In this work we obtain semi-empirical expressions for the probability to find solution of Quantum random walk search algorithm on hypercube when the traversing coin is constructed by generalized Householder reflection and an additional phase shift. Calculations are made for specific relations between coin phases (first from the reflection and second from the shift), that are obtained in our previous works. The quantum algorithm could be made more robust against phase errors, if an experimental implementation preserves those relations. The results from numerical simulations of quantum random walk search algorithm on hypercube are used to find an empirical formulas for algorithm’s probability to find solution. Those formulas, obtained by fitting data to a suitable function, are used to make prognosis about the algorithm’s robustness when dimension of the coin is too large to be simulated on classical computer. Here, we explicitly show all steps of the method used in the paper, while also discuss its advantages and limitations. The goal of this work is to help the experimental implementation of the quantum random walk search algorithm by giving evaluation of the algorithm’s robustness and the obtained probability to find a solution when our design of the walking coin is used.