EA Factory | Page 27 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Calendar as Fundamental Source at Forex Factory 14 replies

Forex Factory Store 0 replies

Journal Factory posters - how much money are you making? 17 replies

Newbie Factory Journal 12 replies

Attachments: EA Factory
Exit Attachments

EA Factory

  • #521
  • Nov 18, 2006 1:27pm Nov 18, 2006 1:27pm
Howard:

Are you speaking in general terms or have you reconciled the backtested results with your manual tested results trade-by-trade?

If not, would you be willing to post your manual backtested results for October (pref. in Excel format) so that someone can run a Strategy Tester backtest to compare against?

I think that would help Aha and others immensely, particularly if there was analysis provided as to the difference. I'd be willing to do the analysis with receipt of the manual backtested trade logs, EA run backtests and clearly stipulated rules document.

To get all of this stuff correlating properly, details details and more details are essential! No surprise as the markets are dynamic!

Quoting howard
Hi

Your result is in line with and confirms my manual backtesting of strategy 1 for the month of October which produced 263 pips, we have to wait for the consultation of aha and micardo to see why they have different results

Please post your results for strategy 2 as well. Thanks

howard
  • #522
  • Nov 18, 2006 7:08pm Nov 18, 2006 7:08pm
Aha,
Thanks for your contribution to this forum.
You have written many experts.
Have you come across any that are profitable, and have been profitable over several years, with low/med risk?

Thanks
  • #523
  • Nov 19, 2006 1:02am Nov 19, 2006 1:02am
Hi FxBabe,

I haven't found a mechanical system that can be profitable in long run. The market situation is always changing, the more I code EAs for mechanical systems, the more I understand the importance of trader's discretion. If you want to be successful in long run, I guess you have to polish your trading skills and find your own system. I don't think there is a shortcut here.... but what I know, maybe I haven't seen enough.

aha



Quoting FxBabe
Aha,
Thanks for your contribution to this forum.
You have written many experts.
Have you come across any that are profitable, and have been profitable over several years, with low/med risk?

Thanks
Target: Trading naked.
  • #524
  • Nov 19, 2006 1:53am Nov 19, 2006 1:53am
Quoting FxBabe
Aha,
Thanks for your contribution to this forum.
You have written many experts.
Have you come across any that are profitable, and have been profitable over several years, with low/med risk?

Thanks
There is a huge misunderstanding in the trading world regarding finding The System, profitable over a 100 years without ever having to be touched again, once turned on. Unfortunately anyone who looks for such a system does not understand that the market is a series of waves that change in width and height over time. Sometimes the surf pounds smaller and harder and faster, and sometimes it's larger and slower and gentler.

Looking for a such a system is like deciding the most efficent vehicle to get from point A to point B(a journey across mountain ranges, rivers and plains) is a car with racing suspension and tires, and one low gear. Well of course it will get there eventually; the racing suspension and tires will help it on the highways and the low gear will keep the car going on the steep terain. But we all know it makes much more sense to give the car several gears, and a more fexible suspension and a set of different tires.

The same goes for your system. It's simplistic to think you'll find one group of settings that will make your system decently profitable year in and year out on the ever-changing waves of the market. So instead of optimizing your system over 12, 24, 36 months, optimize it over 4,3, or maybe even just 6 weeks. And then re-optimize every few weeks to keep your system fitting the market waves most profitably. Your profits will soar while your drawdowns will fall. I promise.
  • #525
  • Nov 19, 2006 5:40am Nov 19, 2006 5:40am
There are elements of adaptive algorithm in my trading system.

I believe a lot of adaptive systems are working far and wide. Were you given such a system you would be coining money. You would hardly discuss profitable adaptive system.

Nevertheless, I believe many of us can build adaptive system themselves.
Prof Herman
  • #526
  • Nov 19, 2006 11:25am Nov 19, 2006 11:25am
Quoting aicccia
So instead of optimizing your system over 12, 24, 36 months, optimize it over 4,3, or maybe even just 6 weeks. And then re-optimize every few weeks to keep your system fitting the market waves most profitably. Your profits will soar while your drawdowns will fall. I promise.
Nice input aicccia (lots of c's in that screen name!). The notion of adapting one's system to every-changing market conditions is definitely sound. While I embrace that concept, I also believe in market cycles which at times can represent abrupt changes from recent history. For example, the Summer months make exhibit a certain market personality which abruptly changes when Autumn hits. Accordingly, if it was the beginning of Autumn and I'm optimizing based on Summer data, it would appear that I'd be out-of-sync with the Autumn cycle.

While I really don't think things are this clear-cut, I do believe in cycles (not necessarily or just seasonal cycles). The question then becomes how do you go about reconciling this with the notion of "smoothly evolving waves based on previous xx weeks of history"? Yes, one might consider waves just another cycle. But, if cyclical study shows that wave x is typically a small wave followed by wave y which is bigger, then, surely optimizing based on wave x would not be "optimal" shall be say.

What are your thoughts on adapting the concept of waves with cycles?

Bill
  • #527
  • Nov 20, 2006 10:46am Nov 20, 2006 10:46am
Well said aicccia. I agree with you. This is why human discretion is important -- a good trader knows when to switch gear.

aha


Quoting aicccia
There is a huge misunderstanding in the trading world regarding finding The System, profitable over a 100 years without ever having to be touched again, once turned on. Unfortunately anyone who looks for such a system does not understand that the market is a series of waves that change in width and height over time. Sometimes the surf pounds smaller and harder and faster, and sometimes it's larger and slower and gentler.

Looking for a such a system is like deciding the most efficent vehicle to get from point A to point B(a journey across mountain ranges, rivers and plains) is a car with racing suspension and tires, and one low gear. Well of course it will get there eventually; the racing suspension and tires will help it on the highways and the low gear will keep the car going on the steep terain. But we all know it makes much more sense to give the car several gears, and a more fexible suspension and a set of different tires.

The same goes for your system. It's simplistic to think you'll find one group of settings that will make your system decently profitable year in and year out on the ever-changing waves of the market. So instead of optimizing your system over 12, 24, 36 months, optimize it over 4,3, or maybe even just 6 weeks. And then re-optimize every few weeks to keep your system fitting the market waves most profitably. Your profits will soar while your drawdowns will fall. I promise.
Target: Trading naked.
  • #528
  • Nov 20, 2006 10:54am Nov 20, 2006 10:54am
We always have the same idea here :-). That's why I stopped coding EAs and started looking at Neural Network solutions. The success of Phoenix in the EA championship proves that a tuned up simple system could profit in short period, for certain pair. And I believe an NN system could do better.
Quoting aha
Well said aicccia. I agree with you. This is why human discretion is important -- a good trader knows when to switch gear.

aha
  • #529
  • Nov 20, 2006 11:38am Nov 20, 2006 11:38am
I agree with you too, Bill. It's possible to design self-adaptive fine-tuned EAs, but considering the time and energy requires, it's hardly doable. We all know market is changing, but how many parameters we need to use to characterize the change of market conditions? what kind of instrument we need to develop to accurately measure the change of each parameter? Then what kinds of strategy should be used at which market conditions? All these questions can be Ph.D. thesis topics and probably worth a lot of years of research... not much simpler than finding the grail.

aha


Quoting billworld
Nice input aicccia (lots of c's in that screen name!). The notion of adapting one's system to every-changing market conditions is definitely sound. While I embrace that concept, I also believe in market cycles which at times can represent abrupt changes from recent history. For example, the Summer months make exhibit a certain market personality which abruptly changes when Autumn hits. Accordingly, if it was the beginning of Autumn and I'm optimizing based on Summer data, it would appear that I'd be out-of-sync with the Autumn cycle.

While I really don't think things are this clear-cut, I do believe in cycles (not necessarily or just seasonal cycles). The question then becomes how do you go about reconciling this with the notion of "smoothly evolving waves based on previous xx weeks of history"? Yes, one might consider waves just another cycle. But, if cyclical study shows that wave x is typically a small wave followed by wave y which is bigger, then, surely optimizing based on wave x would not be "optimal" shall be say.

What are your thoughts on adapting the concept of waves with cycles?

Bill
Target: Trading naked.
  • #530
  • Nov 20, 2006 11:46pm Nov 20, 2006 11:46pm
Sorry for chiming in late guys, but does anyone have clean-running versions of TD Seq. and TD Combo for MetaTrader 4? I've been unable to find them posted here----thread is getting pretty long!
I've been spoiled w/ Bloomberg at work and have no access to the resources at home.

Also, TD ROC....is there a version of that out there as well for MT4?

I'm rather new to the platform but have become very reliant on using the above 3 (particularly Combo and ROC...I find Combo much more reliable in ccy markets).

Is there a website with these indicators professionally scripted and available for download? I saw the site in one of the previous posts but it lacked any of Demark's indicators.

And a side note...I'm a Demark addict if you cant tell and well read on his systems. It's good to see others out there practicing the work.

Thanks! Any help would be greatly appreciated.
  • #531
  • Nov 21, 2006 12:19pm Nov 21, 2006 12:19pm
Hey Billy,

Glad to see you here! We made TD SEQ and Combo but we don't have experts on how to read it. Can you enlighten us a little on that? Also since you have the Bloomberg version, can you check if the indicators here are accurate?

Here are the indicators I made.

aha


Quoting BillyRayValentine
Sorry for chiming in late guys, but does anyone have clean-running versions of TD Seq. and TD Combo for MetaTrader 4? I've been unable to find them posted here----thread is getting pretty long!
I've been spoiled w/ Bloomberg at work and have no access to the resources at home.

Also, TD ROC....is there a version of that out there as well for MT4?

I'm rather new to the platform but have become very reliant on using the above 3 (particularly Combo and ROC...I find Combo much more reliable in ccy markets).

Is there a website with these indicators professionally scripted and available for download? I saw the site in one of the previous posts but it lacked any of Demark's indicators.

And a side note...I'm a Demark addict if you cant tell and well read on his systems. It's good to see others out there practicing the work.

Thanks! Any help would be greatly appreciated.
Attached File(s)
File Type: zip TD SEQ and COMBO.zip   49 KB | 2,024 downloads
Target: Trading naked.
  • #532
  • Edited Nov 22, 2006 12:31am Nov 21, 2006 10:53pm | Edited Nov 22, 2006 12:31am
Good Work! I've just taken a first glance at the Combo. I read in one of your earlier posts that it's still in somewhat early stages. The "Setup" works well and appears to be clean. The countdown, and I speak very understandingly, might use some work, primarily when it comes to that which is probably the most complex component of the indicator - the recycle.

I'm very curious to hear how you have handled them...personally I hate dealing with strategies surrounding them (they're tedious!). It seems the largest difficulty with the indicator is this issue.
I'm unsure as to how metatrader will be able to digest the various adjustment paramaters. What do you mean by "Recycle Based on Range" btw....? Also, TD Termination Count - is it taken into account?

And as a novelty but still very helpful, Bloomberg will draw dashed TDST lines from the highs or lows of the #1 Bar of setup....if you are so inclined.

I'll see what I can gather on Bloomberg and get back to you with some side by side comparisons.

Thanks!!!!!!!!!!!
  • #533
  • Nov 23, 2006 1:25pm Nov 23, 2006 1:25pm
This is an indicator plots previous day's high/low on today's chart. It is a useful tool for testing daily range systems. Place it on your 1M or 5M chart you will be able to see when the price penetrate the high low range, how far it goes, all those information.

Enjoy!
Attached File(s)
File Type: mq4 DailyHighLowHistory AHA 0_1.mq4   3 KB | 1,467 downloads
Target: Trading naked.
  • #534
  • Nov 24, 2006 2:01pm Nov 24, 2006 2:01pm
Hi aha,
Thanks for the info. I downloaded a bunch of screenshots from Bloomberg today for you. When I went to go and upload them onto the thread it said that they were too large to do so. I don't want to shrink them as I might lost some important detail. I've got about 10 or so of TD Combo and Sequential, and the indicator adjustment window screenshots listing all of the adjustment options. If you have an email address you would like to send me via private message I will gladly send them to you.

I did notice some discrepancies with your Combo and default settings in Bloomberg. As I suspected they pertained mainly to the countdown phase. I'm unfortunately under time constraint right now but I can give them a look a little later and see what might be causing these differences. At first glace, the price bars recorded in Bloomberg vary slightly from the ones provided by Interbank FX (they make slightly higher highs, etc)...as would be expected and this would vary from broker to broker. This would nevertheless mean the difference of recording a # or not.

Regarding TD Termination Count, I'll write a little later. It's actually quite simple. DeMark's second book "New Market Timing Techniques" is probably the most comprehensive on these topics, and his option trading book is the most recent and perhaps the most reliable, though most of it is repetition.

One thing I wanted to say was that I noticed some people earlier in the thread were questioning "false signals" given off by these indicators (a 13 was recorded but price "hiccuped" and kept on going). The signals were all actually accurate. Entry techniques on these signals need to be known by the trader well in advance of using the indicator as a profitable means of trading. Because when they fail, they really fail, and opportunity is born to reverse the position and shoot in the direction of the prevailing trend. If it were as easy as reading 13 and going north or south, I would own half of Manhattan by now.

Speak soon,
BRV
  • #535
  • Nov 24, 2006 3:42pm Nov 24, 2006 3:42pm
Thanks Billy! I have sent my email address to your PM indox. I guess one thing on my todo list is to debug the TD-Combo indicator. Then I may add alert mechnism based on entry technique you mentioned (may need to ask your a lot of questions in the process).

Do you skype? Maybe we can chat online sometime.

aha
Target: Trading naked.
  • #536
  • Nov 24, 2006 4:12pm Nov 24, 2006 4:12pm
Thought this might be of use to others...email to aha

Regarding entry techniques for TD Sequential and Combo:

"1. Enter on a '13'
2. 2 lower lows/lower closes followed by higher close/higher high for BUY (TD POQ) - reverse for sell
3. A close higher or lower than the close 4 bars ago (price flip)
4. A combination of confirmations: a. REI or ROC broke ob/os zone downward or upward, b. a qualified TD line break, c. the Combo or sequential count d. heavy selling or buying pressure breaking previous zone of bars (but no more than 6 usually) and/or TDPOQ

By programming them in you're suggesting that it has to be a certain way. People using this indicator aren't your "4X Made Easy" crowd....they are literally making billions upon billions every year using this guys' indicators (trust me I know...there's a reason I study this guy's work so much). There's a lot to digest with this guy, but in the case of option 4 above, you've got an extremely high-probablility trade - you gotta be totally out of it not to make money in a case like that!"

Happy trading,
BRV
  • #537
  • Nov 24, 2006 4:16pm Nov 24, 2006 4:16pm
And another...message to aha might be of use to others

"Hi there...just posted a reply on the thread. ROC I do use actually as I find it to be one of his better indicators. It tracks the raw price movement and clearly identifies when things are abnormal vs. 12 or 8 bars ago....sort of an arithmatic standard deviation, but much clearer. 100.25 and 99.75 are used for the over bought and oversold areas. Probably the best (in my opinion) for identifying change in direction of price. The DeMarker I stay away from (things didn't work out!), but the REI I will use for pin pointing entries in pullbacks in up or downtrends (I can usually nail the tops or bottoms perfectly if conditions are right) . Here's an example, but I try to identify the trend through other means and not a moving average (I typically don't use moving averages too often) http://www.trading-naked.com/efuturevision3.htm (I'll use a series of 4 for this method).

Hope it helps,
BRV"
  • #538
  • Nov 24, 2006 4:53pm Nov 24, 2006 4:53pm
I am trying to make an indicator that marks the bars in the history that is the market opening bar. But I am not sure if I have those market opeing hours right. The websites I found through google seems to give me contradicting results. Could anyone verify this for me? Thanks!

All time below are for GMT-0 time zone.

Tokyo: 0:00
London: 8:00 in winter, 7:00 in day light saving time
NewYork: 13:00 in winter, 12 in day light saving time

aha
Target: Trading naked.
  • #539
  • Nov 24, 2006 5:46pm Nov 24, 2006 5:46pm
Hi Billy,

I did a quick check on those charts you sent to me and most difference I found were due to different data feeds. The default setting of my indicator is to recycle based on range, which is recommended by TD, while your charts seems to use the first complete setup for countdown without considering the setups finished before the combo countdown finishes. I don't remember TD recommend to do that. If you change the recyclebasedonrange in my indicator to false, it will use the latest completed setup for combo countdown, the unfinished countdown started from the old setup will be deleted. This is also recommended in TD's book.

I guess I will spend sometime studying the TD-REI or ROC. It would be great if you can send me some charts (or post it here) on how you use the option 4 in your post to take trades.

Another thing I don't understand is those lines. Some are dashed, some are continuous, some are green, some are red. Are those lines for stoploss or take profit?

Thanks,

aha
Target: Trading naked.
  • #540
  • Nov 25, 2006 1:36am Nov 25, 2006 1:36am
Hi guys,

Here is the DeMark REI indicator v0.1. Enjoy!

Billy: let me know if you see something wrong.

aha
Attached File(s)
File Type: zip DeMark REI AHA 0_1.zip   4 KB | 2,810 downloads
Target: Trading naked.
0 traders viewing now