The strategies  identified to generate experimental designs capable to analyze tens of variables at a time using a full quadratic regression model with the minimum number of necessary runs are shown in this figure, for 10, 20 and 50 variables. As the number of variables increases, many of these strategies become unfeasible.