Basket 14 EA System | Page 77 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

BASKET TRADE Management: Using multiple trade management EA's on basket trades 11 replies

Basket Knights - A Basket Trading Round Table 1,313 replies

Relative Strength Basket Trading System 3,479 replies

Simple Basket Trading system 61 replies

TMS basket: Dashboard for a basket of MA-Filtered RSI Signals 135 replies

Attachments: Basket 14 EA System
Exit Attachments

Basket 14 EA System

  • #1,521
  • Oct 29, 2010 3:51pm Oct 29, 2010 3:51pm
Quoting frazerd
He is joking..
No I wasn't!
Open to new approaches.
  • #1,522
  • Oct 29, 2010 4:59pm Oct 29, 2010 4:59pm
Quoting tazmaan
yep, abandoned that one quite awhile ago, and it was MY mod lol
--
Tazmaan,

Are you going to leave us alone too? I know you can handle it. You have been one of the inner circle in the development of NB/TB (Bob and Steve) and without any doubt contributed a lot in the thread. With that kind of experience I believed you can be of help in the conclussion of this endevour.

Sincerely,

Edd
  • #1,523
  • Oct 29, 2010 5:34pm Oct 29, 2010 5:34pm
A theoretical question:

If we are looking at whether "phantom trades are profitable" when they opened at a given time...then isn't that the same as looking at the difference in price between "now" and "then"?

So conceivably, if at any time "now" we could compute "when was the then", we don't need to store and save in global variables the opening of the phantom trades, we just have to compare prices now vs then.

The strategy is a "trend follower" looking at unsmoothed deltas between now (top of every hour) and certain mark points, and opening trades in the pair & direction of the strongest trend. Real time close happens any time basket profit is sufficiently large.

Is there something I'm missing?

If not, then perhaps the code infrastructure for storage & management of the phantoms are superfluous if one can look up previous prices directly.
  • #1,524
  • Oct 29, 2010 8:55pm Oct 29, 2010 8:55pm
Quoting Edd Ganuelas
--
Tazmaan,

Are you going to leave us alone too? I know you can handle it. You have been one of the inner circle in the development of NB/TB (Bob and Steve) and without any doubt contributed a lot in the thread. With that kind of experience I believed you can be of help in the conclussion of this endevour.

Sincerely,

Edd
Nope, I am used to all the stuff that goes on on this forum I am not going anywhere
Whether that's good or bad, I have no idea (for anyone) lol.
  • #1,525
  • Oct 29, 2010 10:10pm Oct 29, 2010 10:10pm
Quoting mbkennel
A theoretical question:

If we are looking at whether "phantom trades are profitable" when they opened at a given time...then isn't that the same as looking at the difference in price between "now" and "then"?

So conceivably, if at any time "now" we could compute "when was the then", we don't need to store and save in global variables the opening of the phantom trades, we just have to compare prices now vs then.

The strategy is a "trend follower" looking at unsmoothed deltas between now (top of every hour) and certain mark points,...
if you dont store the 'then price', how do you get it ? one way is to store the start time and then go back using 1 min TF and get the close price of that 1 min bar. but from coding perspective its easy to store them in global variables to avoid calculting 'then price' every time. in coding world, performance is given equal importance, thats why store the calculated values somewhere and reuse instead of doing calculations every tick
  • #1,526
  • Oct 30, 2010 2:24am Oct 30, 2010 2:24am
Quoting roundrock
if you dont store the 'then price', how do you get it ? one way is to store the start time and then go back using 1 min TF and get the close price of that 1 min bar. but from coding perspective its easy to store them in global variables to avoid calculting 'then price' every time. in coding world, performance is given equal importance, thats why store the calculated values somewhere and reuse instead of doing calculations every tick
I guess I don't understand the difference. Once an hour you calculate delta (now-then), it just changes where you get the "then price" from---a global variable or MT4 history. I don't think there's a huge amount of difference---other than saving a bunch of administrative code. Am I correct, or not?

On the newest versions which reset every 8 hours won't the start times be nearly always on-the-hour, in which case you could use the the opens of 1H bars as a "then" reference point?

In fact, for me the main reason I'm thinking this way is generalizing and understanding the method: it's just trend-following, and so there may be even better statistics than what is currently being used.

I see the big trick is to consider the multiple pairs together and select trades based on relative strength across currencies, not just each pair independently against some arbitrary "is trending" threshold.

But first, let's get confirmation that my hypothesis about what it is doing is correct, i.e. I'm fully understanding the implications of the strategy.
  • #1,527
  • Oct 30, 2010 2:31pm Oct 30, 2010 2:31pm
"On the newest versions which reset every 8 hours won't the start times be nearly always on-the-hour, in which case you could use the the opens of 1H bars as a "then" reference point? "

Since I dont know squat about coding, I guess you are asking if there is an easier way to code the phantoms. That I dont know. But you are correct in that you are looking at difference in price from start of phantoms to the current price. You would have to ask MC but I dont think the phantoms start on begining of one hour bar. They start when you start up the EA or when there is a equity pop. As for the 8 hour reset, Dont know if its the top of the hour or 8 hours from when the EA started.

As for the stategy you are correct, it is very short term trend trading (from start or reset of phantoms). When it works well you will get a equity pop just about every day or twice a day. The problems arise when the markets get choppy. Thats where the concept of the synthetic hedge comes in. If you have long trades for ex. in EUR/JPY and the market turns short the ea will then probably start to take shorts in CHF/JPY thus somewhat balancing out your exposure. I think the trick is to keep exposure somewhat balanced enough without taking too much drawdown until a few pairs move far enough and long enough to give you an equity pop. At least thats my take on it.

All the testing and new versions now I believe are looking for the best ways to avoid those big drawdowns and get quicker equity pops.
  • #1,528
  • Oct 30, 2010 3:34pm Oct 30, 2010 3:34pm
Quoting mbkennel
I guess I don't understand ...
In fact, for me the main reason I'm thinking this way is generalizing and understanding the method: it's just trend-following, and so there may be even better statistics than what is currently being used.
Having read both your posts, I think your opening statement is correct, you don't understand and you need to go back to the beginning of the thread and start reading.

These EAs do not use statistics or trends. The phantoms are trading in real time and providing real time information on which pairs are the biggest winners.
Frazer
  • #1,529
  • Oct 30, 2010 6:46pm Oct 30, 2010 6:46pm
The account is a 200:1 Standard.

The settings are 0.67 for MaxMarginToUse and 0.5 for ProfitPct.

http://bobthoughts.net/Phantom_AddSix_03b_TADAWUL.htm
Open to new approaches.
  • #1,530
  • Oct 30, 2010 10:50pm Oct 30, 2010 10:50pm
Closing every day.. I love it MC.

I think the lower ProfitPct is a good move.
Frazer
  • #1,531
  • Oct 30, 2010 11:30pm Oct 30, 2010 11:30pm
Quoting frazerd
Having read both your posts, I think your opening statement is correct, you don't understand and you need to go back to the beginning of the thread and start reading.

These EAs do not use statistics or trends. The phantoms are trading in real time and providing real time information on which pairs are the biggest winners.
I have downloaded and looked the source code and am running it, and I've read a significant part of the thread.

Is this correct (please correct as necessary)

At top of each hour, the bot looks at the difference in price at present vs price of the same pair at the saved mark price (the phantoms). Those saved prices are stored in MT4 persistent global variables. The robot computes from all pairs on the list "which one would have been the largest winner", i.e. the hypothetical profit difference assuming a fixed lot size (not sure about) trade would have been entered in each pair. Half are designated to be always buys and another half always sells. The robot enters a real trade in the same direction and in the same pair which is the one with the greatest win among the hypothetical trades.


When I say "statistic" I'm using it in the technical sense that any concrete function extracted from observations is called a "statistic", i.e. an actual number computed from data as opposed to a parameter of a theoretical unknown distribution. In this case the hypothetical profit which is a multiplicative factor times the price difference is what I call the statistic.

I call it "trend following" because once there is a hypothetical winner, meaning the price has changed in a certain direction by the maximum amount among pairs, it places a trade in the same direction on the same pair under the assumption that the price is more probable to continue in the same direction. I call this "trend following" because it profits when there are positive autocorrelations in price differences (between now and the phantom---i.e. the market moved in a certain direction, and the now and the future when real trade is active, in the same direction).
  • #1,532
  • Oct 31, 2010 12:05am Oct 31, 2010 12:05am
Well mb, you are obviously a hell of a lot smarter than I. I can hardly pronounce some of the words you use and therefore I have no idea what you are talking about.

The phantom trades are real time trades and yes, half are always buys and the other half sells - they change with every tick of the pair. At the top of each hour the EA looks at the prices of the phantoms, which are shown in your screen comments too, and places a real trade of whatever pair is the biggest winner at that exact moment. Then when the trades come into profit it closes all trades when the equity reaches ProfitPct above Last Flat Equity (the last time all trades were closed).

That, in a nutshell and in my Tennessee hillbilly language, is how the EA works. If that is what you said above, then you have it right.
Frazer
  • #1,533
  • Oct 31, 2010 1:11am Oct 31, 2010 1:11am
he must be good at scrabble, hehehe.
  • #1,534
  • Oct 31, 2010 1:27am Oct 31, 2010 1:27am
Quoting 2+2=4ex
he must be good at scrabble, hehehe.
No, not that great.

But I do mathematical software & machine learning for a living though, so I probably used some confusing jargon without realizing it.
  • #1,535
  • Edited 4:33am Oct 31, 2010 4:32am | Edited 4:33am
Hi all,

There seems to be a bit of confusion from users about settings and a bit of agro/frustaration from the designers of this system towards those users as it is perceived they are not reading enough.

Could I try and put into perspective. I am not a coder but I do have an average understanding of what basket trading is about and I like it.

I have read the whole thread, some parts twice, some parts three times. I still have difficulty with the settings. I am having good profits then large DD.

My accounts are as follows. A standard 100:1 demo with $10,000. It can trade lots from 0.01 up. I have a live account that I want to use for trading this ea. It is a standard 100:1 but I only have equity of about $1200. It can trade lots from 0.01 up.

I cannot work out what the best settings are for these accounts. I want run both of them and once they (and I through true understanding) both perform consistently I can then transfer over to my other live account which is where I do my serious trading.

I am now using the Phantom_AddSix_02a version. Maybe a simple (for non coders, not the very clever coder people) explanation of which version does what and the best setting for each of them for the different account types. Put it in a word document and if a new version or modification is made, update that portion of the document and attach a prefix which is the date (281010) so everyone can easily establish the latest document.

What appears to be happening in this thread is no different to many threads, as they become larger and different version of this and that start to evolve, we lose track of what is happening and the forum folds.

If the latest versions are updated and placed on the first page of the thread, then everyone knows where to get them rather than trolling through multiple attachment pages.

The coders of this system have done a brilliant job, I believe MC should have power of veto about which versions are placed on the first page and if anyone wants to modify his stuff (without his endorsement) they should start a new thread.

Is it too much to ask? Some of us are not computers and as such we have to punch the info in more than once.

Comments agreeing or disagreeing most welcome.

Regards.
  • #1,536
  • Oct 31, 2010 4:34am Oct 31, 2010 4:34am
Quoting aajduk
Is it too much to ask? Some of us are not computers and as such we have to punch the info in more than once.
  • #1,537
  • Oct 31, 2010 10:08am Oct 31, 2010 10:08am
Quoting MetaCoder
The account is a 200:1 Standard.

The settings are 0.67 for MaxMarginToUse and 0.5 for ProfitPct.

http://bobthoughts.net/Phantom_AddSix_03b_TADAWUL.htm
What has your lowest margin % been, MC?
Frazer
  • #1,538
  • Oct 31, 2010 4:04pm Oct 31, 2010 4:04pm
Quoting frazerd
What has your lowest margin % been, MC?
352.17% on the Tadawul account.
Open to new approaches.
  • #1,539
  • Oct 31, 2010 4:08pm Oct 31, 2010 4:08pm
Quoting aajduk
I cannot work out what the best settings are for these accounts.
aajduk, here are a couple of pretty good guidelines I believe.

If the floating loss goes below -10%, or if the lowest margin goes below 500%, it means your MaxMarginToUse setting is too high.
Open to new approaches.
  • #1,540
  • Nov 1, 2010 2:35am Nov 1, 2010 2:35am
Ok started on two new accounts today, my previous two accounts were at 30% drawdown so I closed them. I believe my max margin settings was too high at 0.6.

Using Phantom_AddSix_03b by the way...

Two accounts are as follows:

Mini account $2000: 100 Leverage
MaxMarginToUse: 0.3
ProfitPct: 0.5

Standard account $2000: 100 Leverage
MaxMarginToUse: 0.2
ProfitPct: 0.1

Will post results every day.

And MC good job !
2 traders viewing now