Advanced Get Efs Studies

[mickson]

I hope this question hasn't been asked a 1000 times before.

I would like to get hold of EFS files for the type 1 & 2 trading strategies od Advanced Get.

I am not looking for the Scanner as I would like to back test the strategy on the markets I follow.

Any assistance will be greatly appreciated.

Regards

Mike

[MR]

Hi Michael,

To be honest I am not an EFS guy. Questions I tend to ask, for someone looking for EFS parameters or ideas for at least a Type 1 setup:

1. Is the 5/35, 10/70, or other oscillators to zero and not below 1.40?
2. Are the oscillators acting in unison in the setup?
3. PTI > 35?
4. W4 labeled or looking for a W4?
5. Any Ellipse picking up the time and price? How far away is the closest Ellipse so far?
6. Any key MOB supporting/resisting the setup?
7. Any 6/4 DMA or RTC confirming or setting up to confirm a W4 breakout coming?
8. Is the risk/reward 1.6% or better under the current W4 setup?
9. Any other time frames confirming the setup? Have you first "cross-referenced" the higher time
frames, applying basic tools for identifying support and resistance areas, then reissued that chart
to the lower, trading time frame for a closer look?
10. Has the W4 corrected enough in relationship to the Fibonacci extension of W3?
11. Is the wave sequence conforming to the "rule of alternation" wave pattern?
12. Are there any other tools or studies supporting/resisting the current W4 setup?
13. Are W1 and W4 close to overlapping, if not, how far away are they?
14. Is the current W4 in progress close to any kind of identifiable, obvious major previous pivot to
support/resist current action?
15. Are there any other similar stocks/futures/indices/sectors/ that are in similar patterns at the
time to help confirm this individual setup.
16. Is the stochastic, or other studies showing any traditional buy or sell signals too?
These are the general questions I would ask. If I can answer yes to many of these questions then I
would say, in general, the odds have increased in my favor we are in a good W4 setup

Type 2's are a little tricker:

Elliott Oscillator confirms a Fifth Wave with clear divergence.
The Elliott Oscillator also pulled back to zero in between the peaks
Crossover of 6/4 DMA or Regression Trend Channel confirms.
Maybe build an XTL into it?

[gspiker]

Let me add a bit to Marc's excellent reply. I am an EFS guy, and while it would be very possible to code **MOST** of the generic rules for type I and type II trades there are some limitations.

1) You don't have access to all of the data stated in the generic rules...though you do have access to most

2) You say you want to backtest. It impossible to backtest AGET the way the backtester currently works. I have pointed this out to eSignal (and in fact pointed out that AGET isn't the only strategy that can't be coded for backtesting). I'm not sure if or when they are going to fix the limitations of the backtester...

3) To get really good results, you need to code in some of the added rules that are not defined in the type I and type II trades, but that most traders do use. Here you have two problems, because it is a more subjective which techniques they use for early entry, and many of them involve line studies which you can't (at this point, but this should change) get data from to use in your EFS.

That said, I know there were a few people looking at coding this...maybe they can share whatever luck they had.

G

[mickson]

Let me just thank you both so far for your wonderful contributions. I am always amazed at the speed of the responses.

If I can add that I would be happy to pay for someone who has the capability of designing the appropriate EFS so that it can be backtested on the ESignal software.

Will be looking out for the answer to my question.

Mike

[eSignal_JasonK]

Hello Mike,

Garth is correct. Developing a back testing formula using the AGET studies could be very complex and is beyond what we can provide as a free service here. You may contact our EFS Consultants directly to discuss the possibility. I would check with Divergence first. I think they have done some work on this already, but I'm not positive.

[Natalie]

I can understand the elliot wave things being very complex to test, but I would have thought that XTL is something that could easily be combined into backtesting. lets face it, it's nothing more than getting the setting and using a condition something like if barcolor = blue...

It would be nice to know how to access XTL at least (within an EFS) for this purpose. Any information greatfully recieved.

Many thanks

Natalie

[gspiker]

Hi Natalie,

You are correct. I don't think that doing an EFS that did backtesting for XTL is out of the question, but it would take a bit of time to put it together. It doesn't use any of the features that, at least for now, limit backtesting AGET Type I and Type II trades.

Maybe someone with more interest in XTL than I currently have wouldn't mind taking a poke at it. It should be workable.

Garth

[eSignal_JasonK]

Hello Natalie,

Here's a code snippet that will show you how to access the return values of the GetXTLStudy().

function preMain() {
    setStudyTitle(" Get XTL ");
    setCursorLabelName("XTL", 0);
}

var study = new GetXTLStudy();

function main() {
    var vXTL = study.getValue(GetXTLStudy.XTL);
    if (vXTL == null) return;
    
    // return values
    // 0 = Red Bars
    // 1 = Blue Bars
    // 2 = Black Bars
    
    return vXTL;
} 

[Natalie]

Many thanks

best

Natalie

[vijay]

Hi

as of this date I am curious if there has been any improvement to the backtester in esignal for backtesting Type 1 and Type 2 trades? I too am interested in going back a couple of years on daily data and backtesting some of these trades.

If the answer is no, can it be done with some custom coding (bpth writing the EFS scripts for Type 1 and Type 2 trades as well as the backtesting)? I am also looking to pull in options data for the actual trade. Is this possible?

Many thanks.

[gspiker]

Vijay,

There are a few issues you raise. I assume here you are talking type-I and type-II.

The problem isn't really with the backtester - it is with the way the aget inetracts with the backtester (and the user for that matter).

But to answer the questions. As long as you are using daily data, they you do have access to years worth of data.

The backtester and aget still don't work togther well for backtesting purposes.

You could try to do this on your own, but you will run into these problems:

1) The PTI value will be wrong most of the time. It will only reflect the latest value.

2) You will have to code in such a way that when aget relabels your code does the right thing.

3) You will have problems if you want to do more advanced analysis for your backtesting (ie: Multiple timeframes for verification of signals).

Garth

[vijay]

Garth,

Thanks so much for your reply. Yes, I am referring to type-I and type-II trades in Aget. I appreciate the points you raise about wave counts changing over time. Hopefully, my entry and exit rules will take care of this problem (meaning I dont really care what the exact wave count is as long as the trend is still in my favor or I haven't been stopped out as a result of a change in trend).

As for the PTI being the most recent value, I am not sure I understand exactly what you mean. As long as the PTI is a specific value at a discrete point in time when I enter the trade (for example > 35 for a Type I), then that's all I care about.

Yes, I see your point about multiple time frames. This is kinda important if I wish to do effective backtesting. Is it completely impossible to corroborate with different time frames or is this doable with some extensive coding?

Thanks again!

[gspiker]

Vijay,

re PTI. I was afraid I wasn't being clear enough on that. Let me try to explain what I mean. But one caveat on this first...I have requested they change this and eSignal said they would look into it. I haven't heard from them that it has been changed, and therefore am assuming (haven't tested it) that it still works as it did 2-3 versions ago. OK here is how it used to work (and I would guess still does): If on todays date the PTI is 25, then whenever I try to get the PTI on the chart, I will always get a PTI value of 25. Notice that with Aget the PTI value, if it is even up, will only reflect the latest PTI...there is no historical PTI values. This will be true with both the backtester, and if you tried to implement your own backtesting. One thing you MAY be able to do (I haven't tried it, but a quick test would be easy) is to run AGET in tick playback mode and see if PTI is reported correctly. In which case you can write a formula that will backtest correctly, but only in tick playback mode.

Wow...that was long. Sorry I couldn't be more concise.


re multi-timeframes. This is possible. You would have to write code to run on the higher interval that would put values into global variable (or to a disk file). You would then read these from the shorter interval chart. This is obviously a pain, but doable. The code will not be pretty either.

For this second point, help is on the way. It may be a few releases away, but easy multi-timeframe support is being worked on.

Garth

[vijay]

Garth,

Thanks again! Very helpful.

On PTI, I have often backtested using playback mode and just assumed that the old PTI values were always available. So, as you suggest, perhaps the backtesting needs to be done in playback mode. Interesting...

Thanks also for the feedback on multi-timeframe analysis. Hope esignal gets it done sooner rather than later...

[pj909]

Mike, Garth anyone,
Has the Type 1 / Type 2 efs been done yet? I don't need the backtesting just something that will put a visible signal (like a text box) on the chart if the rules are confirmed.