Multi Data Stream Analysis (EasyLanguage)

[Tams 20]

This thread is about using the EasyLanguage keyword DATA in multi data stream analysis

 

 

 

Data

 

 

Used to specify a particular data series in a multi-symbol chart;

each data series in a multi-symbol chart has a unique Data Number.

 

 

Usage

 

DataN

 

Where: N - the Data Number of the data series

 

Or (for MultiCharts):

 

Data( N )

 

Where: N - a numerical expression specifying the Data Number of the data series

 

 

Example

 

High Of Data2

will return the high price of a bar in the data series with the Data Stream #2

 

 

High Of Data( 2 )

will return the high price of a bar in the data series with the Data Stream #2

 

 

 

source: EasyLanguage manual

[aaa 10]

TAMS

 

I don't understand

 

Is it a way 2 access to a second symbol in a chart ?

 

Or an other resolution of the same symbol in a chart ?

 

if it's a 2 stupid question, please don't answer

 

aaa

 

------------------------------------------

 

Data (Reserved Word)

Disclaimer

 

Reserved word used to reference information from a specified data stream.

 

Remarks

Data is normally used with a number between 1-50 that allows the specification or which data set is being referred to in terms of price values and functions calculations.

 

Examples

Close of Data3 returns the Close price of Data stream 3.

 

Low of Data10 returns the Low price of Data stream 10.

[Tams 20]

you can use this keyword to access additional data stream.

 

the additional data stream can be of a different symbol and/or resolution.

 

 

for MultiCharts users, the resolution can also be mixed:

i.e. time based chart with non-time based chart

 

e.g.

5 min data1 with 15 min data2

5 min data1 with 2401 contract volume bar data2

5 min data1 with 500 tick data2

[RichardTodd 10]

The most important thing to remember with the DATA keyword, which constantly trips people up, is that you are accessing the other data only when events happen in the primary data stream. (the primary one is the one you apply the indicator to when you add it to your chart, not necessarily DATA1)

 

So, if you apply the indicator to a 5-min chart, update-intrabar turned off... then even if Data2 is a 1-tick chart, you will still only run once every 5 minutes, and Close Data2 will be the most recent close for the 1-tick chart, even if you've missed hundreds of 1-tick bars in between.

 

For this same reason, certain optimizations like AverageFC (which assumes it will "see" every bar it needs to, and only once), and any algorithm that needs to "see" every tick will generally not work across data streams and are best avoided.

 

Things are a bit more intuitive when you have update intrabar on, but it still pays to keep this in mind. Most errors I come across in multi-data indicators stem from this issue.

[BlowFish 10]

Richard makes good points. In MC you can minimise the impact of this by having a 1 tick chart of the fastest instrument as data1 to 'clock' your indicator. Be aware that it's at time like this that having fairly well optimised code is important (yes I know I keep banging on about that )

 

You can hide the 1 tick bars if you don't need them. The problem with this approach is irregular scaling, the charts X scale is done in regards to data1.

[aaa 10]

ThanX to TAMS + Richard Todd + Blowfish 4 your Xcellent info

 

2 illustrate what have been said, here are graphs showing the problem of bar interval

 

it is not smooth + we have 2 wait the end of the bar 2 draw the avg

 

Is it the same problem that ADE which send the info only at the end of the bar ?

 

On chart 3 An other solution is to plot a 240 ( 60/5) * 20) avg directly on the 5 mn chart but it is not accurate

 

 

 

Edited November 28, 2009 by aaa

[Tams 20]

this chart illustrates the data1 and data2 application differences.

 

the resultant calculations are the same,

but the visual presentation is different.

 

 

Here's the code for data2 application.

 

note that you should not use FC functions when calculating secondary data.

 

inputs:  
Price( Close data2 ), 
Length( 9 );

variables:  
MA( 0 ) ;

MA = Average( Price, Length ) ;

Plot1( MA, "Avg" ) ;

 

 

 

[aaa 10]

Is there a way 2 eliminate steps scale ?

 

1 way is 2 plot directly the avg from data2 to data 1

 

That's what I've done here (always with step)

 

http://www.traderslaboratory.com/forums/f56/multi-data-stream-analysis-easylanguage-7165.html#post82170

 

I've read that there is a way 2 compute an avg from a hiher time w/o data2 with a limitation = it must B a multiple

 

May B on a 1 mn chart take the close of all the 5th mn and then compute the avg of the last 20 * 5 mn close ?

 

And then add each mn close from 1 to 4 th mn to the close of the last 5 mn ?

 

PS Sorry I'm not good with maths

 

I'm quite sure that TAMS will ask a mock up :haha:

[Tams 20]

...

I'm quite sure that TAMS will ask a mock up :haha:

 

 

hahaha the steps look nice

 

:-)>

 

 

if you turn the data2 plot from a line to dots,

you will see how the calculations is done at every bar... (at the faster fractal)

Edited December 5, 2009 by Tams

[aaa 10]

inputs:  
Price( Close data2 ), 
Length( 150 );

variables:  
MMX( 0 ) ;

MMX = XAverage( Price, Length ) ;

Plot1( MMX, "Avg" ) ;

value1 = Text_New( D, T, MMX, "~") ;

text_SetSize(value1 , 32 );

 

Tams, Would U mind 2 load this indicator modified ?

 

As U C there is a gap beetween each "dots" and the line 'hides" the gaps

 

Of course the last bar is in RT (not like with ADE) but the "dots" between each step are missing

 

It is not smooth when we plot a 150 avg 60 mn on a 1 mn chart

 

It is impossible 2 trade this way

 

No steps ( Xcept 2 heaven !)

 

In fact I know that it is possible with a function

 

But I'm not able 2 do it

 

Can someone code my "idea", if it is a good one of course !

 

I must say that I don't use any indicators 4 my trading except 3 avg

 

30

 

150

 

750

 

and my 750 avg that I use is not accurate

 

Help will B very appreciated

[Tams 20]

the ~ is nice.

 

you don't need the plot...

 

;-)