Another Golden Cheetah ? R. E. Histograms

The resolution of each histogram data point in watts.

IOW, a histogram or PDF counts up the number of data points that fit within each bin and plots that number against each bin (or it normalizes the data and displays it as a percentage of the total). So for power data the bins could be as narrow as 1 watt so count up all the data samples in your power file that came in at zero watts and then count up all the times exactly 1 watt showed up in the file then 2 watts, 3 watts, etc. The resulting histogram would have very narrow bins and be real spiky but would accurately show the frequency which which each discrete value occurred in the original data file.

More realistic bin widths for power analysis or on the order of 10 to 15 watts per bin, sometimes it makes sense to use narrower bins in the 5 to 7 watt range when looking for a distinct step down and sometimes wider bins in the 25 to 40 watt range are useful as a quick check of time in say sweet spot range or other range that doesn't fit nicely into more formal definitions.

Make the bins wider and you lose some resolution on the data and get some implicit smoothing of the data, make them narrower and you can sometimes pinpoint transition regions (like using histograms from hard rides as a first order guess at FTP via drop off) but the data gets more erratic with more peaks and valleys and typically fewer smooth trends.

Hope that makes sense.

-Dave