A consideration about picking up the `best' values from N times experiences. Mathematically, the author considered "how the maximum value of N values that are distributed normally distributes?". The median values are 0.998, 1.498, 1.998 for N=4,10,30, respectively.
This proposition would be useful because everybody quite often is in need of choosing something best from multiple candidates in their daily or social life. The figures shown above are easy to remember so you can utilize them with ease.
2. Statement :
Assume Norm + ( m,s 2 ) be the distribution of the maximum value of
N
N stochastic variables distributed normally with mean m and variance s 2 .
Then for N = 4, 10, 30, each median value of Norm + (0,1) are,
N
approximately 0.9981, 1.4988, 1.9982, that are rounded down to
0.998, 1.498, 1.998, respectively. These are easy to remember as
1-e , 1.5-e , 2-e for N = 4, 10, 30 with e =0.002, thus possibly very useful
for your daily and social life. These numeric figures are calculated from
billions of Mersenne Twister psudo-ranom numbers using R language.