The initial empiric therapy for pneumonia was started based on knowledge of the epidemiological data of predominant pathogens in specific clinical settings and the local patterns of antibiotic susceptibility. The Japanese guidelines for empirical treatment of patients with pneumonia provide recommendations for antibiotic selections that are primarily based on age, presence of underlying diseases, and severity of pneumonia. Moreover, in cases of old or healthcare-associated patients, multidrug-resistant organisms (MDROs) are also important predictors of patient morbidity and mortality. Younger patients with a mild disease can be adequately treated with a drug that covers either bacterial or atypical pathogens alone, which is diagnosed by criteria provided by the Japanese CAP guideline. Older patients or those with preexisting illnesses, including hospitalized patients, are more likely to have infections with MDROs, and any antibiotic therapy should be aimed primarily against these organisms. For nonresponding patients, empiric antibiotics may need modification to escalation, especially in younger and moderate patients. When a patient is responding appropriately, therapy can be deescalated or narrowed if an anticipated organism is not recovered or if the organism isolated is sensitive to an antibiotic with a spectrum that is smaller than the one used in the initial regimen. Considering the will of the individuals and the aging of modern Japanese society, we must again consider the concepts of “escalation” and “deescalation” of antibiotic selections for treatment of pneumonia.