I'm using daily charts for one of my strategies that has trades that last anywhere from a few days to 2 months. Although I respect the trend, I realized my signals have a higher win rate if I use them on a symbol that has a steady trend on a longer time scale like 1-3 years. While two symbols might ping all the standard trend indicators like MA, EMA, MACD, etc., their 1-3 year charts can look very different to the naked eye. One might be more like a U shape while the other is a steady diagonal line from the bottom left corner to the upper right corner. I'd like to have an objective way of analyzing this aspect of a symbol and the common trend indicators don't seem to do it and Average True Range doesn't really capture it either. To save me from re-inventing the wheel, are there any existing indicators that would help me with analyzing and screening for the steadiness of a trend? I want to say I can use correlation coefficient somehow but not too familiar....
Best indicators for the steadiness of trend
- Thread starter Radiohead
- Start date
%%
Steady trend?? NOT really \SPY benchmark in a bear down\trending market.
But i have used a 50 week moving average, on a 3 year chart= objective......+ price still above 50week ma.
200dma=objective, a bear measure, seldom see a smooth bear trend\so no way a trend would be smooth\ even if a 200day moving average line is smooth+ steady.
Linear Regression Moving Average
This indicator 'linearizes' price movement over any period specified. It's the most 'objective' way to look at a trend. This is due to the fact that each observation is given (an approximate) equal weight using an error minimization algorithm (least squares). You could use a 1h, 1d, or even longer period.
A very reasonable solution for your use case. It is statistical in nature.
This indicator 'linearizes' price movement over any period specified. It's the most 'objective' way to look at a trend. This is due to the fact that each observation is given (an approximate) equal weight using an error minimization algorithm (least squares). You could use a 1h, 1d, or even longer period.
A very reasonable solution for your use case. It is statistical in nature.
That's not really true , to calculate the Hurst exponent it's exponentially weighted moving averages over many timescales as ts a fractal thing
%%
THAT could show [as measured] on a chart as some what a smooth price around a 200dma; not that price in a bear market would be smooth LOL.
BUT a 200 day moving average is fairly steady +smooth \not a that price is, price seldom is smooth.
What, exactly, is not true. You know, for someone as seemingly math literate as you, you're not very good at communicating the literature to non mathematicians.
I'm typing on a goddamn phone I hate it
https://link.springer.com/chapter/10.1007/978-3-642-14350-2_17
It's not true using a single time period will work to estimate the Hurst exponent or stickiness of a series . It would get an answer but not be very rigorous, since there is no reason one timescale is more important than the other