Trackable Travel Rates-Part 6, Baseline

[+shellbadger]

I have examined a specific segment of my trackables to establish a baseline rate of travel that should represent my entire collection of trackables…see the large figure and appended table below. The methods used to assemble the data are detailed in the previous five parts to this post. 

 

As before, the bottom row in the table is the n, or the number of trackables in the series that achieved each successive drop. The sample sizes decline from 3449 trackables at Drop 1 to 154 at Drop 30. The decline results from high attrition between early drops, losses of 11-16% while trackables are in the US, to 4-9% when the oldest-surviving trackables are outside the US.

 

The solid-colored lines in the graph represent the top three rows of the table. They are the maximum (blue line) and minimum (gray) number of days required for a single trackable to achieve a specified drop.  The orange line is the average days (the baseline) for all trackables to achieve a specific drop. The dotted line is the trend line for the average.

 

The baseline is derived from the 24 series of trackables shown in the baseline contributors table below. The series are sorted into potential groupings that will likely be the basis of future comparisons. In those comparisons, the maximum and minimum lines will be omitted. They are retained here to illustrate the huge range of days for trackables to achieve specific drops. For example, for Drop 1, the min/max range is 0 to 3902 days, meaning there was one trackable released, retrieved and dropped again on the same day, whereas another trackable took 3902 days (10.7 years) to achieve the first drop.  At 30 drops, the min/max values are 580 and 3966, respectively 1.6 and 10.9 years. The average to 30 drops is 1552 days (4.3 years).  For convenience, I have also provided a days-to-years conversion table for each drop (see also below).

 

The data at every drop are badly skewed because zero is the absolute limit to the minimum number of available days between drops, whereas there is no limit to the maximum.  For the readers with a statistical bent, the standard (or average) deviation from the mean for each drop ranges from equal to, or more than twice that of the measured mean (average)…for normal (bell-shaped) distribution of values, we like to have values less than five percent of the mean, as opposed to the 100-200 percent seen here. Regrettably, the methods to attach any statistical significance to observed differences for skewed data have receded into the mental fog. Still, I can entertain myself by calculating averages and determining trends.

 

The average rate of travel over 30 drops is 51.7 days (1552 ÷ 30), but that doesn’t tell the complete story. The rate of travel during the first 15 drops (when more than half of the trackables are in the US) is 67.2 days per drop (1008 ÷ 15), whereas the rate for drops 16 through 30 (when more than half of the surviving trackables are outside the US, mostly Europe) is 36.3 days per drop ([1552 – 1008] ÷ 15). While I do believe trackables move more frequently in Europe than in the US, these values do not constitute proof, they are merely suggestive.  The reason being there are unquantified fractions from each region represented in early and late drops. That said, if I ever decide to winnow and compare US-only and Europe-only cohorts, I have every confidence that the difference will be even greater than reported here.

 

That there are differences in the early and late rates of travel can be seen by comparing the baseline with the trend line.  The trend line is straight while the baseline is a gentle arc, reflecting how the rate of travel decreases with subsequent drops.

 

This the concluding part of this post. This project will continue later with another multi-part post comparing series of trackables with each other, and to the baseline.