A simple bit of statistics is needed here. If you have a statistic sample of 100 your chance of error is relativly high. If you run the same code for 1000000 your deviation from the expected 10% will be proportinatly less (probablt between 9.9 and 10.1%)

PS. I think your calculation is wrong as Random * 100 will give a number between 0 and 99 not 1 and 100.