Public discussion of the formula for calculating the cost of resources in the MQL5 Cloud Network - page 24

 

Renat:
А Вы сами покупали рекламу в адвордсе?

Во первых, там продавец не участвует в формировании цены. Он лишь получает часть цены, которую в единоличном виде выставляет гугл во время безумной гонки бюджетов рекламодателей.

..


Instead there will be "government regulation" of the price to balance supply and demand.

Buying :-)

It's even more fun there. The selling price is formed automatically based on demand.

I agree that manual regulation on a standardised product is sensible. There are unlikely to be strong supply-demand waves. The main thing is not to have a situation where there are 10 sellers and 100 buyers, and end up with a perpetual shortage like in the days of government regulation :-)

There is also a psychological moment. IMHO, if the resource sellers can only make $20-50/month on the service, then many just won't bother with the cloud. Perhaps, selling resources to individuals is not the way to go at all, and it is easier to make/rent a server farm and sell their resources. Are there any successful analogues to the project in the world?

 
The current prices are quite high - in less than 24 hours it has already counted $7. I'm not willing to pay more than $5 + I need the network for multicurrency, not just empty responses with initial balance and 0 trades.
Документация по MQL5: Стандартные константы, перечисления и структуры / Торговые константы / Свойства ордеров
Документация по MQL5: Стандартные константы, перечисления и структуры / Торговые константы / Свойства ордеров
  • www.mql5.com
Стандартные константы, перечисления и структуры / Торговые константы / Свойства ордеров - Документация по MQL5
 
masharov:

Buying :-)

It's even more fun there. The selling price is formed automatically based on demand.

I agree that manual regulation on a standardised product is sensible. There are unlikely to be strong supply-demand waves. The main thing is not to have a situation where there are 10 sellers and 100 buyers, and end up with a perpetual shortage like in the days of government regulation :-)

There is also a psychological moment. IMHO, if the resource sellers can only make $20-50/month on the service, then many just won't bother with the cloud. Perhaps, selling resources to individuals is not the way to go at all, and it is easier to make/rent a server farm and sell their resources. Are there any successful analogues in the world?

It's more likely to start with 100 sellers and 1 buyer.


Probably it is worth waiting for 2-3 months or even half a year or even a year

For more or less a comprehensive assessment, we need a real trend of supply and demand prices, and will be seen from the statistics of the use of the cloud.

and then the price will be settled.


Everything will affect here! The number of traders who are developers will agree to use the cloud

The cloud will not be useful without MT5 as most traders work with MT5

Because the majority of traders are working on MT4 and they will make up the audience that will need the cloud

MT5 will become really interesting only when the MT4 and MT5 multiplexes will work in parallel

it will be much more interesting to work with MT5 because of its new features!

and when there is an obvious advantage of MT5 platform over MT4,

for example on MT5 there will be no slippages - instant execution and so on.

i can't tell the difference between mt4 and mt5 ; for example mt5 has no lock, there is a cumulative position and that's what most traders don't like


I don't have orders for MT5 yet, it's more of an enthusiast thing

You can ask on any major trading forum (as an example the biggest one ishttp://forum.masterforex-v.org )

If someone works with real accounts on MT5... I assure you nobody practically, but traders + developers will be the consumers of cloud resources

and the developers can write real orders only when they start getting them

Форекс Форум трейдеров Академии «MasterForex-V»
Форекс Форум трейдеров Академии «MasterForex-V»
  • forum.masterforex-v.org
Школа начинающего трейдера Академия МФ (обучение) Брокеры Биржевой лидер Библиотека Индикаторы форекс
 
papaklass:
That's weird. I've had four agents working almost 24 hours a day. Completed 18565 tasks so far and the pay is $0.06. I understand that the pay for the cloud is very low. From my example $0.06 : 18565 = 0,000 .... per task. Where did $7.0 come from? I, on the contrary, thought the service was not expensive, but it turns out I'm wrong.

No comment so far, just the layout ...

of 66 cores

there are about 23 cores working in test mode right now

7-8 hours per day at night and the rest is a pity to give - some cores are on laptops that do not work full time and not every day

23 cores have been working for 5 days

in test mode 5 days for resources accrued

1.11971$

that's $6.66 a month.

and it if you give 23 cores per day for 7-8 hours, if you give without a PREPARATION for 24 hours, it will be for 23 cores about $ 19 - $ 20 a month

--

to get $100 you need at least 100 cores full time

to get $1000 you need 1000 cores

to get 10000$ you need 10000 cores

--

rough breakdown - one core = $1/month



 

one core = $1 per month, per day $0.03

That means that in 24 hours the processing power of e.g.

7 / 0.03 = 233 cores.

Not taken into account the system fee.

That's normal, it would take all those cores to run continuously. I had a maximum of 30 agents logged in yesterday, and it was only for a short time. That's 9 to 15.

So that's who got them all :)

 

As you can see, for 18.09 counted $10.22. For the evening (hours of 18-19) 17.09 - $4.78

Yes, the resources I could consume a lot, because going full overkill instead of genetics + ekspert tested on the competitive tester 11 minutes, ie, theoretically, I can my 20 thousand passes load the entire network at once (for not need to wait for the end of testing all the passes of a particular population), would be the cores, but apparently there is a limit and the rest also remains :)

It's getting expensive + some glitches - during testing since yesterday none of results has returned (initial balance + 0 trades returned, which is 100% wrong; meanwhile local and remote agents work fine). As a matter of fact, I used a network, but it didn't help me, it only eliminated probably the best passes. But nedeyutsya, developers with this figured out.

A price must be reduced, otherwise no one will use. Better yet, leave the network as it is free or some small monthly fee + benefits for those who provide their agents. But I'm looking at it from my own vantage point, of course

 
YuraZ:

no comment on the layout yet ...

of 66 cores

now running in test mode about 23 cores

7-8 hours per day at night and the rest is a pity to give - some of the cores are on laptops that do not work full time and not every day

23 cores have been working for 5 days

in test mode 5 days for resources accrued

1.11971$

that's $6.66 a month.

and it if give 23 cores per day for 7-8 hours, if give without a PREVIEW for a day, it will be for 23 cores about $ 19 - $ 20 a month

--

to get $100 you need at least 100 cores full time

to get $1000 you need 1000 cores

to get 10000$ you need 10000 cores

--

rough breakdown-- one core = $1 a month


It's hard to say yet, I'll wait a month or two. But I have started to doubt the feasibility of my agents participating in the cloud.
 
notused:
Current prices are quite high - in less than a day already counted for $7. Not ready to pay more than $5 + still need the network for multicurrency, not just empty answers with initial balance and 0 transactions.

10-20$ for a full-fledged expert test is quite normal. Another thing is how often such testing is done.

But in my opinion the current prices are very low, and it would be more logical to think about keeping the cloud free for at least 6-12 months.

masharov:

There is also a psychological moment. IMHO, if resource sellers can earn only $20-50/month on a service, many people simply won't bother with the cloud. Perhaps, selling resources to individuals is not the way to go at all, and it is easier to make/rent a server farm and sell their resources. Are there any successful analogues of the project in the world?

I do not know colu as, but for such prices (according to colleagues - $ 1 per core ) and will not think - Either free for all or for a normal price (I meant "normal", not "robbery") ...
 

A single core consumes about 25 watts. If it runs for 30 days without stopping, it eats about 18 kWh. For example, in Moscow the price per kWh is about 2.66 roubles. It turns out that one core consumes 48 rubles worth of electricity. Don't count the other consumers in the PC, and they will give another 50-100 watts.

In the end, the cost is slightly more than $1. :)

 
Interesting:

10-20$ for a full-fledged expert test is quite normal. Another thing is how often such testing is conducted.

For the Championship, the EA has been tested for about a month, maybe more (multicurrency - first in separate pairs, then together in parts, then together in parts, etc.) - at this rate, it would cost me at least $300 per month. Ow to those who are willing to pay that much. Anyone interested? It may be worth $300, but for that price I'll do with my own resources.

Interesting:

It makes more sense to think about keeping the cloud free for at least 6-12 months.

+1, but free forever :) The thing is, consumer voices are not heard at all, some "manufacturers" say it's "cheap".