SSF5, I went for -0.35) we get an overall haul of 1871 vs 1763 , so up a good 100 pips on the week. We can easily improve on that by closing losing trades at -0.1 (the green vertical stripe). I have looked at some deeper contractions this time, and we could have squeezed out a bit more by letting profitable trades run to anything up to -1. Up to -0.4, we would only have seen 358 pips in losing trades, at -0.7 to -1.2 that goes up to -669, so you have a bit of a choice as to whether you go for caution or pips.

I have coded up a new version of the EA to support this split closure model. There are 2 new params to replace CloseOnContractionTarget, CloseOnContTargetInProfit (defs to -0.7) and CloseOnContTargetInDD (-0.1). If you want to stay with a single contraction target, then obviously set them both to the same value. I have just knocked these up and all I know is that they compile, so caution advised as ever.

Ham, thanks for the heads-up, I had managed to find the exact routine you mentioned, I have history for M15 and H4 true slopes going back to the first half of the year. It is a little tricky to match the times up to my tick data, I've done it for the M15, but need to see some live data tonight/tomorrow to make sure the H4 is right, I am logging H4 tick-by-tick as of open tonight, and have been doing the same with M15 for a couple of days. Pipley, I hope we can get some rough idea of whether shorter TFs might work. At least with close on contraction we hopefully have a universal way of finding an exit, as otherwise we would have to look hard as scaling down SLs and TPs to suit smaller/quicker moves.

Good luck!

J :D
Author:  bshoe24 [ Mon Oct 29, 2012 12:05 am ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
feucht wrote:....Ham, thanks for the heads-up, I had managed to find the exact routine you mentioned, I have history for M15 and H4 true slopes going back to the first half of the year. It is a little tricky to match the times up to my tick data, I've done it for the M15, but need to see some live data tonight/tomorrow to make sure the H4 is right, I am logging H4 tick-by-tick as of open tonight, and have been doing the same with M15 for a couple of days. ...
Good luck!
J :D
feucht,

if another reasonably good broker's data would help you in anyway you can get tick data for Empty4 to backtest for free from MBTrading back to Jan 2011. if you sign up for free account on their website (for demo creation) you can download tick data 1 month at a time per pair pretty easily. use the tools and directions from this website to prep the data >

http://eareview.net/tick-data/download-free-tick-data
http://eareview.net/tick-data/convert-tick-data
http://eareview.net/tick-data/tick-data-backtesting
Author:  feucht [ Mon Oct 29, 2012 1:27 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Wow, bshoe, wish I knew that eareview existed before I started on this journey, what a great site! Now I have got a premise (on D1/H1 at least) that I am relatively happy with, it would be great to be able to backtest it over a bigger dataset, and look at H4/M15 as well.

Thanks!
Author:  bshoe24 [ Mon Oct 29, 2012 5:56 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
feucht,

i did this recently and it was pretty easy. here are the basic steps if i recall.

use a separate Empty4 terminal as it's history will get overwritten in the final step.

1. create account at mbtrading
2. download all the individual tick data files for a single pair to a working folder. you'll have lots of individual files. i did all of 2012 for EU first but, have since done AU and GU for all of 2012.
3. concatenate the files together in proper order (i used http://www.jddesign.co.uk/products/concat/concat-s.htm recommended on website). when using this utility be sure to select them all but, right click on the very first individual file and it should order them correctly. double check to be sure.
4. move the concatenated single csv file to Empty4/experts/files for the CSV2FXT script to pick up
5. use the website's custom Empty4 script CSV2FXT to convert the CSV file to FXT history files (monitor EXPERTS tab for script feedback)
6. at the end the script will ask to move the converted history file to history folder overwriting history
8. close and reopen terminal

i hope this helps you or someone else. it's rare i get to contribute something valuable :)
Author:  Pipley [ Tue Oct 30, 2012 2:57 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Hey! Like bshoe, I never have a chance to contribute, but this might give me an opportunity.(?) I won't have time to investigate (or help) until this weekend, but if I can go through the steps that bshoe outlined and produce some needed data, I will be glad to do it. Will I be able to compress the files enough to send via email?
Author:  bshoe24 [ Tue Oct 30, 2012 5:21 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
Pipley wrote:Hey! Like bshoe, I never have a chance to contribute, but this might give me an opportunity.(?) I won't have time to investigate (or help) until this weekend, but if I can go through the steps that bshoe outlined and produce some needed data, I will be glad to do it. Will I be able to compress the files enough to send via email?
Hi Pipley,

Because having to download the files individually at MBTrading (you can grab a month at a time worth) it's a bit tedious to get a single pair prepared fully. Coincidentally this morning i was thinking it would be nice to have an Empty4 that had been built out with history for most of the pairs. If one had the motivation to do that and was willing to share they could zip up the Empty4 folder after they were done and share on a big file sharing site (there are many just google). You'd want to remember to remove the really big CSV file(s) you staged for importing in /files before zipping anything. Also perhaps deleting any logs. I just checked my Empty4 and it was only about 50MB for the history files for AU,GU, and EU for data back to Jan. 1 2012. That's roughly 20MB per pair so 20 pairs would make the zip maybe up to 400-450MB (before compression) you'd be sharing for data back to Jan 1 2012. Here is something i found in a search that lists the online big file sharing websites http://mashable.com/2007/05/23/7-ways-t ... han-gmail/. If you have the bandwidth to seed it for a short while you could put it onto a torrent site too. I even have a server myself sitting on very high speed connection i could share it from semi-privately with forum members. If someone did build out an Empty4 with all history it could be shared. Where there's a will there's a way.

There is another alternative for the really determined backtester who wants an easier way. I saved this link once before that someone posted as a reputable inexpensive data supplier. This is the least expensive vendor for data i seen before but, it's a very large amount of data going back many years. http://disktrading.is99.com/disktrading/#FOREX. I don't know who (what broker) is the source of their data though.
Author:  feucht [ Tue Oct 30, 2012 7:48 pm ]
Post subject:  Re: Stretchy Slope. Blame Bunker for this one.
I downloaded 23 pairs worth of tick data from Dukascopy going back to 1/1/11, about 40 GB worth! Following the instructions on eareview, I have managed to test a couple of pairs (both losers!) so far going back to then. Will let you know how they all pan out. Thanks for all the help all :)
All times are UTC Page 13 of 14

Powered by phpBB® Forum Software © phpBB Limited