"Warped" trend lines

[fabrizio]

Echelon, Lancer

I have trend lines drawn from June and still there , with no skew or drift at all.

This is my experience

With AGET stand Alone I havd a TL from the 12 of March and did not moved at all till June.

Furthermore its the first time I hear this problem on this BB.

[Echelon]

Get standalone seems fine, it's the realtime that has errors

I too have the GET standalone version and have seen no problems with that software.

It is the realtime version that shows this flaw. Take a simple test, bring up a 30 minute chart, draw a few trend lines, duplicate those lines to creat a few trend channels. Then, simply click on a 2 minute time frame for that chart.

Draw a verticle line between the origin of the trend line channels you drew and move that verticle line to the right and watch the "warping" of your previously drawn parallel lines.

If you don't get that result, I would be interested to know what version of the software you are running, and what specific settings like scale, etc., you have.

David Smith

[Lancer]

David,

Originally posted by Echelon
Are you saying that the only trend lines that are accurate are the ones that are drawn in, and then subsequently viewed in the same time frame? All others are inaccurate?
Please tell me more about your experience with this limitation.

David Smith

That is correct. Lines created in one time frame can't be relied on for slope or position if the chart time period is changed. Also, if you create lines in one time period, change time periods, then copy lines, those copied lines will be skewed when you change back to the original time period. For me, that's not an issue because I use a set of charts with set time periods, and I never change the time periods of charts that have lines.

The thing that gets me is the creeping skew and migration, even just overnight. Also frustrating is the inability to create, modify, or check lines by slope value; (can do in Metastock and others). So, for lines older than a day on an intraday chart, I just can't be sure if they're still good or not. To be sure of good lines for the day, I check reference points and reset all lines each morning prior to the open. I also use standard charts for lines. Unfortunately, lines in standard charts can't be saved, and have to be re-created each day.

I agree, lines (support/resistance) are fundamental and vital to trading. I'll take good lines over 100 technical studies. No doubt eSignal will eventually get to fixing it, but right now I'm thinking they are on a schedule to complete the greatly appreciated major evolution of the product, which I think might reach a plateau at version 8.0. After that, I'm hoping they'll pull back a bit from adding new, and go into product maintenance mode to pick-up all the outstanding punchlist items, including attention to lines.

[Echelon]

Priorities?

Regarding: "No doubt eSignal will eventually get to fixing it, but right now I'm thinking they are on a schedule to complete the greatly appreciated major evolution of the product, which I think might reach a plateau at version 8.0. "

Well, what does someone who represents the Esignal company have to say on this matter? Please read through this thread and let us know where your current priorities lie.

If indeed the basics are going to be neglected for things like paper trading capabilities, intergrated brokerage abilities (all great things to have eventually) over the absolute basics of simple, accurate trend lines, then I will indeed start my search for another charting/data program. I only trade the S&P futures market, so it shouldn't be that difficult to find another company who has software that can draw accurate lines for God's sake.

If, on the other hand this is an issue that is being dealt with right now and in upcomming builds, I think your customers deserve the courtesy of a good business reply. And, not simply lip service. It is amazing to me that you have been pointing out this basic issue to no avail since advanced charts were created, and nothing has been done thus far.

What's the company response?

David Smith

[fabrizio]

Re: Get standalone seems fine, it's the realtime that has errors

Originally posted by Echelon
I too have the GET standalone version and have seen no problems with that software.

It is the realtime version that shows this flaw. Take a simple test, bring up a 30 minute chart, draw a few trend lines, duplicate those lines to creat a few trend channels. Then, simply click on a 2 minute time frame for that chart.

Draw a verticle line between the origin of the trend line channels you drew and move that verticle line to the right and watch the "warping" of your previously drawn parallel lines.

If you don't get that result, I would be interested to know what version of the software you are running, and what specific settings like scale, etc., you have.

David Smith

David

I have followed your hint.

AFAICS no distortions or drifting , sorry, or better happily.

Maybe the problem exist but I'm not able to replicate it or I honestly do not see it .

I have the 7.4 .My built is the 608.

The AGET STANDALONE I mentioned is a RT.

The only thing I can add is that in the past with aget RT standalone feeded by CQG I had some graphic card conflicts ( shadowed ellipses not properly traced etc.) and I remeber that couple of times I had a distorsion of the lines.

That's all

Cordially

[ACM]

Lancer (and Echelon)

Also, if you create lines in one time period, change time periods, then copy lines, those copied lines will be skewed when you change back to the original time period.

While not suggesting it is not happening at your end(s) I am not seeing this and the other problems mentioned and as I said in a prior message I too use trend lines regularly switching between intervals.
Here for example is the process suggested by Lancer. Plotted a trendline on 15 min chart



Switched the chart to 5 minute and plotted a parallel



Switched the interval back to 15 minute. Lines are still parallel.



Switched to 30 minute and lines are still parallel well into the future.



Lancer mentions overnight migration yet I have charts that are months old and the lines are spot on where they were first placed
Again I am not implying that the problems mentioned are not happening elsewhere, but that perhaps other factors may be contributing.

In his message Fabrizio mentions a problem he was having with Ellipses. Here below is what he was showing at the time on his screen



By any standard one could have easily assumed that there was a bug in the software. That was not the case as the display issue was fixed by simply changing the video card.

Another example. I have TS2k installed on two machines. On one of them trendlines plotted past the last bar are skewed whereas they work without any problems on the other. However if I drop the graphic hardware acceleration to its lowest setting on the first machine they work perfectly also on that one.
In this case can one state it is a software problem? Not really as the very same software works without any problems on another machine.

All this to say that before going about pointing fingers we should perhaps try to find out what factors (external or not) may be causing the differences we are seeing.
Just an opinion

Alex

[Echelon]

Video Card?

Ok, perhaps you are onto something.

I checked my video card and settings and found the following: (see attached screen shot).

Please everyone else who can and can't reproduce the problem, post your video card, driver, etc.

David Smith

[Lancer]

FWIW, attached is a detailed report (certain to cure insomnia) on the video adapter, AGP bus, and DirectX for the PC here running eSignal. Next week I can post some chart images if that's useful to anyone.

[Echelon]

No match thus far

Well, from what I can see we are both running totally different video cards/adaptors. So there is nothing there that would lead me to believe that is the source of conflict yet.

Anyone else who does or does not have this error with trend lines should post here there video card/adaptors so that we can see if there is any validity to the notion of a hardware/software conflict.

David Smith

[fabrizio]

Gentlemen

I'm using a ati radeon 9000 driving 2 LCD screen on the machine which gived problems. An update of the driver resolved the issue

The analisys machine is using a Matrox Pharelia 128 still wich not updated drivres ( nov 2002) . Dribves 3 LCD screen 2 Dvi +1 non DVI and It never shwed problems for what in object.

ASAP I'll try to generate a more detailed report for your info.

One simple note: one 17.000 posts yours of more than 6000 BBmembers are the only two showing this particular malfunction.
Do you agree that the stat. relevance of the occurencce shows something? Something interesting but not " 100%" resident on the soft.
I'm confident that the solution will be quickly found.

[Echelon]

Driver issue?

Interesting, you say you had the problem on two computers that were running different video cards/drivers? And, when you updated the driver on one computer, that resolved the problem on that computer?

Did it completely resolve the problem? And, what about the other computer you are running, does it still have the problem? I assume that you tried upgrading the driver for that system as well.

Also, Alex, what video card/driver are you running and not getting the problem?

I took the liberty of upgrading my video card driver to the latest available after reading these post responses, but I still have the problem with trend line warping in time frames other than that which they are drawn.

David Smith

[fabrizio]

Re: Driver issue?

Interesting, you say you had the problem on two computers that were running different video cards/drivers? And, when you updated the driver on one computer, that resolved the problem on that computer?

Actually no. I have 3 machines. 1 had problems (the ATI one) resolved by updating the drivers. The other (Pharelia one) never had problems.
Note: The ATI one is used only for operations and do not run ESIG. now. So I should load Esig on it and see what happens.

The Analisys one (the Pharelia) is totally dedoicated to and only Esignal ( as was before for CQG) and never had One single problem.

Consequently : changing video card =no problems
Updating drivers = partial solution of some heavy problems ( as shown by my posts posted by A. Montenegro)


I took the liberty of upgrading my video card driver to the latest available after reading these post responses, but I still have the problem with trend line warping in time frames other than that which they are drawn.

So maybe the conflict cannot be solved.
I.e.: I stated that i still run on Pharelia a very old driver: this happens because every single time I' v try to update it I had problems, malfunctions, freezing....
So I stick with the old Driver.

Other Example: On Pharelia I had problems on AGET RT% stand alone feeded by CQG........

So -I'm an end user and not an IT specialist- those kind of issues are tangled in finding "why" yet simple to solve: new hardware.

I got a survival rule: I do not want to spend more than (n-1) amount of time on this problems.
So I give a precise "clean " task to each machine and configure it for that, keeping "light" the one that should performe the analisys , leaving "dirty jobs " to the "operational /dispositive "one.
The third is just for redundancy and back-up. ( there is even a fourth lap top which is not in LAN)
Rather than "consume"my attention in finding solutions for hrs and days ( which eventually will draw Time and Emotional Capital from my trading) I prefer to spend couple hundred of bucks and start from scratch.
And recycle what I have on another machine that perform other Minor tasks.

Just for your Info: I'm shifting the Pharelia on the other machine and getting a NVIDIA Quadro for the Analisys one. The ATI will be traded in.

Conclusion: with the nowadays prices of Hardware ( even used) the trade off is always convenient.

[Echelon]

Driver issue?

"Just for your Info: I'm shifting the Pharelia on the other machine and getting a NVIDIA Quadro for the Analisys one. The ATI will be traded in."

That should prove to be interesting as my current card is an NVIDIA GForce4 with two monitors. If I were to get rid of it, I would probably go to the NVIDIA Quadro as well.

Keep me updated as to your experience.

Thanks,

David Smith

[Lancer]

Fabrizio,

Originally posted by fabrizio
One simple note: one 17.000 posts yours of more than 6000 BBmembers are the only two showing this particular malfunction.

I haven't done a BB search so I'll take your word for it, but from that I wouldn't jump to the conclusion that only two eSignal users can see line skew and drift. I've seen line issues discussed here before, so this isn't the first time. Also, not all users seeing line issues post here. I know of one guy running a Matrox G450 that uses a lot of lines and sees the same line skew and drift that we've described here. I've seen it myself on his setup; (same as what I see on my completely different setup). He hasn't posted here about it. Also, not all eSignal users use lines extensively. I'm particularly sensitive to the matter because I use a lot of lines. The user that doesn't use trendlines or only ever draws a few trendlines isn't likely to recognize any problem, and isn't likely to post here about it. That doesn't mean the issue doesn't exist for that user if he/she looked closely.

[fabrizio]

> but from that I wouldn't jump to the conclusion that only two eSignal users can see line skew and drift.I've seen line issues discussed here before, so this isn't the first time

Sorry Lancer but I do. This is the universe made by # of posts & users. If in universe I have (n) occurences ,
[ (n)/ Universe= statistical relevance ] since I do not believe that if someone has such a problem do not drop a line to report it.

I think you share my point.
It is pretty unlike that many people:
do not post or try to solve a problem ,
contestually many people do not use the lines and at the same time many people do not recognize drifting or skewed trend lines.

I'm not saiyng nor I dare to deny your experience, just saying that I had a similar one and was much likely due to a card conflicft.

Much likely I say because I cannot prove it by 100%.

But the important issue is that a solution must exists, for you experiencing the problem and I do not.

Cordially