This paper investigates the statistical characteristics of a box-fitting algorithm designed to detect periodic transits of extrasolar planets in stellar photometric time series. The algorithm searches for signals characterized by a periodic alternation between two discrete levels, with a longer duration at the higher level. The authors present both numerical and analytical results to predict the detection significance at various signal parameters, emphasizing that the effective signal-to-noise ratio (SNR) is crucial. They find that a SNR exceeding 6 is necessary for significant detection. The box-fitting algorithm is shown to outperform other methods, such as the Phase Dispersion Minimization (PDM) and Discrete Fourier Transform (DFT), especially at low signal-to-noise ratios. The paper also discusses the algorithm's performance in different scenarios, including varying transit phases and noise levels, and compares it with other methods like the L-K method and multifrequency Fourier fit. The authors conclude that the box-fitting algorithm is a highly efficient tool for analyzing transit-type signals, particularly in cases with low SNR and large numbers of observations.This paper investigates the statistical characteristics of a box-fitting algorithm designed to detect periodic transits of extrasolar planets in stellar photometric time series. The algorithm searches for signals characterized by a periodic alternation between two discrete levels, with a longer duration at the higher level. The authors present both numerical and analytical results to predict the detection significance at various signal parameters, emphasizing that the effective signal-to-noise ratio (SNR) is crucial. They find that a SNR exceeding 6 is necessary for significant detection. The box-fitting algorithm is shown to outperform other methods, such as the Phase Dispersion Minimization (PDM) and Discrete Fourier Transform (DFT), especially at low signal-to-noise ratios. The paper also discusses the algorithm's performance in different scenarios, including varying transit phases and noise levels, and compares it with other methods like the L-K method and multifrequency Fourier fit. The authors conclude that the box-fitting algorithm is a highly efficient tool for analyzing transit-type signals, particularly in cases with low SNR and large numbers of observations.