Blessing Multipair EA | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Trading multipair triangles, no stoploss 78 replies

'Blessing' EA Modification Wanted 3 replies

Single, Few Pair Traders (1-4) vs. MultiPair Traders (8-20) 0 replies

Coding Request for Multipair MTF ADR Dashboard 2 replies

trying to make a multipair buy and sell script 6 replies

Attachments: Blessing Multipair EA
Exit Attachments

Blessing Multipair EA

  • #41
  • Jan 12, 2020 10:39pm Jan 12, 2020 10:39pm
Quoting ursinho4711
@ll! Here is another reason why I wanted to have the EA as multipair-version: now I can monitor and control all open trades via one dashboard. I can set a basket TP, basket SL, close all manually... {image}
nice dashboard to have
LUCKY golden trades, LUCKY gold investor trading style
  • #42
  • Jan 12, 2020 11:12pm Jan 12, 2020 11:12pm
Quoting fx4money
{quote} will test this on demo to see how it work and if it really work and if need to be revised
It works (basically), but please wait: I am just testing the bugfix for closing all positions instead of just closing positions for the TradePair

still did not sleep...
  • #43
  • Jan 12, 2020 11:19pm Jan 12, 2020 11:19pm
Quoting fx4money
{quote} will test this on demo to see how it work and if it really work and if need to be revised
Try that one, I introduced the TradePair into the ExitTrade... please see if the exits now work better. Still some work ahead...
Attached File(s)
File Type: ex4 BlessingLite_3_v3.9.6.26_beta_odersymbol.ex4   187 KB | 761 downloads
File Type: mq4 BlessingLite_3_v3.9.6.26_beta_odersymbol.mq4   127 KB | 992 downloads
  • #44
  • Jan 13, 2020 12:44am Jan 13, 2020 12:44am
Quoting ursinho4711
... is anyone else besides me enjoying the EA? I think it's fun watching. Better than TV... {image}
Ok i've started yesterday with setting file you posted and suppose are you using.... (bollinger) But on 5min tf i have totally different pairs traded....
Broker ic markets
Sure not the same setting i suppose...
Thanks for effort
Best regards
Attached Image (click to enlarge)
Click to Enlarge Name: Cattura.PNG Size: 29 KB
  • #45
  • Jan 13, 2020 1:07am Jan 13, 2020 1:07am
Quoting cescof
{quote} Ok i've started yesterday with setting file you posted and suppose are you using.... (bollinger) But on 5min tf i have totally different pairs traded.... Broker ic markets Sure not the same setting i suppose... Thanks for effort Best regards{image}
unless you did not start the EA exactly at the same time, with exactly the same environment... then it is almost a must that you have different pairs to trade! Don't forget that we are on a fast timeframe... what did you expect?
I think I will sleep now...

Bugfixed (closing wrong pending positions) version comes today in the evening...
  • #46
  • Jan 13, 2020 1:12am Jan 13, 2020 1:12am
Quoting ursinho4711
{quote} unless you did not start the EA exactly at the same time, with exactly the same environment... then it is almost a must that you have different pairs to trade! Don't forget that we are on a fast timeframe... what did you expect? I think I will sleep now... Bugfixed (closing wrong pending positions) version comes today in the evening...
Thanks
Take care
Regards
  • #47
  • Jan 13, 2020 2:32am Jan 13, 2020 2:32am
betabugfix looks good, I will let it run for some more time. Behaviour as expected (or at least hoped): here and there small but steady wins, until now low DD. Just like a squirrel ;-)
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 52 KB
  • #48
  • Jan 13, 2020 3:44am Jan 13, 2020 3:44am
Is this normal? Just order for jpy related pairs?
Attached Image (click to enlarge)
Click to Enlarge Name: Cattura.PNG Size: 134 KB


no it wasn't... now i changed chart (AUDUSD) and load again Ea and i have some order placed for all pairs a part jpy ones....still confused
Regards
  • #49
  • Jan 13, 2020 3:46am Jan 13, 2020 3:46am
Quoting cescof
Is this normal? Just order for jpy related pairs? {image}
The same here on H1
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 86 KB
  • #50
  • Jan 13, 2020 3:51am Jan 13, 2020 3:51am
Sticking with the original Inputs / Properties names would have been nice. Especially for compatibility with the Blessing manual and for those of us that have optimized the original a few thousand times ... :-)

Also nice when you're looking at the code and making adjustments.

That's what all those double semicolons were for.
  • #51
  • Jan 13, 2020 4:54am Jan 13, 2020 4:54am
Quoting richard96816
That's what all those double semicolons were for.
Ah, thanks for this info! When I looked at the code from versions .10 on following, I thought: "what the heck...". Now it makes sense...

Quoting richard96816
Sticking with the original Inputs / Properties names would have been nice
I think I did it, or at least tried to as much as possible. Many variables just got an pairinfo[]. in front of their original name, as they moved into the big struct. I tried to avoid as much work as possible ;-) ...
... and to keep the blessing "heart" as original as possible

(did not get much sleep the last days before Sidney open..)
  • #52
  • Jan 13, 2020 5:13am Jan 13, 2020 5:13am
Quoting richard96816
Especially for compatibility with the Blessing manual and for those of us that have optimized the original a few thousand times ...
If we wanted to, we could restart from version 09 (or better 13?) again. I stuck with most of the delta-files, that's why I already counted to 26 in the version number.
The first approach was to see whether it makes sense to follow the idea of a multipair-Blessing. When I look at the results of my last beta-version (with the 5-minutes set-file) I have to say: certainly yes!
The big difference from the original version to the multipair version is "just" the struct and the big for-to loop. And cleaning the code from parts that are not needed or even contraproductive in the multipair version (e.g. ForceMarketCond or hedging on another pair).

That the compatibility to older versions is given, can be seen in the fact that older set-files still can be used.

Have a great start into the week @ll and thanks for any further helpful input!
  • #53
  • Jan 13, 2020 6:11am Jan 13, 2020 6:11am
Quoting ursinho4711
{quote} Try that one, I introduced the TradePair into the ExitTrade... please see if the exits now work better. Still some work ahead... {file} {file}
Hello i would like to report the following bug i guess
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 3 KB
Attached Image
  • #54
  • Jan 13, 2020 7:35am Jan 13, 2020 7:35am
Quoting richard96816
Sticking with the original Inputs / Properties names would have been nice. Especially for compatibility with the Blessing manual and for those of us that have optimized the original a few thousand times ... :-)
as to set-files and optimizing, you may want to take a look at Andreas Clenow's new book:

Rules and Variations
Generally speaking, you should aim for as few rules as possible and as few variations as possible.
Once you have arrived at a model purpose, you need to figure out how to formulate this purpose in terms of trading rules. These rules should be as simple and as few as you can muster. Robust trading models, those that work over the long run, tend to be the ones that keep things simple.
Any complexity you add needs to pay off. You should see complexity as something inherently bad, something which needs to justify its existence. Any complexity you want to add to your model needs to have a clear and meaningful benefit.
Moreover, any complication or rule that you add needs to have a real life explanation. You can’t just add a rule just because it seems to improve backtest performance. The rule needs to fit into the logic of the model purpose and play a clear rule in achieving that purpose.
Once you have arrived at a set of rules for testing your market theory, you probably want to try some variations. Note that there is a world apart between testing variations and optimization.
As an example, let’s assume that you want to test a mean reversion type of strategy. You believe that when a stock has fallen four standard deviations below its 60 day linear regression line, it tends to bounce two standard deviations up again.
Now you already have multiple parameters in play. Modeling and testing these rules is a fairly simple task. You could try a few variations of this, perhaps to expect the bounce by three or five standard deviations, using 30 or 90 day regression or a variation in the target bounce distance.
Making a few variations like this can be useful, both for testing parameter stability and to actually trade some variations of the rules to mitigate over-fitting risks.
What you don’t want to do is to run an optimizer to figure out that the optimal entry is at 3.78 standard deviations, on a 73 day regression, using a target of 1.54 standard deviations. Such data is absolute rubbish.
Optimizers will tell you what the perfect parameters was for the past. They will also con you into a false sense of security, and make you believe that they have any sort of predictive value. Which they don’t.
No, skip the optimization. But make a few variations of the rules, using reasonable, sensible numbers.
From: Clenow's new book Trading Evolved - Anyone can Build Killer Trading Strategies in Python
See also: Carver, Systematic Trading, 2015
  • #55
  • Jan 13, 2020 7:52am Jan 13, 2020 7:52am
Quoting josi
{quote} as to set-files and optimizing, you may want to take a look at Andreas Clenow's new book: Rules and Variations Generally speaking, you should aim for as few rules as possible and as few variations as possible. Once you have arrived at a model purpose, you need to figure out how to formulate this purpose in terms of trading rules. These rules should be as simple and as few as you can muster. Robust trading models, those that work over the long run, tend to be the ones that keep things simple. Any complexity you add needs to pay off. You should...
Oh jeez.

We're so far past that book it's not funny. We have Blessing. Yes, it's complex and a hassle to optimize. But god does it work! It was originally designed to change fewer rules, and that works. Or you can change more rules and make more money. You choose.

We're not starting from scratch, which is what books like that one usually address. We have a tool that can be relatively easily tailored to almost any currency pair and made to lock into useful repeating patterns. It works! Backwards and forwards! Money in the bank.
  • #56
  • Jan 13, 2020 8:02am Jan 13, 2020 8:02am
Quoting richard96816
{quote} Oh jeez. We're so far past that book it's not funny. We have Blessing. Yes, it's complex and a hassle to optimize. But god does it work! It was originally designed to change fewer rules, and that works. Or you can change more rules and make more money. You choose. We're not starting from scratch, which is what books like that one usually address. We have a tool that can be relatively easily tailored to almost any currency pair and made to lock into useful repeating patterns. It works! Backwards and forwards! Money in the bank.
well, I wasn't trying to undermine your project.
What you write tells me, though, that you haven't read any of the books mentioned - but nevertheless feel inclined to criticize their content. I can only hope that you are a bit more rational and diligent in your optimizations.
Best of luck.
  • #57
  • Jan 13, 2020 12:30pm Jan 13, 2020 12:30pm
Any idea on this please.
I keep seeing this message continuously. Anyway to stop this message please.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 42 KB
  • #58
  • Jan 13, 2020 12:41pm Jan 13, 2020 12:41pm
Quoting Mcxtrader
Any idea on this please. I keep seeing this message continuously. Anyway to stop this message please. {image}
in the source file, around line 245, find "bool Debug = true;" and set it to "bool Debug = false";
To make it more comfortable, you can write "input bool Debug = false;"
In this way you can switch the option on/off via the EA settings. Will change this in the next version.
  • #59
  • Jan 13, 2020 3:25pm Jan 13, 2020 3:25pm
Quoting josi
{quote} well, I wasn't trying to undermine your project. What you write tells me, though, that you haven't read any of the books mentioned - but nevertheless feel inclined to criticize their content. I can only hope that you are a bit more rational and diligent in your optimizations. Best of luck.
I've read many of the books. Useful and helpful, for writing EAs. All the important (and high-falutin) issues regarding optimization of new EAs. Blessing is not new. It's already written. It works fairly well, better than any others I've tested or written.

One issue I left out. Optimization needs to be fast. Because you need to forward test five or 10 or more instances and choose the good ones. That's the main reason many of the classical works suggest a small rule-set, to speed up the terribly slow optimization process. (And to make the results last a long time, which I find less important.) Forward testing is the real test.

My apologies if my response was overly strident. It was late. I've read quite a few good books and lots of online discussions with lots of smart people. I'm starting to get the hang of this stuff. :-)

I keep having ideas on how to enhance Blessing's trading abilities. And then I remember how effective it is already and how I really don't want to damage it in the process.

I've had the idea of deploying a fair number of instances on difference pairs to hopefully make the results more robust. Or perhaps to turn her into a multi-pair system. That seems like pretty good idea. But keeping the core as it is seems like a great plan, at least for now. That means that most of the advanced and foundational optimization techniques are unnecessary and already done.

Thanks very much for your contribution to the discussion. I should not have been so dismissive. You put a lot of thought and effort into your post and are very much appreciated for that. Many will likely learn from it, and that's a good thing.

Best of luck!
  • #60
  • Edited 11:41pm Jan 13, 2020 11:29pm | Edited 11:41pm
first day of trading, demo, let see how it goes
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 10 KB
1 trader viewing now