Less than 20% robots approved for ATC... What went wrong?! (Survey)

 

I'm sitting here, somewhat in shock and disbelieve, looking at the approved number from all the 1,726 applications submitted to the ATC and can't help to wonder... What went wrong?! ...Why such a low number of robots (less than 20% so far) have passed the infamous test?! ...Is Metaquotes happy with this turn of events?! ...Did the platform come with too many changes in its API that made Mt4 robots translation a pain in the neck?! ...Is the new trading paradigm too confusing?! ...Did we underestimated all the changes to the API?

I can only talk about myself and the reasons why my robot didn't make it to the START line but obviously I wasn't the only one experiencing problems here. Apparently 1,385 participants experienced technical difficulties as well.

I'm opening this thread so you can share your particular experience with all of us and shed some light into this puzzling enigma which is the ATC2010.

Before you start pouring all your interesting stories into this thread I have to say that I love the change in the language. It's not C++ but it is close enough and powerful enough to create really elegant robots. This is not an attack on MQL5. I want you to treat this thread more like a survey to help Metaquotes understand why things didn't turn out the way they might have expected... I Personally won't consider a 20% approval rate an overwhelming success.

Now carry on...

 
With the bugs that still are in MT5 now, and with the changes that the correction of the bugs do to backtests, we get very volatile backtests, so the same test changes from one week to another. Things that now don't work will work and impact next week. Things that now work won't work anymore next week. It's volatile. That's why I didn't even take it into consideration. The keyword is actually premature.
Get in touch with developers using Service Desk!
  • www.mql5.com
We therefore attach great importance to all user reports about issues in our programs and try to answer each one of them.
 

i start to hate mt5 now, i made an ea work fine on backtest, no error at compile time, but when i run it it on demo. This msg " CTrade::PositionOpen::OrderCheck:: Invalid stop(s) request " keep appearing on expert tab, and not trade was executed since last 3 days..

i can't even find out wtf the problem is..

 
Reason here: their buggy auto testing system has some concurrency issue, or even worse, logic bug. Loaded the wrong code and approve/disapprove EA without a second look. They suddenly became silent when I pointed out the problem and gave proofs as well in my private chat. No response, no respect, nothing.
Get in touch with developers using Service Desk!
  • www.mql5.com
We therefore attach great importance to all user reports about issues in our programs and try to answer each one of them.
 

That i agree. They maybe too busy now to fix the system for the competition...too few people finally worked out because of a premature product is negative to the promotion.


TheEconomist:
With the bugs that still are in MT5 now, and with the changes that the correction of the bugs do to backtests, we get very volatile backtests, so the same test changes from one week to another. Things that now don't work will work and impact next week. Things that now work won't work anymore next week. It's volatile. That's why I didn't even take it into consideration. The keyword is actually premature.

 

Come on men. Championship 2010 is just an opportunity.

 I would see lamentations if in example the ATC were shifted for another year, 2011, because of the currently updating

 of the terminal. Maybe tenths of lamentations posts opened if it were not organized for 2010, but for 2011.

 

If you write too complex EAs, highly customized EAs, with the only purpose to won, this may lead to

EA not completely in harmony with the rules and simplicity of the ATC.

 

I'm sure those people behind Metatrader5 and Metaquotes are doing their best, with best effort for ATC and MT5.

We all should not forget that Metatrader is an opportunity , it's the best terminal for retail internet trading in the world,

on the point of view of IT and ICT is a great software (64 bit, memory and cpu optimized resources, agents, object programming

strong language, support, JOBS, etc.., and it will grow more and implement more functions),

and mainly we should not forget that it is for free.. Imagine if you should pay it 50$ or 500$ or 5000$ (it worth more than 5000$ ..).

 So don't spit on it. and don't scorn on the ATC and MT5 (at least this is what personally i do).

 

I can understand your rage and maybe frustration for the ATC trouble, even i have had problems sending my own

EA that revealed not so profitable in my backtestings, and i even i would like to take more and more weeks in addiction, to perfetionize

optimization, or to improve my code, but deadlines are deadlines.

 We have had all the opportunity to follow some guidelines on how to write the code, simply and linear,

and not complex and unstable, by the support team (the coders team support has always been

fast and precise, and available), free knowledge on the forum and articles. It's all about studying, and trying.


 

We have had months to write our own EA for ATC: some have had hungry of Max Profit, some others have had hungry

of stability and simplicity by sacrificing Max Profit. Those latter will not won the ATC, but at least tried.

It's just a matter of software engineering, max performances but unstable, or stable and not max performances?

I would try not max performances, but at least have a chances to use it, by stability. Writing clean code

is a part of an EA. 

 

I think we cannot claim anything about closing of the doors of ATC, since it will start within a week, and since

we have had many amounts of times-shifting of closing time, for modifying our code, and register, and we have had

montly emails in our mailbox that advise us about the fact ATC is coming to a close registration, and suggestion

to hurry up and hasten the times!

 In the world of job and business, respecting times/deadlines and strong rules is a must, in order to go in harmony with it. 

I wont appear as a godfather or moralizer, that's just what i think, and i think we cannot pretend anything. 

 

I am not pretending nothing,  neither if i won 1st place or if i result the last classified of ATC.

For me is just an opportunity, and i see at it in this way.

I will relax, and continue with my work. ATC is an add-on, on the mql5 universe.

Regards 

 

 

 

You know that Christmas is on December 25. But nobody bolted the Championship dates. 

Now I personally don't give a damn about the Championship, but I think it should probably begin mid next year.

Until then, all bugs will be solved, new assets will be introduced. Not even the goddamn CFDs work now! That, in the conditions when options were promised after that meeting in Kazan, april 2009.

So why, why have a Championship in these conditions ?

There is a lot of stress on MetaQuotes too, because the deadline is getting near. As the stress increases, their ability to repair the bugs decreases.

The Championship has to be delayed, so that people will make their EAs on a reliable station. At the same time, MetaQuotes will be able to work at a normal pace.

 

The Championship doesn't mean too much. The priority is MetaTrader itself. Because people might use MetaTrader in 1000 ways, while the Championship is very limited. The rules are generally to prohibit competition by leverage,

but a lot of people say that the rules are limiting very much their options. What is important then? Make an indicator betting station to run competitions with it ? either by MetaQuotes or by brokers? Or to make a reliable trading

station , offering people a full trading palette? Because when it will be completed, if done the way it was initially imagined at that meeting in Kazan, MT5 will be a "one platform to rule them all". It will overshadow all the existing

retail platforms, and it will be at the midpoint between retail and institutional. The major drawback vs institutional would be the backtester, forced to invent ticks instead of using tick data.

 

But you say no, screw the functionality, give us the Championship! Championships are all the time. A bunch of EAs / traders cutting their heads until one remains. Championships are all the time, hosted by several MT4 brokers. Did 

any of them change the retail trading landscape? Nope. Would a completed MT5 do the change? Yes. Imagine...options introduced...  literally thousands of symbols available, volatility games, older strategies revisited, perhaps bad

strategies turned profitable... It will be a completely different retail trading world.

 

So this is what I say : let the Championship be pushed to the next year, and have a fully operational "battlestation" by then... 

 

 

 

 

Hi all,

 

This is where mine failed:

5. Statistics
   4 kb of log files
   there are no trading operations

 

I was surprised by this, as my EA does generate operations on my computer.

Further investigations lead me to the conclusion that the Strategy Tester starts feeding ticks even if the OnInit() event handler hasn't returned.. As my OnInit() function precalculates many things, it takes a long time to return which, on slow systems, result in the tester feeding the EA with all the backtest period, while the EA isn't even initalized.. As a result, ticks are just discarded and no trade is opened...

I'm not fully sure that this is really what's happening (it may aswell just be a bug in the EA, but after spending much time i fail to see what else it could be).

However, if this behaviour is indeed what is really happening, then please someone point me to where in the docs this is documented..!!!!!!

 

In the RULES they said they can reject anyone without a reason. That is the reason that we are all here.


spamdrop:

Hi all,

 

This is where mine failed:

 

I was surprised by this, as my EA does generate operations on my computer.

Further investigations lead me to the conclusion that the Strategy Tester starts feeding ticks even if the OnInit() event handler hasn't returned.. As my OnInit() function precalculates many things, it takes a long time to return which, on slow systems, result in the tester feeding the EA with all the backtest period, while the EA isn't even initalized.. As a result, ticks are just discarded and no trade is opened...

I'm not fully sure that this is really what's happening (it may aswell just be a bug in the EA, but after spending much time i fail to see what else it could be).

However, if this behaviour is indeed what is really happening, then please someone point me to where in the docs this is documented..!!!!!!

 

 


 

It is not the rules, but many were cheated when the the initial auto testings gave false passes/approvals. And they should differentiate  backtesting and real trading, as backtesting with multicurrencies take up significant resources but not so in real multicurrency trading (depending on the algorithm)