n=7
k	CRBN	ARBN	DGARBN p=4	DGARBN p=7			particular networks do not obey any order: different updatings can be more or less "ordered", although the averages do change
0	-0.14286	-0.14286	-0.142857143	-0.142857143			high variance at the edge of chaos… (we do not exhaust all possible initial states)
1	-0.12143	-0.06393	-0.097857143	-0.091785714			chaotic or ordered nets have low variance, except for chaotic determ.-asynch.
2	-0.03929	-0.03714	-0.061428571	-0.023214286			***check inflexion point for different n's (k=2..3)
3	0.077857	0.03	0.055	0.042857143			as we increase n, the variances decrease, so actually we need less examples, which is good…
4	0.193929	0.105357	0.187142857	0.133571429			larger nets, even for the same k, since they have more elements, can affect other elements as well, since they can do so indirectly… therefore large nets in general are less stable than smaller ones
5	0.213929	0.1075	0.207857143	0.156071429			as n increases, asynchronous become more and more similar, independently of determ… not so for synch.
6	0.285	0.114286	0.254285714	0.2025			larger nets with k=1 almost like k=0, very few dynamics (explains results of Mesot and Teuscher, all nets, for n=200 have behaviour almost static (k<=1), or chaotic (k>=2))
7	0.290714	0.157857	0.241428571	0.238928571			higher variance always around k=3… does this mean higher potential for diversity? (k=3 most skewed distribution of point attractors (Harvey and Bossomaier, 1997))
							non-synch RBNs have positive delta for high n, k=2… would this still be "complex"? Transitions order/chaos smooth, since based on averages (for k=5 you still have small prob. Of finding ordered nets)