Slope and Linear Regression

[DaveSkender]

Community discussion about Slope and Linear Regression

[DaveSkender]

From @ray2k:

Any chance of getting a standard dev channel indicator? It's basically a best-fit linear regression and (typically 2, sometimes 1) standard deviations up/down from the line where (in the cast of 2 deviations) something like 95% of price action tends to take place. I think this can be pieced together now between the 'Slope and Linear Regression' and 'Standard Deviation' stuff but seems less than ideal.

[DaveSkender]

Have you seen this somewhere I can reference? We usually don't do custom indicators, but I think this is doable as an extension of Slope, much like we do for GetSma and GetSmaExtended.

[DaveSkender]

Were you thinking something like this with standard deviation for the reference window:

Or like this, using the rolling standard deviation:

[ray2k]

I'm most familiar with the StandardDevChannel as it exists in thinkorswim:
https://tlc.thinkorswim.com/center/reference/Tech-Indicators/studies-library/R-S/StandardDevChannel
also:
https://www.incrediblecharts.com/indicators/standard_deviation_channels.php
https://www.interactivebrokers.com/en/software/tws/usersguidebook/technicalanalytics/standardDeviation.htm

As far as your question, more like the first one. In thinkorswim it defaults to basing the trendline over the entire chart which is how I find it useful (then can switch chart timeperiod for different views/etc). Don't know enough if that is the standard/default or not.

[DaveSkender]

Ok, I see. Since this is used in reputable broker sites, I think this would be a good indicator to add. Please go ahead and create an Issue and I'll put it in the queue.

[DaveSkender]

I think what we can do is make the lookbackPeriod optional; that way, we can do "entire chart" by default. See #361 for development notes and status.

UPDATE: This indicator has been implemented as Standard Deviation Channels

[Snappee]

There are situations where the function does not compute the LINE value in the result but returns everything else. I have a special use case where I first compute a momentum value, then pass it through linear regression to calculate the line, slope, etc.. I have not had time to look into your core code to figure out why a NULL exists, so currently I check for a NULL value in the line and do the standard "line = M*X + B" calc.

[DaveSkender]

It only computes the line over the last N periods for quotes.GetSlope(N), so the Line value will be null for all periods in your results before the last N bars.

[DaveSkender]

Add the Linear Regression option here to test it out: https://charts.stockindicators.dev/