Some 90% of antibiotic use is in the community, rather than hospitals, and much of that is to treat acute bacterial infections that would normally be cleared by the patient's immune defenses. Nevertheless, the "rational" (as opposed to purely empirical) design of antibiotic treatment regimes focuses on the worst-case situations, immune compromised patients who, in the absence of these drugs, may well succumb to these infections. Moreover, as is common in clinical medicine, the rational design of antibiotic treatment regimes suffers from what can be described as parametric reductionism, the use of a single parameter, the minimum inhibitory concentration of the drug, the MIC, as both a measure of the pharmacodynamic (PD) properties of antibiotics and bacteria, and the criteria for classifying bacteria as susceptible/resistant to that antibiotic. In this presentation with the aid of mathematical and computer simulation models and a some data, I will argue: (1) that a more comprehensive consideration of the PD of antibiotics that includes the phenomena of hetero-resistance and tolerance may well account for treatment failure of infections that, by the MIC criterion, the target bacteria are deemed susceptible. (2) For the treatment of acute infections in patients with adequately effective innate immune systems, (a) neither acquired (evolved during treatment) inherited or phenotypic resistance (persistence) will lead to treatment failure, (b) low doses and short terms of antibiotic administration can be as effective as high doses and long terms of administration, and (c) bacteriostatic antibiotics can be as effective as highly bactericidal drugs. I will conclude with a discussion of how the hypotheses presented in this talk can be tested (and rejected).