• Please try to select the correct prefix when making a new thread in this folder.

    Discuss is for general discussions of a financial company or issues related to companies.

    Info is for things like "Has anyone heard of Company X?" or "Is Company X legit or not?"

    Compare is for things like "Which of these 2 (or more) companies is best?"

    Searching is for things like "Help me pick a broker" or "What's the best VPS out there for trading?"

    Problem is for reporting an issue with a company. Please don't just scream "CompanyX is a scam!" It is much more useful to say "I can't withdraw my money from Company X" or "Company Y is not honoring their refund guarantee" in the subject line.
    Keep Problem discussions civil and lay out the facts of your case. Your goal should be to get your problem resolved or reported to the regulators, not to see how many insults you can put into the thread.

    More info coming soon.

Intensive Review Genetic Builder Intensive Review

Long and detailed review proocesses. Ask AsstModerator if you want to leave one.
Status
Not open for further replies.
If I may, Mark, I'll be adding the issues I come across here in bits and pieces, without trying to differentiate too much between the fundamental and the cosmetic. Pharaoh is taking your software for a very thorough spin. I suspect my contributions will often be more superficial...

1) Looking at the Build tab, having chosen and started the build of a strategy:
What is missing for me is a running indicator of how much time is passing and how many strategies have been generated. Possibly a strategies/minute gauge that can give me a sense of how well my hardware is performing. I expect people will want to maximise their number crunching capabilities in the long run. It's true that once I press Stop, the log publishes how many seconds have elapsed, but "DONE in 9691.578 s." is not very user friendly. And "Elapsed Time:" displays "N/A". Some more live feedback from the programme regarding its performance would be good.

2) The numbering of the strategies:
Each strategy is given a number. That number starts with "0.". I'm guessing that what looks like a decimal point is actually something else? Does the zero refer to mode - in this case Random Generation? Be that as it may, it leads to a very confused looking count, with 0.999 being followed by 0.1000. From a user point of view, this could be improved for the sake of clarity.
 
I've just been informed that GeneticBuilder 2.0 has been released. Now I just have to cancel a few hundred other items from my schedule to play with it.
 
You got ahead of me Prahaoh :)

Yes, I'd like to announce that new major version 2.0 of Genetic Builder
was just released.

The main new features are:
  • Support for tick data
    enables to test the strategies with best possible precision

  • Strategy Editor
    allows you to edit or create a new strategy in an easy-to-use wizard

  • Custom Indicators
    enabling you to use your favorite indicators in your strategies without the need for them to be build-in Genetic Builder

  • Number of other enhancements
    like configurable Risk-Reward ratio, support for candle patterns, better EA source code generation and so on

You can see a complete list of new features here;
Genetic Builder What's New

I'd like to thank all users for their ideas and feature requests. We weren't able to put everything to the new version, so some of the smaller features wait for the next minor version.

What's most important is not the program features, but its ability to create real, profitable trading strategies. I believe the new version moved the possibilities of the program in this to a new level.
 
Let's see...

Had a few technical issues with some upgrades. Lost a lot of time due to vacation, but I'm back in the saddle now.

I think I may have taken too much advantage of setting those automatic deletions of EAs that don't meet my standards. Over 200,000 generated and not one survived yet.
 
Let's see...
Had a few technical issues with some upgrades. Lost a lot of time due to vacation, but I'm back in the saddle now.
I think I may have taken too much advantage of setting those automatic deletions of EAs that don't meet my standards. Over 200,000 generated and not one survived yet.

yes, it takes some time to figure out the best settings. And it can be different for every timeframe or symbol.

If you'll get all the strategies filtered out check the log file what was the reason for dismissal and change the custom conditions to be less strict.
 
Found the problem. I told it I wanted stability above 50, both in and out of sample, not above 0.50.

Between that and wanting profit above 0 in and out of sample, there are very few survivors.
 
Woohoo! One of my feature requests was included and I didn't even know it.

I'd left my spare laptop running about a 5 days and 9 EAs survived my automatic deletions - then the laptop locked up. I was about to write a long, whining email to Mark about needing a recovery option, but when I restarted Genetic Builder, it gave me the option to recover those 9 lost strategies.
 
this looks like a great product...
couple of questions...
forex and metals symbols both? and some mt4 brokers offer indices and even individual stocks... these, too?

does the generator put out advisors actually ready to trade...
that will deal with real world trading conditions,
like trying again for a certain time if orders don't get executed?

thanks...
 
It should work on anything you can get MT4 data for. If I ever have a free moment, I want to test it out on Gold and Silver.

I'm not sure about retries. Send a PM to Mark and ask him to come in and answer that part.
 
GB

Hi Pharoah & Mark,
Been running trial unit for a few weeks. I only have smaller memory CNS server (made for MT4 traders) controlled via a Galaxy SIII smartphone, but am seeing how GB manages RAM, fortunately for me, so GB has ongoingly maintained stability. Am wondering though if such manner of running reduces effectiveness??? Can't really tell, but as mentioned in this thread there is sticking with certain kinds of strategies, still am probably learning and shall have to go over this thread more in future after I purchase GB in 3 months.
Big question/problem I have is seeing on 5minTF attempts strategies that avoid hitting stops and stay open throughout a trend swing before and during out-of-sample period.......previously saw slew of strategies with empty set for exits, what am I doing wrong to invite such?......or.....is this a filter needed that says dismiss quantity of coding = 0?
Now I'm seeing such large occurences of no out-sample trades occur with ok exit code, hmmm, since I'm on 19mo in-sample and 2mo out-of-sample maybe another exit has not arrived? Will have to check.
So far, regardless of my limited mode I have couple 15minTF EA's that were good in OOS in GB (ending in Jan1st '12) and equally good in IBFX mini recent live 4mo backtest, quite an amazing thing to see such robustness can happen from "garden variety" indicators and moving the EA over to MT4!!!
Also found a profitable & ultra simple manual Daily (open on one day of week, close on another) strategy that I posted on a genetic builder thread I started on pipcop.com.

Heard there are low spread brokers like ILQ w/ 1.3 pip spreads, so be quite something to find a scalper mode strategy someday that can do well w/ smaller stops. I know, dream on, suppose it be like Pharoah's dreams of Holy Grail......don't think THAT can ever exist, market has a varying randomity to it, GB finding anything that can work decently....for a period?......is about best new hopes "we've" arrived at, like we gotten "to the moon Alice".
Mark has acknowledged my hopes for automatizing the custom indicator integration, so I look forward to that after purchase......some forums have advanced indicators, that section of that forum does costs per month to partake, but such new crops of indicators in GB on a more rapid employment basis could be a "gas".
OkYdOkY!
Jerry
 
Status
Not open for further replies.
Back
Top