Coding Cumulative Delta & Volume Delta Indicators for Multicharts

[tucciotrader 10]

Is there anyone who figured out how to code Cumulative Delta & Volume Delta indicators for Multicharts??

 

In Multicharts prebuild indicators we have "Volume up" and "Volume dn", that are only uptick or downtick, they don't consider the real volume at bid or volume at ask.

An uptick can be made by a 1ctr or by 2000ctr that's not definitely the same thing.

 

DELTA VOLUME: it is the difference between volume traded at bid and volume traded at ask in a definite time frame (es 1min 5min and so on)

 

CVD: The cumulative volume delta is the algebraical sum of the single delta volume values

 

By the way what I'm telling now is, I'm not a programmer but just a trader but in Multichart we have:

- the possibility to plot the bid and the ask historically

- the possibility to have a tick by tick "last trade" with volume

 

It seems it's all we need to build those indicator I think...

 

I would love a line that tells the cumulative delta tick after tick!

 

Do you think there is any difference by setting the chart from 1 tick to 1 contract (Build Volume on: Trade Volume) and use UpTick and DownTick so that I can understand better Volume at Bid and Volume ad Ask??

[Tams 20]

EasyLanguage is not the best tool to do those jobs.

[tucciotrader 10]

can it actually be done anyway? I mean something like On Balance Volume

[Tams 20]

it can be done... but...

it is not just the language, the datafeed has to be accurate as well.

Most retail datafeed do not have the resolution to make such an analysis meaningful.

[tucciotrader 10]

ok so, if I have for example: date,time,price,volume on a tick-by-tick basis that is not enough? I mean by using uptick and downtick...have look at the attached file, please

[tucciotrader 10]

Also, I have noticed that an Italian trader (and programmer) have stopped by here and managed to code a CVD for Multicharts (along with the volume profile hisotgram I am still searching for the code)

[Tams 20]

ok so, if I have for example: date,time,price,volume on a tick-by-tick basis that is not enough? I mean by using uptick and downtick...

..

 

how could you know the uptick is an uptick,

and a downtick is a downtick?

 

have you heard of aggregation?

 

:-)

[tucciotrader 10]

yeah you are right :doh:

[natedredd10 10]

"Gomi" has done this in Ninja straight out..

Surely it can be done in MC...at least kinda.

Surely not a 2 day project though.

Before you waste massive amounts of time I would REALLY run through the logic that volume delta makes any sense at all.

To me its basically nonsense. Thats not to say that something like Market Delta has not figured out a better chart to visualize tick data than century old candle shit.surely they have..

You should however question that the bid - ask means jack shit in reality in a double auction.

[tucciotrader 10]

yeah I noticed that...I read the entire discussion about this in a thread over here and I have come to think the same...also the high frequency trading mess all it up!

[tucciotrader 10]

my last question is: how much difference is between On Balance Volume and Cumulative Volume Delta?

 

On Balance Volume can still "feel" buy/sell pressure in a way? (allthough it must increase or decrease in ticks)

[playafh69 10]

I beg to differ. I trade using Bid/ask, and it actually works quite well. Its true that you need to have a quality data feed, but most people don't realize that you have to use it on the right instrument. Bid/ask does not work well at all on the YM or TF, simply because there isn't enough institutional activity.

 

Also, i hear a lot of people complaining about how HFT messes everything up. But, again, I fail to see the validity in this. If a bank buys 2 x1000 lots or 2000 x1 lots, they're still buying 2000 contracts, and you will see that regardless. HFT is simply the speed or rate that the trades occur, thats it. There might be increased momentum because of it, but they're still buying and selling 2000 contracts, and you will see that.

Banks and institutions don't do HFT trading to compete with newbie day traders, they do it to compete with other banks and institutions to get better price improvement.

[BlowFish 10]

It can not be done accurately with easy laguage. There are some architectural issues with multicharts that I have made Tssupport aware of. I wrote some code that you will find in the volume splitter thread that 'works' with this caveat.

[Tams 20]

It can not be done accurately with easy laguage. There are some architectural issues with multicharts that I have made Tssupport aware of. I wrote some code that you will find in the volume splitter thread that 'works' with this caveat.

 

MultiCharts was re-written in 2007.

If the issue was pre-2007, probably they have taken your report into consideration and fixed it.

[ochie 10]

There is an architectural issue (which Andrew Kirillov confirmed) reported to the TS Support forum in 2009. It appears it has to do with the async reading of the CurrentBid/CurrentAsk while incoming ticks are still being processed. It was originally thought that the problem was with IntraBarPersist. Don’t know the current status of this and it may still be an issue. I moved to an alternative. Here is the link to that thread. Post subject: Why I don't trust "IntraBarPersist"

[BlowFish 10]

MultiCharts was re-written in 2007.

If the issue was pre-2007, probably they have taken your report into consideration and fixed it.

 

It still appeared to be there a couple of months ago. bestbid & bestask are non starters for this sort of thing (they peek at current best bid and best ask not the value it was when the tick occurred). Most of the code out there (including stuff I have posted) use that approach. Using multiple data streams appear to have odd sequencing issues now and then. From what I understand it can not be guaranteed that all processing in every study in every chart is done for tick n before tick n=1 gets some attention. It's all a bit tricky as ticks can occur at exactly the same time. I dunno all I can say is that I wrote some logging stuff and there where rather too many anomalies for me to be completely happy.