It was supposed to demonstrate how easily fooled we, as a species, are by probability.
Probability just doesn't work as we intuitively think it does. Improbable things happen all the time, and very probable things often fail to happen.
You can't judge the probability of winning on Clixgrid over the course of a few weeks, your results will only average out over the long term. Until you reach tens of thousands of clicks, your results will be swingy - one run of either good or bad luck will skew your results dramatically until you have a large range to average them out - that's why in the other thread I suggest aggregating the stats shared by the entire community to find your expected long term value.