Requests & Ideas - page 188

 

Gents,

Hope you all are enjoying your weekend!

Quick question: i really like the divergence capabilities of some of the indicators being developed here. Was wondering if there have been attempts to code an EA which detects (or 'reads') the divergence from the indicator with recent price.

If so, would love to see it as an example. Can/has this been done?

I am well aware that divergence can only be established several bars "after the fact"...and that's OK!

Cheers, San.

 

San

Divergence needs to detect peaks which means that it is a bit similar to a zigzag : it always can be detected only after the fact. So, if an EA is concerned, working with "after the fact" indicators is out of the reach of regular EAs (it would need to "know" what we human know to be able to use divergence in a right way)

Snowski:
Gents,

Hope you all are enjoying your weekend!

Quick question: i really like the divergence capabilities of some of the indicators being developed here. Was wondering if there have been attempts to code an EA which detects (or 'reads') the divergence from the indicator with recent price.

If so, would love to see it as an example. Can/has this been done?

I am well aware that divergence can only be established several bars "after the fact"...and that's OK!

Cheers, San.

 
mladen:
David

The "cloud" can not be converted in the form it is (metatrader 4 does not have a fill option). Anyway, it is a regular QQE just presented differently, so if you use any normal QQE the results are going to be the same

As of ccitema - you already have it in elite section (cci with pre filtered price using 18 types of averages - just chose tema and you will get that result) CCI tema trend for some reason does not work OK on my since some of the new upgrades PC but judging from the number of drawing buffers it uses, it can not be converted to metatrader 4 either

thanks for your kindly and rapid reply.

 

Hi guys,

Implemented an adaptive lookback indicator (alb rsi smoothed) in an EA.

However, when running it on a backtest, the EA seems to load the indicator each tick...and thus slowly but seriously slows down the computer.

See screenshot (I renamed the indicator for practical purposes):

Is this due to the adaptive lookback or is there a solution to prevent the EA from doing this.

I've implemented other indicators and they all load once, all fine.

Thanks in advance for tips and suggestions!

San.

Files:
 

San

That is an almost sure sign that parameters in the iCustom() call are not correct

Check the parameters for iCustom()

Snowski:
Hi guys,

Implemented an adaptive lookback indicator (alb rsi smoothed) in an EA.

However, when running it on a backtest, the EA seems to load the indicator each tick...and thus slowly but seriously slows down the computer.

See screenshot (I renamed the indicator for practical purposes):

Is this due to the adaptive lookback or is there a solution to prevent the EA from doing this.

I've implemented other indicators and they all load once, all fine.

Thanks in advance for tips and suggestions!

San.
 

Doh...! Found it.

In the indicator, half-way through settings, there were some notes:

extern string infoa = " Mode settings ";

extern string info1 = " 1 = Period of Dominant Cycle";

extern string info2 = " 2 = Amplitude of Dominant Cycle";

extern string info3 = " 3 = Phase of Dominant Cycle";

I simply "skipped" them in my parameters call....but that did not work. Moved them down solved the problem.

Thanks for the tip mladen, much appreciated!

 

Hey lads,

I really like working with Absolute Strength indicator. Sometimes however, when not zoomed in, it is kinda annoying to really find out where lines broke out, or touched.

Would it be possible to draw the symbols I painted in the picture on the main chart?

I think the chosen symbols would be perfect to interpret.

Also dots, stop signs and arrows should have their own drawing height from the candles position, as they can occur on the same point of time.

Maybe the option to choose which symbols to draw (e.g. show stop signs = fault etc.) would add to the usability of this feature too.

I would be happy if someone is able to do that. Thank you Regards.

absolutestrength_v1.2_mtf.mq4

 

Here's an initial version.

As you see, the chart gets crowded quickly, so I decided to just do the 'entry' signals for now. You can turn these on/off in the settings, change color, etc.

Cheers, Snow

 

Market-Condition Range/Trend Detector Based On Einstein Power Law

Market Condition Range/Trend Indicator

Hi Mladen/Mrtools,

Having seen many papers evidence of markets asnon-random and with significant autocorrelation effects, i wish that we can codethis range/trend indicator. While Polychromatic momentum was good fordetecting divergences very reliably using Einstein's Square Root Rule,Einstein's power law could be used in a better format to detect range/trendconditions as i propose and believe below. If this is done, this will be the2nd publicly-available market condition indicator based on a power law (other than fractal dimension/hurst).

-------------------------

Idea:

1. Bill Dreiss's "Choppiness Index" orSD divided by ATR

2. Apply autocorrelation, ideally InverseProportionality with 4 Corrections (IPN4) which doesnt give significantlybiased estimates like the usual Autocorrelation OLS procedure. IPN4 is betteras it takes estimates in proportionality to sample sizes via 4 coefficientcorrections (instead of a discretionary manner like adding more self-determinedcoefficients to a low-pass filter for better smoothness). IPN4 is superior tothe common autocorrelation method. (if IPN4 is too difficult to code correctly,the basic autocorrelation function shall have to suffice). Formula as stated in the last few pages of the PDF file.

3. Apply Einstein's Particle Square Root Rule,akin to Polychromatic Momentum. Hence, final product has fixed levels between 0and 1 for us to use heuristically to determine range/trend conditions as usersees fit.

----------------------------------

Conclusion:

Final implementation mode is free for you to doas you deem fit as im not mathematically/scientifically/programming inclinedand can only provide a skeletal idea. If this idea has no merit/pretty crappy, feel free tosay so/ignore this. Regardless of decision to code or not, i would like to thankyou for both your presence at TSD as i've learned alot from you all.

Yours sincerely,

Wintersky

 

I like your idea Wintersky111. Read that paper ("Climate") already some time ago, very interesting.

Thanks for posting and let's hope mladen can make some edible soup out of all this.

Cheers, Snow.