Signal clarity

[fabrizio]

Hi

Being honest I have followed your thtread since the very beg, making at the early stages some backtesting.
I must Share the Matt point, I'm a little confused.

I have a sub folder full of efs but actually I'm not sure which ones sssss are is the right to drop in chart.

Is it possible to group them in a zip file and post it? Just the necessary last releases ones....

Thanks. a lot

[FibbGann]

This system is a work in progress. So at this point there is no need to post the current filesbeing worked on. When we get near a finished product one will post and give the details.

Fibbgann

[mattgundersen]

You know what would be nice

EFS has a function you can put in preMain called "checkVersion".

What could be done is designate which EFS's are the ones to use. Then as the group agrees on what the latest/greatest/official changes are, post those to the fileshare.

Then as people run the EFS's from their hard disk, checkVersion looks to see if there has been an update and automatically grabs the latest EFS! from the fileshare.

I will make one disclaimer though: This is a function that has existed for some time but needs more cleanup. Alexis and I worked on it for a while cleaning up some items that caused confusion. It still needs more refining...

Anyone want to give it a go? I'll provide details and help clean up stuff as needed...

FibbGann - can you refresh me on the question? I forgot.. Sorry.



m.

[FibbGann]

Thanks Matt... Will look into it. About my question... I was wondering when Merlin (your avatar) was going to be linked to esignal? (ESP)


Fibbgann

[fabrizio]

Matt

If it is a matter of testing it hard I'm on your side; for script evaluation I'm too bad so I'm useless.

Ps. Actually as you remark, the sistem never worked properly since the very beginning. And would be great to have it tuned up.

[FibbGann]

Here are the new efs's for the MA's.(will upload them to fileshare) I don't have the new overall efs'.(Fgallfalse.efs) Have to clear some things up with Alex and David first. I did not incorporate the ability to quickly change the settings in edit studies for each efs. (maybe some can do that or fill me in on how?) Have'nt made it that far through my books yet. LOL

I like what I see, I guess it all depends on if someone is looking for an automated trading system or a helper?

Anyways, I will get these posted to the file share after a few more tweaks.
What would really be nice is to be able to average all 3 together and use signals based on the degree of angle. (Im a dreamer!)

P.S. I need to run the RSI through the ringer, but will wait til the final efs. is finished.

Fibbgann

[dloomis]

Re: Reply to post 'Signal clarity' - rise/run question

Now that you brought up "degree of angle"...

This never made any sense to me - I always thought slope was related to the
rise/run. When I think about this re ma's, I get confused.

To me, rise is the change from one bar to the next - or group of bars to the
next group of bars - and run is the number of bars the rise is measured
over.

So if the ma increase by 5 points in 15 bars, the slope is 5/15 = 0.333.
Then this value of 0.333 is compared to another time period when the rise is
3 and the run is 15 and the slope is 3/15 or 0.2. Then I figure you want to
compare slopes, so you divide 0.333 by 0.2 to get 1.6 - meaning the original
rise was 1.6 times the latter rise - which is exactly the same as dividing
the original run - 5 - by the latter rise - 3 - for 1.6 again.

So how does the slope tell anything.?

Sorry for the diversion while we sort out which efs is which.

[FibbGann]

Now thats a nice mind boggler for a friday night. As you know by now my mind never stops. I have so many senarios that run through my head about how to do things. What I was thinking out loud about was, I was wondering if some how I could take the 3 MA's (not to get off the main purpose here, just for testing purposes) and weight them together (some type of average directional movement) into one MA. and then get signals based on the degree of angle the MA's is pointed.

Note: I am just running of ideas. It does'nt necessarily have to pertain to this project.

Fibbgann

[dloomis]

Re: Reply to post 'Signal clarity'

If I wanted to average the three indicators, I would average the rate of
change of the three indicators - bar to bar.

Then I would export the rate of change average figure into excel and sort
then to find what is a rapid rate of change - in both directions.

Maybe take the highest 10% of signals and see if those quick rates of change
told you anything.

The idea is that when the market is changing direction rapidly, trading
opportunities arise.

[FibbGann]

The 3 efs's for using the Ma's of RSI, CCI, and STo have been uploaded to file share here


Fibbgann

[FibbGann]

Now that I have a clue (efs building) I have set-up 3 different scenarios in testing this strategy. Now that I can fly through building these formulas I am going to run a bunch of test on different set-ups.
Will let you all know my results and post them in file share. Dloomis I will run tests on Alex's latest strategy and the one I was talking about.

Thanks all,

Fibbgann

[ACM]

All
Matt makes reference to a checkVersion function available in EFS to be used for automatic updating.
If interested in seeing how this function works click here and download acmupdatedemo.efs (saving it to your Formulas folder) and acmupdatedemo.doc which are the instructions.
Once you follow the instructions you should be able to apply the automatic update to the acmupdatedemo.efs you have just downloaded. FWIW the updated efs is located here under the same file name.
Alex

[ACM]

David
I see that you added all the strategy logic to FGallfalse3. Here is the same efs with a correction to an error that was in my original framework (changed STOMALength to Slow in vMAofStoch).
I also added the checkVersion function and have set the internal version number to 1. Further revisions should be internally numbered 2, 3 etc while leaving the file name unchanged for the EFS Update function to work.
Still needing to be added are the OB/OS user adjustable levels should anyone want to try the original OB/OS strategy using MAs. I will take care of that and post the updated efs.
Alex

[ACM]

FibbGann

Alex? when you coded the Fgallfalse(2b).efs. Were you intending to use the median of each indicator for the B/S signals.

Yes. The strategy in fgallfalse2b.efs (and in FGallfalse3.efs) does the following
IF CCI crossing over 0 AND RSI crossing over 50 AND Stoch crossing over 50 THEN Sell
IF CCI crossing under 0 AND RSI crossing under 50 AND Stoch crossing under 50 THEN Buy

Alex

[ACM]

Matt (or anyone else tuning in at this point)
A brief recap of what has been done up to now (no comments or considerations are made here)

FGallfalse2.efs - This is the original strategy that Buys when all three indicators (CCI, RSI and Stoch) are crossing over the OS level and Sells when all three indicators are crossing under the OB level.

FGallfalse2b.efs - This is a revised version of the above that Buys when all three indicators are crossing under their corresponding centerlines and Sells when crossing over the corresponding centerlines. So far the best parameters I have found are

FGallfalse3.efs - This is a revised version of the above but uses moving averages of the indicators to generate the signals.

All three EFS(s) are available in the Three Signals group and mirrored in the ACM Test group.
All the EFS(s) now include checkVersion set to 1. Further revsions of each should be internally numbered 2, 3, etc.

Alex