need to add pox of having one or two regressor parameters per node

[domenicodigangi]

Done, from first tests the estimates seem to work reasonably well even with missp dynamics for the fitnesses