Mql5.com Improvement Proposal 03

New comment
 
The MQMGuard or the MetaQuotes Market Guard 

For each product category a vendor can publish , add an additional one that is the category +With MQMGuard.For example "Expert advisor +MQMGuard")

What is the MQM Guard ?

It is a "guarantee" that the product is lacking certain scam tricks and it is mostly what it claims to be.

How does it work ? 

  • The vendor has to deposit an amount , for this example let's say $100 for a new product with the MQMGuard.
  • The mql5.com commission for an MQMGuard product is 25%. 
  • For MQMGuard the product is uploaded with all its source files and not the ex file.
  • The structural rules are similar to those of the CodeBase . 
  • Now , as there's moderator volunteers there can be MQMGuard check volunteers (at first because this is insane to do with a gpt , but the gpt can be learning while the checks are manual).
  • Per product 5 volunteer checkers are randomly assigned .(difficulty here is to have a decent amount ,there may be difficulties too if you want to use the service desk as per employment laws + salary that is why i mention "volunteers" here).
  • Goes without question the product must be commented out in english only , and the volunteers must be able to understand english .
  • The 5 volunteers assigned to the product check for known scam types , the claims in the description , the title and anything else the administration does not want to happen . For instance , a random popup with a message from the vendors server .
  • That means that the volunteers must also be well versed in mql5 . 
  • Now , if all volunteers allow the product to pass , it is compiled on the server , the source files are deleted.($100/5 or amount/5) is released to the volunteers who checked it .7Days clearing for these payments.
  • If the product has at least one volunteer who objects publication (with their remarks) then the source is marked for a check by an "entity" you trust (if the desk is familiar with mql5 then the desk) and , if the product is indeed rejected then the volunteers who objected its release share the $100 . Those who green lit the product are banned/suspended including any pending payments they have from the MQMGuard. 
    If the rejection reason is some inconsistency then the remarks are sent to the vendor . If its a suspected scam it is checked further and if it is indeed a scam attempt the vendor is banned.
  • After publication the vendor cannot edit the overview
  • For a product upgrade the whole process must restart.


Now , the 5% from the commission is assigned to ALL the volunteers that did checks for that product to be published with a clearing of 45days.

So the better the product the more the volunteers who checked the product receive in the long run.

If however a complaint arises and misdeeds are discovered for that product, then all pending volunteer payments on that product halt .

The volunteers that green lit the product are revoked their volunteer status and no longer receive commissions for ANY product they checked.

If a volunteer communicates with a vendor because they recognized the product while checking , if the vendor responds he is banned ,and the volunteer rights are revoked + they are also banned too.

Unfortunately the checks must be performed via remote connection of some sort so that code cannot be copied by the volunteers.

Now , the service in its first stages should be optional with a payment by the vendor .

In the future as the market progresses and the quality output is evaluated (meaning no abuse arises etc) it can be made mandatory and it can also accommodate 3 free products for new vendors because they may not be able to afford the service at first . 

Ranking of MQMGuard products : a product with MQMGuard is visibly highlighted , there is a logo indicating it has the guard and it gets 10x more rating points than products without the guard .  

So if the guard works in the end the guard products will outrank the others . 

If it works as intended you are looking at a 90% eradication of scams from your market + absolute application of any rules you have.

The goal of course in the end is to automate that process to eliminate the human "corruption" factor from the process ,and to achieve that the reward for those participating (who essentially will be teaching the algorithm indirectly) must be enticing.

Pros : 

  • If some vendor claims DeepLearning , ML , NeuralNetwork you can check if it is indeed used .
  • History reader eas are decimated , gone . No more of that.
  • You can detect martingale , grid etc
  • You can detect if there is indeed news filtration occuring for the trades.
  • It's still a gray area with gpt though as you don't have access to the server of the vendor.
  • All the things you detect in the code , features , can be now added as properties of the product allowing you to have precise search filters for MQMGuard products . 

Cons : 

If all the volunteers decide to just allow products you won't realize until a scam emerges from an MQMGuard product .

So you could use mql5 familiar users you trust first for the pilot of the service , moderators for instance .

Edit On 30.07.2023 17:45 GMT : 

  • The pool of candidate volunteers can be freelancers who meet certain criteria and are active . 
  • If the service desk cannot accommodate the role of the "final decider" for products that the vote is not unanimous on , after some time the most trusted volunteers may be offered to decide on these products (again random selection and cut in the %) . 
  • I left out what happens when one of the checkers objects per the publication of a product , but , their claims (for the rejection) are invalid . They get their volunteer status revoked and all commission rights that have accrued that far and pending payments from the checks.

 
Lorentzos Roussos:
The MQMGuard or the MetaQuotesMarket Guard

For each product category available when publishing a product add an additional one that is the category +with MQMGuard.(ex : Expert Advisor + MQMGuard)

What is the MQM Guard ?

It is a "guarantee" that the product is lacking certain scam tricks and it is mostly what it claims to be.

How would it work ? 

  • The vendor has to deposit an amount , for this example let's say $100 , for a new product with the MQMGuard.
  • The mql5.com commission for an MQMGuard product is 25%. 
  • For MQMGuard the product is uploaded with all its source files and not the ex file.
  • The structural rules are similar to those of the CodeBase . 
  • Now , as there's moderator volunteers there can be MQMGuard check volunteers (at first because this is insane to do with a gpt , but the gpt can be learning while the checks are manual).
  • Per product 5 volunteer checkers are randomly assigned .(difficulty here is to have a decent amount ,there may be difficulties too if you want to use the service desk as per employment laws + salary that is why i mention "volunteers" here).
  • Goes without question the product must be commented out in english only , and the volunteers must be able to understand english .
  • The 5 volunteers assigned to the product check for known scam types , the claims in the description , the title and anything else the administration does not want to happen . For instance , a random popup with a message from the vendor's server .
  • That means that the volunteers must also be well versed in mql5 . 
  • Now , if all volunteers allow the product to pass , it is compiled on the server , the source files are deleted and is published.($100/5) is placed as a pending payment on each volunteer's account , 7day clearing).
  • If not , then ($100/5) is placed as a pending payment on each volunteer's account , 7day clearing , and the remarks
    are sent to the vendor and the service desk is notified if the rejection reason was a suspected scam.
  • Now , the 5% from the commission (of sales of the product) is assigned to ALL the volunteers that did checks for that product to be published with a clearing of 45days.


So the better the product the more the volunteers who checked it make .

If however a complaint arises and misdeeds are discovered for that product :

  • All pending volunteer payments on that product halt .
  • The volunteers that green lit the product are revoked their volunteer status and no longer receive commissions for ANY product they have checked.

If a volunteer communicates with a vendor because they recognized the product while checking :

  • if the vendor responds he is banned
  • he volunteer rights are revoked + they are banned too.


Unfortunately the checks must be performed via remote connection of some sort so that code cannot be copied by the volunteers.

Now , MQMGuard in its first stages should be optional with a payment by the vendor only.

In the future as the market progresses and the quality output is evaluated (meaning no abuse arises etc) it can be made mandatory and it can also accomodate 3 free products for new vendors because they may not be able to afford the program at first. 

Ranking : a product with MQMGuard is visibly highlighted , there is a logo indicating it has the guard and it gets 10x more rating points than products without the guard .  

So if the guard works in the end the guard products will outrank the others . 

If it works as intended you are looking at a 90% eradication of scams from your market + absolute application of any rules you have.

The goal of course in the end is to automate that process to eliminate the human "corruption" factor from the process ,and to achieve that the reward for those participating (who essentially will be teaching the algorithm indirectly) must be enticing.

At first due to low volume of volunteers there must be an availability indication for vendors of whether or not the MQMGuard is available (meaning if there are available volunteer checkers).

☕️


OMG.

Why not simply, instead of all this "nonsense" add a special type of signals to the signal database. No subscription possible. The vendors EA gets to set up the EA he uploaded to the market for signal service to a demo account.

Use the statistics of that signal to increase the ranking of the product.

Everyone gets to see detailed history tracking, everyone can make his own conclusions to what the product is capable of.

It were not as complicated as MGM were, and it's automated, without source code disclosure or anything else that nobody will agree to.

Don't forget MQ Market is not the only place to buy EAs, they are in competition with others. Nobody would sell their EAs under your layer out conditions.
 
Dominik Christian Egert #:
OMG.

Why not simply, instead of all this "nonsense" add a special type of signals to the signal database. No subscription possible. The vendors EA gets to set up the EA he uploaded to the market for signal service to a demo account.

Use the statistics of that signal to increase the ranking of the product.

Everyone gets to see detailed history tracking, everyone can make his own conclusions to what the product is capable of.

It were not as complicated as MGM were, and it's automated, without source code disclosure or anything else that nobody will agree to.

Don't forget MQ Market is not the only place to buy EAs, they are in competition with others. Nobody would sell their EAs under your layer out conditions.

That is a good idea too . 

Like an MQMTracked or something . 

Hosted by MQ in their demo accounts ? 

What about utilities and indicators though ? 

Also we may see the "launch 20 of them" and the one that shows a nice graph on the demo monitor , unhide it . 

The MQMGuard will make the mql5.com market -if not abused by the checkers of course- the least likely place for someone to fall victim to a scam . Your proposal reduces the scam likelihood as well , i like it . There is however something they may not like in the demo signal solution . Sales will likely plummet .

Yes at first the "human looking at source code" may make vendors cringe but as the algorithm learns to check on its own , when it gets automated mq will end up with a massive advantage over other markets . 
 

I understand your point, however, I believe that might be too strict. Some of us here might be students, studying finance, 100 USD can be a budget for anyone. Moreover, I wouldn't be personnally okay with having someone else getting access to my EA's source code, especially volunteers.


I do not understand as well why would you like to detect "Martingale and Grids". These are trading systems as well, there is no need to discriminate them. The only case were that would be legitimate, is if the EA itself if claiming to not use either of those systems. However, even in that case, there are a LOT of definitions for "Martingale" for example. And, in some case, your EA is made to be a Martingale, which is the case with some of mine, would I be penalized too? Even though the point of my EA is to be a Martingale?


I do agree on something however: scams are a huge issue on the MQL5 market, I can't stand seeing extremly pricy EAs on the market that are total scams yet getting bought like cookies, while individual sellers pain to climb up the ranks of the market.


Here is a solution I thought of: most of the scams EA, take form as "AI CHAT GPT" EAs, while just being Martingale. You can see that on the signals that many attach, that are just Martingale as show by the curve. I would suggest a system that would calculate a similarity ratio in terms of performance between backtest and live signal. Of course, I know backtests are no guarantee of live (forward) results, but when you see an "AI CHAT GPT" EA having a 0.5% drawdown durinbg 7 years and yet magically having an enormous drawdown on the signal, it's probably that something fishy is happening. You could have some kind of "signal accuracy" rating, that would check the accuracy between the signal and a backtest report.

The signal system is also not that good. At the moment, most of those EAs would for instance have a small balance signal, with a very high leverage, therefore showing huge % of growth, and the signal would get deleted / replaced as soon as it blew up the account.


All of that to say, I agree with you on the fact that something should be done to check the EAs a bit more, however, I personnally wouldn't be comfortable sharing a source code I took a lot of time to work on, even with NDAs or Trust Contracts.

 
Zaky Hamdoun #:



I understand your point, however, I believe that might be too strict. Some of us here might be students, studying finance, 100 USD can be a budget for anyone. Moreover, I wouldn't be personnally okay with having someone else getting access to my EA's source code, especially volunteers.

Yeah that's why in its pilot phase it could be optional , but hey if you do it you get a 10x in your rating over the others . Why is the price in place in the pilot phase and the 3 free ones are not ? because there would be a flood while everyone tries to get the 10x rating (i mean if you have a simple tool , you break no rules , the 10x boost is enticing) and mq would have to "hand in pocket" those free guards at first .

I do not understand as well why would you like to detect "Martingale and Grids". These are trading systems as well, there is no need to discriminate them. The only case were that would be legitimate, is if the EA itself if claiming to not use either of those systems. However, even in that case, there are a LOT of definitions for "Martingale" for example. And, in some case, your EA is made to be a Martingale, which is the case with some of mine, would I be penalized too? Even though the point of my EA is to be a Martingale?

It detects it and it becomes part of its properties , so when you look at the properties table it tells you and furthermore you can search by it and when you search martingale you really get martingale for instance . Now if the vendor says no martingale and it has martingale ,you understand.

Here is a solution I thought of: most of the scams EA, take form as "AI CHAT GPT" EAs, while just being Martingale. You can see that on the signals that many attach, that are just Martingale as show by the curve. I would suggest a system that would calculate a similarity ratio in terms of performance between backtest and live signal. Of course, I know backtests are no guarantee of live (forward) results, but when you see an "AI CHAT GPT" EA having a 0.5% drawdown durinbg 7 years and yet magically having an enormous drawdown on the signal, it's probably that something fishy is happening. You could have some kind of "signal accuracy" rating, that would check the accuracy between the signal and a backtest report.

The signal system is also not that good. At the moment, most of those EAs would for instance have a small balance signal, with a very high leverage, therefore showing huge % of growth, and the signal would get deleted / replaced as soon as it blew up the account.

Yeah this is a nuanced case too that this solution may solve (the guard) , and also Dominik's idea ties in well here because i've seen numerous times (you too) the monitoring signal url being changed on the ea overview . One of them was caught on video by one of the "youtube ea reviewers" .

 
Lorentzos Roussos #:

It detects it and it becomes part of its properties , so when you look at the properties table it tells you and furthermore you can search by it and when you search martingale you really get martingale for instance . Now if the vendor says no martingale and it has martingale ,you understand.

Yeah I get that. However, I still stand on the fact that there's a lot of ways of implementing a Martingale (keeping positions opened? closing them? doubling lots? using a formula? etc...), therefore making it difficult to detect 100%.

I still however would not be ready to provide my source code to someone outside of my organization / team however...

I believe the signal thing might be the best alternative.

 
Zaky Hamdoun #:

Yeah I get that. However, I still stand on the fact that there's a lot of ways of implementing a Martingale (keeping positions opened? closing them? doubling lots? using a formula? etc...), therefore making it difficult to detect 100%.

I still however would not be ready to provide my source code to someone outside of my organization / team however...

I believe the signal thing might be the best alternative.

Yes of course . This proposal is a hard sell . For instance the first thing that crosses your mind is 

"The volunteer must be well versed in mql5" , which implies someone who possibly can launch products will look and understand your code . 

But the initial phase will be very "appealing" to both sides .

On one hand you'll have the vendor who'll consider and think to themselves :

  • -i'm not a scammer
  • -my code is well formatted
  • -anyone with experience can reproduce what i've done anyway
  • -i like the rating boost , i'll do it !

Then as these vendors start requesting the service there will be volunteers (who are approached to be offered to participate) who will think :

  • i know mql5
  • if i do a decent job and examine the code carefully ,i can earn a lot in the future because these products will be boosted and will be fewer in the beginning gathering -relatively- more sales and attention

Then it can spread .

If it spreads , besides the skeptics -which is understandable- there is one more group that will not submit their code for the service . The scammers. 

 
Lorentzos Roussos #:

Yes of course . This proposal is a hard sell . For instance the first thing that crosses your mind is 

"The volunteer must be well versed in mql5" , which implies someone who possibly can launch products will look and understand your code . 

But the initial phase will be very "appealing" to both sides .

On one hand you'll have the vendor who'll consider and think to themselves :

  • -i'm not a scammer
  • -my code is well formatted
  • -anyone with experience can reproduce what i've done anyway
  • -i like the rating boost , i'll do it !

Then as these vendors start requesting the service there will be volunteers (who are approached to be offered to participate) who will think :

  • i know mql5
  • if i do a decent job and examine the code carefully ,i can earn a lot in the future because these products will be boosted and will be fewer in the beginning gathering -relatively- more sales and attention

Then it can spread .

If it spreads , besides the skeptics -which is understandable- there is one more group that will not submit their code for the service . The scammers. 

Even if that does happen, I do not agree with the mandatory formatting using MQL’s coding standard.

I personally hate it, I might not be the only one as well. 

As for the point « anyone with experience can reproduce it », that’s not the case. When I sell my bots, I sell the code, of course, but I also mainly sell the thought process and strategy behind it.

Not everyone can come out with a trading strategy that works. 
 
Zaky Hamdoun #:
Even if that does happen, I do not agree with the mandatory formatting using MQL’s coding standard.

I personally hate it, I might not be the only one as well. 

As for the point « anyone with experience can reproduce it », that’s not the case. When I sell my bots, I sell the code, of course, but I also mainly sell the thought process and strategy behind it.

Not everyone can come out with a trading strategy that works. 

Yes i don't like the formatting either but it is suggested per the automation of the service in mind.

Theoretically (its insanely difficult tho) since the CodeBase + Articles have this layout it could be used for an initial model

and then as the volunteers interact with code and find issues and log their remarks then it will keep adjusting and learning .

The end goal , if possible , is automation . No volunteers .

Hence the suggestion for the format , because there is a database of that already.

I meant that these vendors will be the first "movers", i'm not saying everything is not unique.

And yeah you (and any decent vendor) has the users that trust them . For them whether or not you use the MQMGuard it won't make a difference because they know you .

But those who end up buying the junk , they will gradually turn to preferring only MQMGuard products .

But look at it this way , you will be taking a risk but the end goal will benefit all of the vendors . (Those who don't look for ways to slide in the clients wallet with something mq has not considered or has not seen yet).

 

Response from service desk 

New comment