are there any 'common' means to draw an automatic trendline and trend channels?
trendlines and trend channels such as these
https://school.stockcharts.com/doku.php?id=chart_analysis:trend_lines
i found an old thread from a web search, but i refrain from posting in that thread as that is some 9 years old
https://www.elitetrader.com/et/threads/mathematically-computing-a-trendline.224044/
it is easy to visualize on charts the highs and lows and draw trendlines/trend channels on them connecting the edges of the local highs and lows. but the reality is that the trendlines do not necessarily touch the edges even if they seem so. statistically they are likely off by some points. sometimes we subjectively ignore an edge or draw it without touching the edge but close. but on a chart it still looks like a trendline intuitively.
i've thought about linear regression (that's probably a start), but that linear regression mostly draws an average slope that cuts right in the middle through the series of points rather than the edges. and that the slope of the linear regression does not necessarily coincide with its edges
trendlines and trend channels such as these
https://school.stockcharts.com/doku.php?id=chart_analysis:trend_lines
i found an old thread from a web search, but i refrain from posting in that thread as that is some 9 years old
https://www.elitetrader.com/et/threads/mathematically-computing-a-trendline.224044/
it is easy to visualize on charts the highs and lows and draw trendlines/trend channels on them connecting the edges of the local highs and lows. but the reality is that the trendlines do not necessarily touch the edges even if they seem so. statistically they are likely off by some points. sometimes we subjectively ignore an edge or draw it without touching the edge but close. but on a chart it still looks like a trendline intuitively.
i've thought about linear regression (that's probably a start), but that linear regression mostly draws an average slope that cuts right in the middle through the series of points rather than the edges. and that the slope of the linear regression does not necessarily coincide with its edges