Kai Klaas wrote:
-------------------------------
The only way to reduce the uncertainty is to increase the amount of data points, while repeating your experiment again and again, and, of course, by having this experiment repeated by many colleguae teams all over the world.



very right,

infact stastically we assume the highest probable value is MEAN (gaussian peak). and distributing other vales both side of the mean with exponentialy decreasing probability.

kai i did an experiment during my UG degree the assignment was to simulation and study of event of tossing a coin and verifing the stastical fact that deviation from the probable outcome decreases as number of experiments increase. i used the 'C' function rand() to return the random value in the interval [0,1). it was also assumed that random numbers are eqaly likly. toss function was like this the best i can remember


int toss(void)
{
return(rand()*2.0)
}

i.e. its head if the value is left of the peak and tail otherwise. statisticaly if this exp is done N number of time head is expected N/2 number of times


float expr(long N)
{
int i,head;
for(int i=0;i<N;i++)
{
if(toss()) head++;
}
return(head/N)
}

exp was done with increasing number of N(sufficiently large) and results were saying exactly what you did.



abhishek