Global drawing lines etc on tick charts

[pjsmith]

Hi,

Not sure there is much that can be done without some new more intelligent line re-drawing on the other time frames, but I'm recently struggling with drawing lines on 2 different tick charts via global drawing object. I've attached an example. The issue of course, in the simplest form (I believe), is that the x axis on tick charts is not consistent, nor predictable into the future, so simply re-rendering the chart on the other time frame, you could expect some oddities. However, we must be able to do something better? For example, in the attached image, the trend line, drawn on the higher timeframe touches all the bottoms of the bars (normal candles), yet in the 100 tick, they are way off. At first, when you are drawing lines on historical charts, it does not look so bad, I think because start and end time are already recorded. But if you are trying it live, it's completely useless

Anyone have any suggestions to work around this? Perhaps Ninja in the future could implement something with a bit more intelligence to resolve this.

[EDIT] - I thought I'd try on time based bars which I rarely use on short time frames. Expecting perfection, but actually, it's a bit rough there, too, which I truly don;t understand! Further example attached. I've attached a 3/4 example too. This one is drawing a line on current price on the larger timeframe, seeing what it looks like on the smaller time frame, then, without moving the line on the larger time frame at all, just wiggle it a bit, and see how it has changed on the smaller time frame! Again - I kind of understand some of the reasons this would happen, but hoping Ninja might be able to do something more intelligent in the future,.

Thanks.

[NinjaTrader_ChelseaB]

Hello pjsmith,

Global drawing objects will have future anchor values which do not match between charts because future anchors adjust per chart and are simply estimates.

Unfortunately, this is known expected behavior and I am not aware of any work around.

Do you have specific logic you would like to request be implemented?

This thread will remain open for any community members that have suggestions.

[pjsmith]

Thanks. I'm not sure how it would be implemented, but given most trend lines have anchor points on the bar highs/lows, etc., perhaps any future tool could use price anchor points instead of time, to more accurately reproduce the lines on other timeframes? I've not really thought that through carefully, but just throwing it out there...

[brucerobinson]

#ChelseaB
"
Global drawing objects will have future anchor values which do not match between charts because future anchors adjust per chart and are simply estimates. "
Could you kindly elaborate on this statement?
I spent many hours with this in the past without a proper explanation and still face the same problems today.
It means such lines cannot be relied upon.
Kind regards,

[NinjaTrader_ChelseaB]

Hello brucerobinson,

We are not able to predict the future or which bar the anchor will be on.

How long does it take for 100 ticks to be received? (It's not a question that can be answered unless you can see the future)

This means when an anchor point is past the last char bar in the future, we are making an estimation of how long in the future that is.

[pjsmith]

NinjaTrader_ChelseaB But, if a bar touches a line at 56.50, does it not touch the line at 56.50 on all time frames? I am thinking that sort of intelligence is one thing that could help.

Even with time based anchoring, all you would need to do to dramatically improve the situation is re-render the line as new data comes in, and the future time is mapped. You'll find that if you plot these types of lines historically, they are not that bad. If you run them into bars not yet plotted is where the issue is. BUT, if you simply wiggle the original line without moving it, hey, much better... Perhaps NT could simply 'improve' it's positioning, by re-rendering lines based only on the portion of the line that has anchor data (like a ray - In fact, I'm going to test rays instead of lines to see if that is a work around, thinking about it), and re-rendering/re-positioning lines drawn into the future as new data becomes available. It does not seem too hard, if you put a but of thought behind it...

[brucerobinson]

Both - I've got no intention of muddying the waters, off at a tangent or 'hijacking' & if I'm misunderstanding, either or both of you feel free to ask me to bow out.

But Chelsea's reply to my querie, & PJ's further comment, and my own experience of a similar/same findings makes me think rather not.

On Chelsea's reply re: predicting the future for anchor points, in my 'example' I was not drawing lines with future anchor points beyond the last bar.

I don't see that PJ was either, but may have misunderstood.

My experience was as pj reports 'if it hits 56.50 on one shouldn't it on all?'. Line drawn with all anchor points drawn in 'history' not beyond current bar.

And my experience was uncannily similar - seemed something to do with whether originate the global line on LTF or HTF, and 'wiggling it' caused it to change.

Anyhow, by all means tell me to duck out if not helpful.

There's an extensive thread on it somewhere with lots of charts showing the issue I had, irrefutably.

But may be I misunderstand

Edit: ah, I think I've figured out my misunderstanding yet I think we end up at the same place. Pj was originally talking about Lines hence future anchor point which Chelsea correctly responded to, I confused with Rays which I use/was my experience and don't require anchor points in the future hence not the cause.

Pj now considering Rays - the bad news: I can confirm an unresolved similar issue on Rays for which the cause is not the 'future approximation' issue 'cos can draw with all anchors in the past, & still get the 'doesn't hit 56.50 on all charts' problem. It can be so scarily out after not much time as to be useless & worse, confidence sapping.

As mentioned earlier, 'fiddling around' could correct, or at least ameliorate it, but I was not able to isolate to step by step a workaround.

Pj, if you're minded to pursue it, with Rays, I'd search my name for the thread on it - I put a lot of time into it before hitting the blocks, might help/save you a ton of fiddling if it is the same topic

[NinjaTrader_ChelseaB]

Hello pjsmith,

When you mention:
"If a bar touches a line at 56.50, does it not touch the line at 56.50 on all time frames?"

This is not a question I can answer with a yes or no. If an anchor is on a bar (and not in the future) and is at 56.50, then yes, since that anchor has that price we can know the line would be touched at 56.50.

If the anchor is not on that bar and we are considering slope and one of the anchors is in the future, then no, both charts are not guaranteed to hit the slope of that line at the same time as the slope is calculated by both x (time) and y (price). If the x and y are not the same, the slope is not the same, and a point of the line that is not the anchor may be at a different price.

When you mention:
"given most trend lines have anchor points on the bar highs/lows, etc., perhaps any future tool could use price anchor points instead of time, to more accurately reproduce the lines on other timeframes?"

Drawing objects do have a price anchor point. The price of the anchor does not change. Only the slot position changes as NinjaTrader estimates which bar slot position corresponds with the anchor time.


brucerobinson,

Is your inquiry unrelated to drawing in the future?

Are you drawing a global line where all anchors are in the existing bars and the price and time of the anchors on each chart is different?

[brucerobinson]

Hello Chelsea,
Apologies for the delay, not getting Notifications.
I'm drawing Rays, so extend into the future, but the anchor points are in the bars, I can't recall if the anchor points were different or not but the line ahead of the anchor points would intersect the bars at very different places.
It was that aspect of PJ's post that caught my eye.
I'll search for my original Post. I gave up on it at the time but continue to notice the issue.
thanks & kr

[brucerobinson]

Here is the thread on what appears to be similar.

https://ninjatrader.com/support/forum/forum/ninjatrader-8/platform-technical-support-aa/109423-ray-global-drg-obj-appears-to-be-incorrectly-drawn

Suggest to take a look at Post #1 and Post #11, particularly their Attachments as they make it plain to see - same charts, same run & rise, anchors in the bars, wildly different intersections ahead.

[pjsmith]

brucerobinson Thanks for the pointer to that thread! Explains a lot. Disappointing it is not better, but there we go...