N_Faker wrote: ↑01 September 2020, 19:04
So there are statistical differences between pre-generated and real-time generated random numbers?
There could be yes. It depends on how a site is using the predetermined sets of random numbers.
If sets of numbers were used more than once then that would ultimately skew the stats.
If no sets were re-used but some sets were never used that too might sway the numbers.
But this is somewhat academic. As previously stated everything rests on how any given site uses the random numbers.
Consider this. A simple game of Backgammon. Instead of calling the RNG program at the point of each dice throw lets say the site instead calls 30 dice rolls for each player BEFORE the game starts. Totally randomly. You now have 2 sets of 30 dice throws, one for each player, each with genuine random numbers.
Now from a randomness perspective those dice rolls will always look right when analysed. They have to. They are pure random numbers.
However the fact that they were generated BEFORE the game makes a huge difference. Because of that the system can then choose whichever one of those dice throws out of the 30 that it wants at any time.
Suppose you have within your set of 30 rolls the numbers 6-6 1-1 4-4 and 27 others and your opponent has 2-1 3-2 4-1 and 27 others.
If I give you the 6-6 as your first throw you get the perfect first throw. I give the other player 2-1, an awful throw. I then give you the 4-4 allowing you to unload your outside stack of 5 and your other stack of 5. I give the other player 3-2. At this point you've been handed the game. You have a huge initial start, your 2 back men are out of opponents quadrant and you have a nice wall built that is hemming opponent in your home quadrant.
All the dice rolls used were totally random but by picking and choosing when to assign them you can swing just about any game.
The algorithm wouldn't be particularly difficult. If system wants a player who is winning to get held up, and that player gets a blot knocked off and the weaker player only has the 6 point covered in his home quadrant then the algorithm just looks at the remaining available pre-rolled dice throws to see if there is a double 6 available. If there is, choose that, then the good player remains stuck on the bar.
This whole thing can be as simple or complicated as a programmer cares to make it. You can give a player a huge edge or a slight edge. All the time using true randomly generated numbers. This technique applies to just about any online game.
Generate the random numbers for dice or cards or whatever BEFORE the game starts and then allocate them as needed to make certain people win or lose and thereby maintain a level of control of balance on across the membership.
Again not saying this is happening here. Just giving examples of how sites can rig outcomes whilst keeping the numbers random.
