After the Clinton administration dropped SA (Selective Availability), on 01 May 2000, a marked improvement in positional accuracy was noticed. The GPS WAYPOINT REGISTER took the opportunity to measure the accuracy that can be expected when taking a GPS reading.

We marked a specific spot and took GPS readings of the location's coordinates between 01 June 2000 and 03 October 2000. In total a number of 143 measurements were taken. We then averaged these readings in order to get a better estimate of the true GPS coordinates. After that we calculated the positional errors made with each reading and arrived at the following conclusions:

1.   "most" of the time, a single GPS reading will give you a reading within 8 to 20m from the actual point.

2.   90% of the time, you can expect the error to be larger than 8 m

3.   90% of the time you can expect the error to be less than 20 m

4.   The WORST reading had an error of  37 m

5.   The BEST reading had an error of 7 m

6.   Taking 3 readings over a period of 12 hours, and averaging them out, reduced the expected error to about 5m

7.   The EPE (Estimated Positional Error), as indicated on the GPS when taking readings, is not accurate, but changes in relation to the actual positional error.

The conclusions can be displayed graphically as follows:

The above results, do not correspond with the statement that dropping SA would results in an accuracy of better than 15m, 95% of the time. HOWEVER, we concluded that many other factors influence accuracy, other than the dropping of SA. These factors includes DOP (Dilution Of Precision), which contributes to the overall errors that we've encountered.