Bitcoin Zero Crossings

For any time series you can compute can compute moving averages. These averages are typically distinguished by the number of samples of the underlying data and the weighting (if any) that is given to each sample. Weighting typically involves weighting each sample in proportion to how close in time it is to the latest sample.

We can see this in action for Bitcoin on the cognotrend home page.

The moving averages always lag, of course, behind the current daily data sample. When the daily data is trending upward the averages will be below the daily value. If the daily data is trending downward the averages will be above. Eventually, the averages and the daily averages will cross when the trend swings from up to down or down to up.

Sometimes these crossing are termed zero crossings for the following reason: If you subtract the daily data value from the itself, you get a null time series, and if you subtract the average from the daily value, the result will be positive or negative depending on whether the relative trend is upward or downward. The swing from positive to negative or negative to positive is called a zero crossing (or null crossing).

You can also subtract the different kinds of averages from each other and examine those zero crossings. All of these zero crossings tell you something about the underlying trends. Shorter term averages will tend to make zero crossings with the underlying data series more frequently than longer term averages.

Today's BTC charts are exhibiting some zero crossings with respect to the 30-day EWMA (exponentially weighted moving average) that seem to indicate a cooling off of the recent upward trend and possible period of sideways movement or possibly even the beginning of a downward trend. Compared to the 30-day EWMA, the BTC price movement to the upside is running out of steam.