The most developed concept for realizing continuous energy (electricity) production from fusion reactions is the tokamak, based on an axisymmetric toroidal plasma. In recent years there has been considerable development of databases and accumulation of knowledge of the behavior of tokamak plasmas around the world, and these make it possible to design an experimental fusion reactor for energy production. However, a degree of uncertainty still exists in predicting the confinement properties and plasma performance in such a device. A precise theory of the classical collisional transport losses has been developed (Section 2.2). Since this does not completely explain the transport processes across magnetic surfaces, additional processes driven by plasma turbulence are required to be taken into account. Significant theoretical efforts are being devoted to understanding the cross-field transport in tokamaks due to the turbulence and a few models are broadly consistent with present experiments. On the other hand, since tokamaks with a range of sizes, operating parameters and heating powers have been constructed, empirical scaling laws derived from these are useful for predicting plasma performance in any new device. Furthermore, the empirical scalings serve as a benchmark for theoretical models. One expects that predictions with such scaling laws will be improved if one imposes dimensional constraints on the form of these laws in the scaling studies. It is also recognized that transport codes solving radial transport equations numerically are also useful for obtaining quantitative predictions. As a result, three approaches are being pursued at the moment for providing predictions for confinement: these are (a) derivation of empirical global scaling laws, (b) non-dimensionally similar studies, and (c) one dimensional transport modeling codes.