With 20 sets of simulation data, random numbers that fill the chip dimension,
the standard deviations for means and medians of the moving windows were calculated.
The standard deviations present the ideal value,
which will be obtained when noises are absent in microarray analyses.
| cycles | mean | median |
| 1 | 0.1984723 | 0.2479306 |
| 2 | 0.1989189 | 0.2466279 |
| 3 | 0.1986852 | 0.247983 |
| 4 | 0.2024613 | 0.2502793 |
| 5 | 0.1970177 | 0.2450545 |
| 6 | 0.1998656 | 0.2469877 |
| 7 | 0.2008643 | 0.2497612 |
| 8 | 0.200276 | 0.2476841 |
| 9 | 0.1995815 | 0.2485011 |
| 10 | 0.1995136 | 0.249285 |
| 11 | 0.2006275 | 0.2493826 |
| 12 | 0.2013696 | 0.2494893 |
| 13 | 0.1996653 | 0.2482006 |
| 14 | 0.2013769 | 0.2502287 |
| 15 | 0.2011599 | 0.2502651 |
| 16 | 0.1993147 | 0.2472288 |
| 17 | 0.1998732 | 0.2499465 |
| 18 | 0.1991683 | 0.2472938 |
| 19 | 0.1976704 | 0.2472706 |
| 20 | 0.2000082 | 0.2493452 |
(average for the median) = 0.2486063
The R program for this simulation