Non-pharmaceutical interventions are the current central strategy to stop transmitting the novel coronavirus disease (COVID-19) globally. Despite remarkably successful approaches in predicting the ongoing pandemic's spatiotemporal patterns, we lack an intrinsic understanding of the travel restrictions' efficiency and effectiveness. We fill this gap by examining the countries' closeness based on disease spread using country distancing that is analogical to the effective resistance in series and parallel circuits and captures the propagation backbone tree from the outbreak locations globally. Our method estimates that 53.6\% of travel restrictions as of June 1, 2020, are ineffective. Our analytical results unveil that the optimal and coordinated travel restrictions postpone per geographical area by 22.56 [95\% credible interval (CI), 18.57 to 26.59] days of the disease's arrival time and protect the world by reducing 1,872,295 (95\% CI, 216,029 to 23,606,312) infected cases till June 1, 2020, which are significantly better than the existing travel restrictions achieving 12.87 (95\% CI, 10.59 to 15.17) days of arrival time delay and 861,867 (95\% CI, 238,250 to 3,879,638) infected cases reduction. Our approach offers a practical guide that indicates when and where to implement travel restrictions, tailed to the real-time national context.