ORIGINAL: Cohen
Has the quality of the generator been tested statistically ? Not only the mean, standard deviation, and more generally output repartition, but also the correlation between successive rolls ?
Sorry but here there is little to test - on a tabletop game I pick the dice and roll it.
Here a number between 1 to 10 is to be picked each time.
That's all that is required for that department.
No, that's not true.
You would not want a generator that generates 1-2-3-4-5-6-7-8-9-10-1-2-3-4-5-6-7-8-9-10-1-2-3-4-5-6-7-8-9-10...
In the same way, you wouldn't appreciate an opponent that would know how to roll the die to get, say twice more tens than you roll.
Or, in an air-air combat, you'd not like a generator that rolls appropriate means and standard deviation, but reserves the 8-13 range (about 50% occurence) for your rolls, and the other rolls for your opponent, such as:
8-14-10-7-12-6-9-15... looks correct, but is really badly twisted.
You want something that has the look and feel of randomness. Here, rolling four 1 in a row looks alittlke strecthed, but not impossible : it still calls for an investigation. The same happens when I roll a sequence of attacks and in five or six combats, don't get anything above 11. Possible, but this let my eyebrows rise: that's within a 1% chance of occurence ; it smells. Yet, it cannot be discounted just because it looks weird : with randomness, weird things happen.
