Dear Mahmood,

looking at your nice graphic, I think it happened, what I feared to happen. There is a certain repeatability!

Some examples:
If '5' occured, there's a high probability, that again '5' occurs. Nearly every second '5' is followed by a '5', resulting in a 'twin-package'.

Also, there are lots of 'twins' and 'trios'. Too much than looking actually randomly. Especially the trios will annoy the people, which have to work with these random numbers. These many trios are just too unprobable to occur so often...

Could it be a solution just to suppress the twins and trios? Means, to generate a new random number, whenever a twin or trio occurs? I'm aware of the fact, that the total absence of repeatings is also unprobable, but may be, it's not that annoying like those many trios?

May be, what Bartosz reported can help. But keep in mind, that these random numbers are always fabricated by a certain routine. And by masking a serious loss of pseudo randomness can occur. When I experimented with modulo concept, I was very surprised how much simple masking or dividing can damage this randomness... And I was surprised, how many firstly good looking routines showed certain patterns after some analyzing. It's a very difficult thing to produce really randomly looking pseudo random numbers!

Good luck,
Kai