More Mesa Bugs!

[jbergeron]

I'm getting a little distressed. I opened a previous thread several releases ago regarding the bug in the Mesa package's sinewave and leadsine indicators. This thread can be found here: http://forum.esignal.com/showthread....highlight=mesa

I was notified that is was taken care of in build 564 (I think, it may have been a later release). Regardless, it wasn't. And I have also found bugs in the sines, phase, and mode now.

Please refer to these images. Note that none of these indicators are plotted with EFS code. These are the stock indicators that came with the MESA subscription that I have been paying for and unable to use for almost 5 months now.

This image is of the MESA indicators running in realtime throughout the day on the chart. As you can see they are not working as they should.


This image is of the same chart above, with the indicators reloaded. If these indicators worked correctly, both charts should be identical. The image below is how these indicators *should* look.




So, essentially none of these mesa indicators are being calculated correctly realtime. But seem to work okay when initially loaded for the previous bars. I do not know if the Kalman Filter, Instantenious Trendline, or Dominant Cycle indicators are effected. I would assume they are, since the mode is incorrect.

I look forward to your prompt response, as well as a prompt resolution to this matter. I am unable to develope my systems until this is fixed.

For what it is worth, I highly advise anyone using the MESA indicators to not use them until these issues are fixed.

Information for the developers: I am running 7.3 Beta (Build 573) 3/26/2003.

Sincerely,
Joshua C. Bergeron
jcb@othernet / #ESignal

[ScottJ]

We think we've isolated the realtime skewing problem to the way the MESA dll files operate w/in eSignal. Not sure if it impacts historical data but we have more testing to conduct next week.

So far, the issue seems to crop-up whenever multiple instances of the same MESA indicator are being used during market hours. If you run just one instance, it seems to function as it should or if you do your analysis after hours. This includes running the same study on different charts. See if you can confirm that as well for us on Monday.

We've also tentatively learned that this skewing problem occurs on at least one other platform that MESA is integrated with. We are in contact with MESA engineering and we hope to have more definitive information next week.

Thanks.

[jbergeron]

Great! Thanks for the information. This explains why this system did so well on Thursday (when I only had them implemented on one chart) and poorly on Friday (when they were on two charts).

I thought for sure I didn't notice any problems with the indicators on Thursday. I'm glad I'm not crazy.

I guess will continue using them on just one chart, until this is resolved.

Cheers,
Joshua C. Bergeron

[jbergeron]

Well. The last few days I ran MESA only on a single chart, with a strategy that calls MESA functions for only that particular symbol.

Is this bug related to multiple symbols of the same indicator (IE, on different charts), or strictly multiple invocations of the same indicator (and even the same symbol)?

Here's some screenshots from today that indicate problems when running on even the same symbol.




On the second chart, the phase indicator was reloaded, and note the differing values. Not just the phase indicator is different either! Also note the placement of the sines (and I didn't even reload those! It's impossible to trade with issues like this.

I really hope this is resolved soon. I have notable success with my MESA system (more so than any system I have worked with), and need to research this much more so that it can be optimized.. But I can't do it until this problem is fixed.

If Ehlers is planning on a complete MESA trading system for ESignal with 7.3 (which I am to understand, he is), someone needs to get their act together... no offense.

For what it's worth, this issue seems (to me) to be a threading problem in the DLL's. I'm not really sure though. These bugs are getting more and more awkward. (and more annoying, too).

Sincerely,
Joshua C. Bergeron
jcb@othernet / #ESignal

[ScottJ]

I think the current problem is anytime you use the same indicator twice, whether it be on the same symbol or a different symbol. We're working with Mesa engineers to find a solution.

Thanks for your patience.

[jbergeron]

I just downloaded 7.3; before I bang my head trying to figure this out, have the MESA issues been resolved in this release?

Optimistically,
Joshua C. Bergeron

[AndyS]

Hi Joshau,

Mesa continues to work on this problem. There is no fix within the 7.3 release. However, when there is a fix, we'll provide it here.

Regards,
Andy

[jbergeron]

Thanks for the information. I'll be anxiously awaiting (to say the least)

Cheers,
Joshua C. Bergeron

[jbergeron]

It has been almost a month now.. any new information?

With the release of a complete MESA based trading system in 7.3, I would think this could be more of a priority, especially when people are paying to use these indicators that have been disfunctinal since what I can tell to be their initial inclusion in ESignal.

When are they going to be usable? Does the MESA crew need some help with code? If there is anything I can do to help, please let me know.

I, like many others, have been patiently (well, almost patiently) awaiting these bugfixes for months and months now. With all due respect, this is getting kind of rediculuous

Cheers,
Joshua C. Bergeron
jcb@othernet / #es-traders

[ScottJ]

From what I understand, the problem with using multiple MESA indicators in the same layout has been addressed in version 7.4. If you'd like to test the beta, you can get beta 2 here. Please let us know how it works. We should be going gold with 7.4 by early July in case you don't want to test the beta.

Thanks.

[jbergeron]

Re: Reply to post 'More Mesa Bugs!'

Great, thanks. I will post my results in the thread.

Cheers,
Joshua C. Bergeron

----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Wednesday, June 11, 2003 5:14 PM
Subject: Reply to post 'More Mesa Bugs!'


> Hello jbergeron,
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>

[jbergeron]

I've been testing the MESA indicators under build 595. For the most part, they seem to be working correctly. However, I have noticed issues with the MESA Mode indicator.

This is a chart of a few mesa indicators, and also my system which uses them during the day of 2003-06-25 (today).



Here's that chart with the Mode indicator removed and reloaded onto the chart (for some reason I can't simply reload the mesa indicators ; they have to be removed and re applied. perhaps that's something else that needs to be fixed



The differences between the images may be kind of hard to catch for some, so for example, take a look at the mode at approximately 13:00 in each screenshot. You'll notice the mode inside the first screenshot is '1' a bar or two longer than it is after the indicator is reloaded

While this is a very subtle bug, it makes a big difference in the system's performance, and I hope it's a quick fix.

As for all the other bugs, they all look to have been fixed. Thanks.

Cheers,
Joshua C. Bergeron
jcb@othernet / #es-traders