Testing 'CopyTicks' - page 21
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
On the almost empty Alpari marketwatch, the copitix brakes just as much as on BCS.
On Robo, everything is more than an order of magnitude faster.
Results on USDCHF M1, CalcLength from the indicator https://www.mql5.com/ru/code/16537:
The results are constantly fluctuating + single threads rule is not observed - on one of the streams without the pick and mix for CopyTicks, on the other with the pick. The demo is not opened on BCS, but the main thing is that there are no stacks on RoboForex.
Strange measuring code. A lot of unnecessary things were measured, but not the CopyTicks request time:
{
const ulong StartTime = ::GetMicrosecondCount();
int X0;
int Y0;
BARS bars(this.Chart);
TICKSPICTURE TicksPicture(this.Chart, &bars, X0, Y0);
this.SetProperty(::OBJPROP_XDISTANCE, X0);
this.SetProperty(::OBJPROP_YDISTANCE, Y0);
TicksPicture.Fill(ColorBid, ::BID);
TicksPicture.Fill(ColorAsk, ::ASK);
TicksPicture.Fill(ColorSpread, ::AVG);
TicksPicture.SendToResource(this.Resource);
::Comment("LastCalcTime = " + (string)::TimeLocal() +
", Ticks = " + (string)bars.GetAmountTicks() +
", CalcLength = " + (string)((::GetMicrosecondCount() - StartTime) / THOUSAND) + " ms.");
::ChartRedraw(this.Chart);
return;
}
In any case, we optimise the call to copy ticks. It takes a lot.
Results on USDCHF M1, CalcLength from the indicator https://www.mql5.com/ru/code/16537:
Problems on Alpari Real and BCS Real. It is very easy to open the real
Forum on trading, automated trading systems and trading strategies testing
Limit order slippage statistics at the exchange
pivomoe, 2016.08.25 15:15
Start making a demo mt5 account with bx. You will receive an email with a link to the distribution. At the stage of server selection choose not a demo server, but for real trading. Create an account with arbitrary data. Create a certificate. All you have a real account with zero balance with real quotes and history.
The results are constantly floating + the rule of single threads is not respected - on one without stack the threads (where there is no need to do selections and mergers for CopyTicks), on the other with stack. The demo does not open on BCS, but the main thing is that there are no streams on Roboforex
Weird measurement code. Measured a bunch of extra stuff, but in no way the timing of the CopyTicks request:
That's not up to me. It's all measured together. The only difference is the trade server. So the brakes are only related to CopyTicks.
In any case, copyTicks call is optimized. It takes a lot.
Copyix in its current form is very inconvenient. For example, it's absolutely unclear how to get the tick that was before from.
Why cannot we return the index on the base array and work with the base as with an array? Let's leave the problems of adding new data to the database entirely up to the user. Let him sort it out himself, if something goes wrong. Right now, working with Copyix is, well, very inconvenient. I seem to be one of the few people who uses it very actively. And I can say about this quite responsibly.
Results on USDCHF M1, CalcLength from the indicator https://www.mql5.com/ru/code/16537:
Results are constantly fluctuating + single threads rule is not observed - there are threads without picks and merges for CopyTicks on one of them, on the other one with the pick. On BCS the demo does not open, but the main thing is that on RoboForex there are no stacks
Strange measuring code. Measured a bunch of unnecessary things, but in no way CopyTicks request time:
{
const ulong StartTime = ::GetMicrosecondCount();
int X0;
int Y0;
BARS bars(this.Chart);
TICKSPICTURE TicksPicture(this.Chart, &bars, X0, Y0);
this.SetProperty(::OBJPROP_XDISTANCE, X0);
this.SetProperty(::OBJPROP_YDISTANCE, Y0);
TicksPicture.Fill(ColorBid, ::BID);
TicksPicture.Fill(ColorAsk, ::ASK);
TicksPicture.Fill(ColorSpread, ::AVG);
TicksPicture.SendToResource(this.Resource);
::Comment("LastCalcTime = " + (string)::TimeLocal() +
", Ticks = " + (string)bars.GetAmountTicks() +
", CalcLength = " + (string)((::GetMicrosecondCount() - StartTime) / THOUSAND) + " ms.");
::ChartRedraw(this.Chart);
return;
}
In any case, we optimise the call to copy ticks. It takes a lot.
Please clarify what you mean about single threads?
Another question, in order for the ticks to come as fast as possible, it is necessary that the cup is not open in the terminal and there is no subscription to the cup update event from the EA/indicator?
From reference:
Output speed: the terminal stores for each symbol 4096 recent ticks in the quick access cache (65536 ticks for symbols with the stack running), queries to this data are fastest.
Please clarify what you mean about single flows?
Forum on trading, automated trading systems and strategy testing
Testing 'CopyTicks'
fxsaber, 2016.10.13 10:18
There is an MT nuance with ticks where the tick history itself is overwritten retroactively because of multiple sources of ticks received.
Thank you, but do you know about the glass and the speed of receipt?
No, unfortunately. Renat argued that the glass is permanently streamlined for the whole marketwatch. But that's not an appropriate solution for most situations.
This is how you should test CopyTicks:
//+------------------------------------------------------------------+
//| |
//+------------------------------------------------------------------+
void OnTick()
{
ulong from =(TimeTradeServer()-1200)*1000;
ulong ticks =GetMicrosecondCount();
int records=CopyTicks(_Symbol,ExtArr,COPY_TICKS_INFO,from,2048);
ticks=GetMicrosecondCount()-ticks;
Print("Time: ",ticks," msc for ",records," records");
}
Here's the output in microseconds: 95 microseconds per sample of 2048 INFO ticks over the past 20 minutes
Please clarify what you mean about single threads?
The bid/ask flow and the full ticked flow with volumes and last/traded prices are two big differences.
Cardinally big differences.
Another question, in order for the ticks to come as fast as possible, it is necessary that the glass is not open in the terminal and there is no subscription to the event of updating the glass from the Expert Advisor/Indicator?
Openness in the terminal or subscription in the EA does not matter.
If the symbol is in marketwatch, then the terminal receives all the full tick stream with the glasses unconditionally.
But most importantly, the sample rate calculations above are irrelevant. They are so clumsily done (anything but CopyTicks time is measured) that it is even surprising.