EFS Drives CPU to 100%

Carlton,

Just poppin' in, I assume you are really referring to Garth, eh?

[gspiker]

Hi Carlton,

I suspect you mean me, and not Steve, you shouldn't insult poor Steve like that, he never did anything to you ;-)

OK..1st. You shouldn't get the same results between the original and the one you modified to add "sym" to, unless sym and the chart symbol are the same. If you do, they it shows how badly the script was broken in the original. Why - because in the loop where they access volume information you never added the sym paramter. So you were using the underlying chart volume instead.

2nd - Because I use the user supplied sym for all paramters you should always see different results from you modifed script...unless sym and the underlying symbol are the same.

3rd - The original script has some problems with the way is used data such that it could (and usually would) generate NaN's during some math opperations (NaN = Not a Number). This means that some of the results would be off. Since your modfied script has these same errors it would cause the same errors. If your script run on say IBM, generated the same numbers as the original run on IBM - that implies that the bugs were bad enough that even though sym is set to $INDU and looking at $INDU for values...it got the same result as IBM. This is not a good thing.

The way I solved the NaN problem is admittedly a bit of a kludge. If I understood the formula well enough there would be a more graceful way of handling it (But since this really should only effect it during the initial few bars - it shoudn't be a problem overall.

The best bet is if you could get a snapshot of a verified DMI run on the same symbol and interval run on an other platform and compare the results to the one this script provides. It should be telling as to the impact of the changes I made.

If you want, I could change it back and reintrduce the bugs, but I doubt that would be a good thing.

Garth

[cpatte7372]

Hee hee, I get little confused sometimes - and its getting a little late here in England ;-)

I am going to run with your modified script. It was so frustrating trying to trade today - it just kept on crashing.

Anyway, you guys are great.

Thanks a lot.

Cheers

Carlton

[gspiker]

Carlton,

Maybe someone here has another platform with the DMI indicator available for it. If so, I bet they would post a snapshot of it on a symbol/Interval you use (it looks like you use $INDU...and this a symbol everyone should have on any trading platform).

I would encourage you to verify the results. It is better to use something that can be proved to be correct, rather than use something that may be wrong...

Garth

[cpatte7372]

Garth,

I have re-read your previous post, and it would appear that I didn't explain myself - I think you guys are mind readers :-)

Anyway, I did run the script on the same symbol (coincidentally it was IBM) with your very first version of demand index and the one you recently modified. I did change the symbol from $INDU to IBM hoping that I would get the same result.

So, where you said "If your script run on say IBM, generated the same numbers as the original run on IBM - that implies that the bugs were bad enough that even though sym is set to $INDU and looking at $INDU for values...it got the same result as IBM." I agree but I did change symbol to IBM.

Now, with that in mind would you say that there is a discrepancy?


Cheers

Carlton


Hopefully, someone has read the post from this forum with another platform with the DMI indicator.

[cpatte7372]

Hi Garth,

Following up from my last post, I was just wondering if you could think of any other novel ideas I could use in order to verify the results.

I couldn't concur with you more when you say "It is better to use something that can be proved to be correct, rather than use something that may be wrong..."

Cheers

Carlton

[gspiker]

Carlton,

Calculate the results youself for a few bars and see if they match.

Liberal use of debug (or Steve's trace) statements.

Garth

[cpatte7372]

Hey Garth,

I just want to take the time, while the markets are relatively quiet, to say thank you very much. The indicator works like a dream.

Cheers mate.

Carlton.

Steve, if you're check this post thank you too.

[its1111]

Hi Carlton
Is the working EFS the same one as at the start of this thread or have you tweaked it. I am always interested in new EFS to try. Can you post the finished product . We all want to see what works like a dream.
Mark

[cpatte7372]

Sure thing mate...


Here you go.

Carlton

[cpatte7372]

Garth,

There is just small hiccup with the efs - I'm having to manually reload it to get the latest value. This generally, isn't a problem after market hours, as I don't mind reloading it, however, I'm using the efs to assist me with my entry during trading hours so it be great if it would update/reload normally.

I remember someone else suggesting that the original Demand Index didn't reload and you fix it, I wonder if you could do the same with this latest demand index?

Thanks again mate.

Carlton

[cpatte7372]

Garth,

My mistake i do believe it does update.

Cheers

Carlton

[its1111]

Thank you Carlton for posting the EFS
Mark