Or is a separate EA for every _Symbol or maybe _Period better or more precise for some reason?
Logic suggests using one EA is more efficient than attaching the same EA on 20 different charts because the simple act of loading many different charts seems to lag the platform.
Is that all there's to it?
However seems also that having a multi strategy (like 10+) EA that runs on every timeframe on every instrument (70+) might lag things up and become inefficient somehow (even using magic numbers etc).Even if you had a supercomputer, would it be worth it?
Is there any conclusive evidence that trading every instrument/timeframe from the same EA better (at least generally/most of the time) than attaching 20 EA's to 20 charts instead?
Like for example if I trade an EA on 1m and 5m should I attach it to work individually for each period or should I code the 1m version to also trade on 5m? Any advantage either way?
Depends on your coding skills / level.
You have to love what you do.
lol yes, and do what you love.
So its generally better (if you have the skill) to just code everything on the same EA?
When EA is designed to perform a single task (trade plan, symbol/timeframe settings), and invoked multiple times, this can be considered as somewhat more reliable, because any unpredicted situation in a state will damage only once instance, while all the others left unaffected and running (and you can take it for granted that every EA contains bugs, not to mention the terminal itself). But you're right about efficiency. One EA can make the job with less overheads.
There are cases, when EA should handle multiple symbol/timeframe configurations, for example, because its trading strategy relies on hedging (I mean basket trading, spread trading, etc) and multiple positions are required to get net profit.
any unpredicted situation in a state will damage only once instance, while all the others left unaffected and running.
BTW, if your question is about MT4, it uses a so called trade context internally, so efficiency of both solutions are limited, but when you run a single "multipurpose" EA you can probably adjust order priorities. When many EAs are running they compete for the trade context, and you don't have control on how every one of them preempt the others - this decides the terminal itself.
P.S. To be more precise, the trade context is not limited by one order, but 8 simultaneous orders (if someone's multisymbol basket is larger, [s]he is in trouble ;-).
It's actually 12/16 simultaneous orders.
Forum on trading, automatic trading systems and testing of trading strategies
Renat Fatkhullin , 2015.07.04 12:38
In MT4, you can simultaneously conduct up to 16 transactions (12 simultaneously from different experts), so that trade context busy is very difficult to obtain.
In MT5, there are many more transactions, especially taking into account the asynchronous mode.
That's strange. You're referencing a post from 2015, but here is another one from 2016 (in russian), where it's 8. Also I saw posts from others, who tested this matter and got exactly 8 limit.
I never tested, so they probably changed after 2015 or it was wrong information. I will update my bookmark. Thanks.
Unfortunately you will still see just the one instrument in the tester.
That is a limitation when you try to test those hedging strategies.