Build-1-of-finantic-IndicatorSelection-now-available

Amazing!

P.S. Typo in average: "Cumulated Avgerage Target"
https://www.wealth-lab.com/images/extensions/finantic.IndicatorSelection/Contribution.PNG

I think the https://www.wealth-lab.com/extension/detail/finantic.IndicatorSelection#screenshots link needs much more discussion about what the plots are all about. As it stands now, I'm not sure what I'm looking at in many cases.

And the profit-target vs indicator-value scatter plot is unnerving. It shows an R-squared relationship that must be under 0.1. What kind of relationship is that? The model needs more terms (maybe even more dimensions) to boost the R-squared correlation to be useful. Part of the problem is that the initial indicator entry value may not correlate well with the "overall performance" of the strategy itself; these are two separate (and independent) entities.

The fundamental problem is the "entry" merits must be analyzed separately from the "exit" merits because they are independent. You can't be grouping the two together and have a meaningful correlation.

Eugene would you mind fixing DrKoch's typo?

QUOTE:
needs much more discussion about what the plots are all about.

Yes, of course. The screenshots are just appetizers...

The extension comes with a very verbose collection of Help Files:

There is:
* Introduction
* Getting Started
* Frequently Asked Questions
* Glossary

and much more pages that walk you through all the screens and have quite a bit of explanation.

QUOTE:
Eugene would you mind fixing DrKoch's typo?

DrKoch has to fix it since it's in his code, not just a description thing.

QUOTE:
And the profit-target vs indicator-value scatter plot is unnerving. It shows an R-squared relationship that must be under 0.1. What kind of relationship is that?

What you see here is a quite common (i.e. normal) relatoinship between an indicator (and this one is one of the best) and the profit of a trade starting the next day.

Most people never look at this relationship when plotted as a scatter plot.

In fact there is a very low correlation. That is why I always say: financial time series are mostly noise.

QUOTE:
The screenshots are just appetizers...
The extension comes with a very verbose collection of Help Files:

Okay, I didn't know. You want the screenshots to "sell" the extension. So to make the screenshots more enticing, I would add just enough discussion to get the viewer to dig in deeper.

QUOTE:
[Concerning the profit-target vs indicator-value scatter plot] What you see ... is the relationship between an indicator ... and the profit of a trade starting the next day.... In fact there is a very low correlation.

I didn't realize this was the "next day" profit following the indicator value. I thought it was the overall profit of the trade after the position was sold. I thought each plotted "dot" represented a completed position. This should be clarified.

But I would never include a plot with an R-squared below 0.4 under any circumstances. I appreciate you're trying to say there's no correlation between performance from one day to the next, but we already knew that.

QUOTE:
Typo in average: "Cumulated Avgerage Target"

Will be fixed in build 2 of finantic.IndicatorSelection.

Thanks for pointing this out!

Looks like another well thought out tool! Thanks @DrKoch

Hello! I just installed IndicatorSelection and noticed that when I click "Getting Started" in the Help, it says:

QUOTE:
Empty Topic: finantic.Eval.GettingStarted

I'm trying to get started as best I can without the "GettingStarted" Help section. I haven't modified any settings or config other than making "DataSet and Scale" and "Intervals" match my backtest settings. I'm getting lots of errors in the log after clicking "Start calculation":