possible future new EA: "Invisible FX Landscape"

[Batchboy]

RE: possible future new EA: "Invisible FX Landscape"
yep, by me eeee-ven! LOL
After crashing with some EAs I bought and recently even my friend's y-oww I recommended to him (embarrasing!), and since best here is Para_SAR being only one TA in it (and how market change could make it become worse than best), I felt quite completely at loss about hope with EAs.
I have the above plus pipcop site to be thankful for first having this desparation followed by confronting a chart sufficiently ["facing the music"] (5 years experience/knowledge with fx market) to right away plop TAs on it that is looking good for few days, VERY good indeed. I have already contacted Yaroslav of forex-experts.com and Witold of mqlsoft.com for a quote to program this complex EA plan. This is a sophisticated combination of TAs, if it ever loses it's a small amount of pips vs the gainers, and it keeps on top of things with 1Min bars. It takes advantage of retracements on any sort of move that is strong enough and fact that most of the time there is some sort of retracement, but if there isn't 1Min bars and the TAs are fast enough to respond to "not much retracement". It be an extremely rare thing fo hit a 50-pip SL !
It is timely that I re-encountered my favorite New-Age tune by Tom Vedvic "Invisible Landscape" on my XM mini radio upon which I was there to hit record to capture this 17 minute aesthetic number, thus I name this EA "Invisible FX Landscaper" after it.
Thanks to pipcop I leaned about HOW back-tests are bad, thus I have written into the plan necessary points to make sure all TA decisions and SLs are made at the exact bar closing, thus other in-between tick moves won't be part of the picture in optimizing/back-tests.
I have also planned-in start-stop-trades that will allow me in different manner than ever seen before a sliding trade-time window of different sizes hours-wise during optimizations, thus hope to see revelatory most ideal least drawdowntrading/gaining periods of the day!
Thanks to pipcop.com for elucidating me on frailties of back-testing.
I got to say, it's the best combo of TAs & methodology of tradfing I ever seen on a chart.
Any recommendations for this development is appreciated, as you can see above I have tried to think of everything, we want to arrive. Sorry if I can not reveal any further TA & set-up details to avail such help, help in any other way is appreciated.
Thanks, knock-on-wood this continues to be great!
jb

[Batchboy]

I still watch these two indicators, one showing a cross and the other requires two of same color same direction bars. EURUSD_1Min. RSI or RVI as set-up minimum value to show a trend has occured, thus a retracement is most likely in order. I see it works most often, only small losses at times. I've got the quotes and lowest is $180 to have Yaroslav of forex-experts.com program it. I can't afford right now because I over-spent on my previous outrageous EA/FX endeavors, so my financial breathing has been needing rest and recuperation. Give it 1-2 months. But this is good because instead of jumping into it with loot$ in hand I just keep observing this method on chart, so far so good. Whence I run optimizations with sliding different size trade-time windows I'll have notion of best times to have it "look" for trade opportuinities. He says it will take a week to program. Like I say I'll have the SLs done with bar_closing market orders so back-tests match live/demo performance. It never seems to even come near 50-pip stops, method closes out well before then. Who knows, may find 2-4 minis per Thou on account could ber possible, will see.
Whence I have, run, and test, then I can run publishment on a url posted here. Whence that looks good then I hand demo unit to pipcop.
jb

[1ka]

I would offer my services to program this EA, free of charge, but if it's as complicated as you make it sound then I'm not sure I can give it the time it would require to code it. It would be a part time endeavour for me.

I've coded a few EAs and am currently forward testing one here. It's not doing so well

[Batchboy]

Originally Posted by 1ka

I would offer my services to program this EA, free of charge, but if it's as complicated as you make it sound then I'm not sure I can give it the time it would require to code it. It would be a part time endeavour for me.

I've coded a few EAs and am currently forward testing one here. It's not doing so well

I message you to discuss. Thx
jb

[Batchboy]

"We" are still in logic dis-entangling, but just turning-on computer/MT4 at 8pm Pacific_time with this initial simple indicator (v.002, attached) shows a green dot area (trade open) with a 20 pip bagger that ..... seeing later closed at about that much.
jb

[Batchboy]

hi Karl,
(posting this at new EA's thread):
[EA now basically operating]
For the 1week optimization 11:00 hour is actually 3am for my time zone, which is fascinaating that was one of my ideas I wrote last week to you thinking "who knows, maybe it could be 1 hour at 3am!", and Lo, optimization zoomed-in on that! Not that I'm any genius, just common sense that between most of the major times for economic announcements was around then, so I mentally thought of that...."what would be in the middle?" re: things lightly swinging up 'n down between the Econ impacts.
I'm running longish all-time-frames on optimizing and RSI/RVI range for 2-12/2 months, have finally come across something that has 38% gain and 19% DD , but it's not finished running.
Would be nice if something could be for relatively permanent settings, but not holding-out big hopes as month to month and even week to week Econs and market and news all changes all the time, so it's probably going to be daily weekly optmizations, to run certain hour of the nite from those results set in parameters, which makes sense, as who can really say any fully-auto EA is ever going to cover all market conditions, no such thing as holy grail in such modern noisy markets. This is not good for fully-auto EA's of pipcop. In this regard maybe this is flag for change overall in manner of running EA's in reality.
jerry

[Batchboy]

To quote myself from last post:

Originally Posted by Batchboy

Would be nice if something could be for relatively permanent settings, but not holding-out big hopes as month to month and even week to week Econs and market and news all changes all the time, so it's probably going to be daily kind of weekly-optmizations, to run EA at certain hour of the nite from those results set in parameters, which makes sense, as who can really say any fully-auto EA is ever going to cover all market conditions, no such thing as holy grail in such modern noisy markets. This is not good news for fully-auto EA's of pipcop. In this regard maybe this is flag for change overall in manner of running EA's in reality.

I.E., Manual intervention combined with "full-auto", a kind of "hybridly-auto"???
Here's a idea along that line to vote for: Each author run his own EA in his way recommended way on his computer or server. How can there be manual-intervention cheating!? If author does manual adjusts not to the trades betterment then he be cutting his throat. If he does per his own instructions that will be in manual and equity & DD is ok, then it be behooving of what he has overall as an EA. Folks can check authors performance statement-url against their own EA performance-trades, and if different then naturally the author is "on the line".
In this possible future reality mode that trashes old dieing concept/dream of "ooooo, auto is heaven!" the over weight to such site as this running servers for all the EA's would be gone and trust per above would be appropriate. Simply have each author with his own set-up server and MT4 with the publishing tab, such as what I have done for "$18.01 P.SAR EA". Simple, and expensive overload of such as here for EA servers is gone! And the author is running it like interested parties would should they like that EA, ...........and this is the way of operating EAs I have come to be struck with myself, cognizants of Economic-announcements and changing market conditions, etc., despite up 'til now being a hard believer of fully-auto myself........... both fully-autos not anywhere near the dream cracked-up-to-believe, vendors selling/scam us on the dream , and now myself working on being responsible author like Bernard of Excalibur (for example), all pointing now to responsible operation of EA's is the way.

jerry

[Batchboy]

hi folks,
Optimization followed by back-tests show 157 trades at certain hour of the nite with $450 gain in 3 months and only 15% DD. I tried to post the performance html, but upload rejected html's? Anyone with recommendation as to how to post? So in essence it can double an account every 3 months, SHOULD (!!) walk-forward concur.
jb

[Batchboy]

Originally Posted by Batchboy

hi folks,
Optimization followed by back-tests show 157 trades at certain hour of the nite with $450 gain in 3 months and only 15% DD. I tried to post the performance html, but upload rejected html's? Anyone with recommendation as to how to post? So in essence it can double an account every 3 months, SHOULD (!!) walk-forward concur.
jb

ok, so made a page from my server for back-test of best optimized values for past 3 months 1Min data performance page:
http://206.71.180.236/index/Strategy...XL_2_10_09.htm
Bear in mind I took steps to make sure EA "does its doggily duties" only at closing of last forming bar to have back-tests correlate with live, even the stop-loss is done in such a "dynamic SL " way.
Thanks!
jb

[Batchboy]

Making progress folks! Some checks to confirm back-test correlates with wal-forward-live are breing done as well as further 3 month back-tests. Things look good so far. Just want to say that we have a real EA pro programmer aboard this forum, 1ka, w/o whom this could not have been done, he'sa been there throughout my painful feeling momentsd and promptly helped us get thru it.
Of interesting thought is, if author can tell that all calculations are commited at bar's-closing using static values [like this EA uses C(-1) that just became that from C(0) closing, C(-9), & C(-10)] and he uses "dynamic-SL " (bar difference calculated and done per C(-1) and origianl opening price to see if greater than SL ), AND he can show reportts how a week of walk forward is replicated by back-test, then I "think" (??) with pretty good certainty (??) that back-tests are proof enough to bypass 3 month demos (!!) and go ahead with small size live testing, like $100 for excample, which as soon as my further replication tests are done I will check when my finances will aloow $100 for 1/10th mini commitment. This would vastly shorten the time it takes to bring live$$ to a show on pipcop for any EA author willing to demonstrate said info!! Just my considerations after bouncing back 'n forth with ika about.
Anyhoo!!! Onward ho
jb