For test #0, there were a total of 125 attempts, of which 7 were successes and 118 where failures. This is a 5.60% success rate. This test was (as guessed by Monder) a control against random selection of points.
Test #2: 155 total attempts, 151 successes and 4 failures. That is a 2.58% failure rate. Since the isometric map makes the height of the hills very apparent, the user should be able to determine the highest hill every time. This test is therefore a control against user error, as well as the unfortunate habit of the program to put the hill against the edge of the map (an easy fix, I only just noticed it today).
Test #1: Out of 150 attempts there were a total of 127 successes and 23 failures, an 86.67% success rate.
For anyone that explicitly said they had problems with the program, I removed their data point. This data is considered corrupted. If we consider people who got 0 or 1 successes as statistical outliers, then the success rate for test type 1 jumps to exactly 90%.