There are many different reasons to perform a testing simulation. The following types of simulations are the most common; we’ll provide some examples of these in this chapter.
Testing to measure the range of accuracy for a specific filter
Testing to measure the speed of adaptation in chaotic environments
Testing to compare the effectiveness of multiple filters
Testing to compare features in a single filter
Not all simulations require the strict quality assurance discussed in the previous section. Tests in which the results themselves are not as interesting as how the results change between tests generally permit a more lenient test simulation (and sometimes a chaotic one). However, these types of tests are usually effective only at measuring features in a homogeneous environment (that is, where the same filter is being tested against itself). Since all filters react differently to the data, the only true way to compare filters or to measure the real-world accuracy of a filter is to use data that takes all test challenges into consideration.