A large number of various acceptable examples assures a sufficient basis to evaluate the relevance of all known factors. If acceptable variants mistakenly are not included in the example base the relevance values tend to be higher than adequate.
Adding those missing variants to the example base would make the fulfillment of specific values less compelling. Those unregistered values that lead to success would make look specific values more tolerable and thereby these factors less important. The exact proportion of the variants in the example base compared to all those theoreticly observable is not known. An unsufficient example base only can be stated in retrospect by delivering a faulty analysis.
The step of including formerly ignored factors is an essential part of the analysis. The range of possibly interesting factors begins with those factors that are known to be important what is of little further help. The range ends with a random selection of untested factors. That costs much effort. As the number of factors to be analysed increases the number and diversity of examples required to come to conclusions increases as well.